Seismic Performance of Bonded Post-Tensioned

Slab-Column Connections with and without

Drop Panels

by

Unnop Prawatwong, Pennung Warnitchai and Chondro Hidayat Tandian

Reprinted from

Advances in Structural Engineering

Volume 15 No. 10 2012

MULTI-SCIENCE PUBLISHING CO. LTD.

5 Wates Way, Brentwood, Essex CM15 9TB, United Kingdom
 Seismic Performance of Bonded Post-Tensioned
Slab-Column Connections with and without
Drop Panels

Unnop Prawatwong*, Pennung Warnitchai and Chondro Hidayat Tandian

School of Engineering and Technology, Asian Institute of Technology, Thailand

(Received: 20 July 2011; Received revised form: 1 February 2012; Accepted: 7 February 2012)

Abstract: Two 3/5-scale models of bonded post-tensioned slab-column connections,

one without drop panel and one with drop panel, were tested to failure. The study’s
goals were to: (1) develop information on the seismic performance of typical bonded
post-tensioned slab-column connections; and (2) improve the seismic performance
using a drop panel. A lateral quasi-static cyclic loading routine, simulating earthquake
actions, was applied to the models’ top columns. Overall performance was examined
and compared in terms of lateral load-carrying capacity, maximum drift, stiffness,
energy absorption, and failure mechanism. Relevant design equations suggested by
ACI 318–08 Building Code provisions for preventing stress-induced and deformation-
induced failures as well as previous similar tests by others were compared with test
results. Results indicate that ACI provisions could be used for these connections, and
the presence of a drop panel is an effective way to greatly enhance the overall
performance of this type of connection.

Key words: slab-column connections, drop panel, post-tensioning (PT), flat plate, punching shear, seismic
performance, drift capacity, bonded tendons.

1. INTRODUCTION formwork. The post-tensioning (PT) systems can be

A slab-column frame structure is a popular choice for
the gravity load carrying system of multi-story
buildings in many countries, primarily due to its ease of
construction and architectural and serviceability
reasons. There are two types of slab, reinforced
concrete slab and post-tensioned concrete slab. Post-
tensioned concrete slabs have become increasing
popular since they have several advantages over
reinforced concrete slabs. Post-tensioned slabs are more
economical because they use prestressed steel with a
much higher tensile strength than normal reinforced
steel, they allow for larger spans and greater
slenderness, they make it easier to control deflections
and cracking, and they allow considerable reduction of
construction time due to earlier striking of slab
further divided into bonded and unbonded systems,
depending on whether the post-tensioning tendon ducts
are filled with cement grout to provide a bonded
structure or coated, wrapped, greased, or plastic covered
to prevent bond. In some countries such as the USA
and Canada, the majority of PT flat plates are
constructed using the unbonded system due to
construction speed and other economic considerations
(Hawkins 1981). In other countries, the bonded system
is preferred since the bond between the concrete and the
tendons offers more flexibility regarding structural
modifications such as making new openings in the slab
for stairwells, utility access, and future expansion.
Since the early 1990s, PT flat plates with bonded
tendons have been widely adopted in Thailand for

*Corresponding author. E-mail address unnop@sut.ac.th; Fax: +66-2524-6059; Tel: +66-2524-6415.

Advances in Structural Engineering Vol. 15 No. 10 2012 1653

 Seismic Performance of Bonded Post-Tensioned Slab-Column Connections with and without Drop Panels
medium to high rise buildings. This is similar to the
situation in many other countries. The primary
reinforcement for this type of structure is draped
tendons that are banded in one direction within the
column strip and uniformly distributed in the orthogonal
direction. To ensure flexural continuity and to control
cracking, top deformed bars are provided as
supplementary reinforcement in peak negative moment
regions, i.e., in the vicinity of the columns. Slab bottom
reinforcement bars are arranged in an orthogonal mesh
to satisfy a minimum reinforcement for temperature and
shrinkage effects. There is no specific requirement for a
continuous bottom bar passing through the column to
protect against progressive collapse after punching
shear failure. Furthermore, due to the large size of
tendon ducts prepared for cement grout and the
congestion of reinforcement bars in the column section,
PT tendons are normally arranged so that none of them
passes through the column. These flat plate structures
are commonly designed primarily for gravity loads,
while they are coupled with concrete shear walls for
resisting lateral wind load. The flat plate structures are
normally neither designed for lateral seismic load nor
checked for deformation compatibility with the shear
walls to ensure their ability to undergo the maximum
earthquake-induced lateral drift without losing their
gravity load carrying capacity.
Lessons from past earthquakes such as the 1964
Alaska, 1967 Venezuela, 1971 San Fernando, and 1994
Northridge earthquakes have demonstrated the
vulnerability of slab-column connections and flat plate
systems to strong ground motion and raised serious
concerns about the adequacy of similar construction. It is
widely known that slab-column connections are the most
critical regions in a flat plate system. Under a strong
earthquake ground motion, brittle punching shear may
occur at a slab-column connection region due to a
combination of direct gravity shear and eccentric shear
from earthquake-induced unbalanced moment between
slab and column. In addition, extensive cracks in the
connection region caused by repeated reversals of large
lateral deformation may significantly deteriorate the shear
strength of the connection. Hawkins and Mitchell (1979)
have shown that the punching failure at an interior slab-
column connection can sometimes initiate a progressive
collapse throughout the entire structure. In addition,
Mitchell et al. (1986) pointed out that such failure was the
primary mode of failure in many waffle-slab and solid-
slab buildings during the 1985 Mexico earthquake. A
thorough understanding of the seismic performance of
slab-column connections is therefore necessary.
Over the past four decades, most experimental research
studies on the seismic performance of slab-column
1654
connections have focused on conventional reinforced
concrete (RC) slabs. Several scholars such as Pan and
Moehle (1989, 1992), Robertson and Durrani (1992), Luo
and Durrani (1995), Moehle (1996), and Hueste and
Wight (1999) have pointed out that the magnitude of
gravity shear stress at the slab critical section (adjacent to
the column) has a major influence on the lateral drift level
at which a punching failure occurs. Meanwhile, a number
of tests have indicated that the lateral drift capacity of the
slab-column connections can be markedly increased by
the use of shear reinforcement. Moreover, a superior
lateral drift capacity and ductility over those of
conventional stirrups can be obtained by stud-shear
reinforcement (SSR), which is a series of vertical bars
with mechanical anchorage at both ends. These findings
have formed the basis of current code provisions and
guidelines for the design of slab-column connections
under a combination of gravity and lateral loading.
There have been far fewer experimental investigations
for PT slab-column connections than for RC slabs, and
all PT slab-column tests have been carried out on the
unbonded system. Therefore, the seismic performance of
bonded PT slab-column connections has never been
examined. Few guidelines and little information are
available to designers. In addition, the design of slab-
column connections for punching shear can be governed
by the thickness of the slab. To achieve an economic
design, drop panels are frequently used to increase the
punching shear capacity of PT slab-column connections.
However, there are no experimental investigations of PT
slab-column connections involving drop panels
subjected to earthquake-type loading. There are only a
few investigated seismic responses of RC slab-column
connections using drop panels or shear capital, such as
those reported by Wey and Durrani (1992), Megally and
Ghali (2000), and Dovich and Wight (2005). It is clear
that investigations are required to establish information
and guidelines for the design of bonded PT slab-column
connections subjected to earthquake-type loading.
In this paper, an experimental investigation into the
seismic performance of two PT interior slab-column
connection specimens with a bonded tendon system is
presented. The first specimen, named S1, was carefully
designed and constructed to represent a typical PT slab-
column connection with no drop panel designed in
Thailand. The second specimen, named S2, was an
improved design of S1 by adding a drop panel. Each
specimen was subjected to a lateral quasi-static cyclic
loading routine to investigate its seismic performance
through the elastic and inelastic ranges and finally until
failure.
The results show that the presence of a drop panel
greatly improves the overall seismic performance of this
Advances in Structural Engineering Vol. 15 No. 10 2012
Unnop Prawatwong, Pennung Warnitchai and Chondro Hidayat Tandian

type of connection. Overall performance in terms of
hysteretic response, lateral load-carrying capacity,
maximum drift, stiffness, and energy absorption were
discussed. Furthermore, the tests provide data on
connection behavior (concentrating on the observed
failure mechanisms and their zones of development
during the different loading up to the ultimate
resistance) as well as the relative influence of gravity
shear stress (at the slab critical section) on the lateral
drift level at which punching failure occurs. Finally, to
assess the reliability of current code design methods, the
relevant design equations suggested by ACI 318-08
Building Code provisions for preventing stress-induced
and deformation-induced failures were compared with
test results. Results indicate that ACI provisions could
be used for these connections.
2. TEST STRUCTURES AND EXPERIMENTAL
SETUP
2.1. Key Structural Indices of Slab-Column
Connections
To investigate the seismic performance of typical
bonded PT slab-column connections, it is important to
make sure that the model S1 represents typical
connections as much as possible. For this purpose, the
construction drawings of five typical buildings in
Bangkok, the capital city of Thailand, were collected,
and key structural parameters associated with the
seismic behaviors of slab-column connections were
computed and are presented in Table 1. These key
parameters are herein called “structural indices.” They
are: gravity shear ratio (Vg /V0), critical section
perimeter-to-depth ratio (b0/d), side ratio (b1/b2),
column width ratio (c1/c2), prestressing ratio (fpc / fc′ ),
and negative moment reinforcement ratio (ρ).
In the above, Vg is the direct gravity shear force
acting on the perimeter of the critical section at d/2 from
the column faces [Figure 14(a)]; the gravity load (Vg)
here is the dead load (without load factor) plus the live
load expected to be present at the time of earthquake
(likely live load), which is set equal to 30% of the design
live load on the floor according to ATC-40 (ATC 1996);
V0 is the slab punching strength in the absence of
moment transfer as defined in ACI 318–95; b0 is the
perimeter of the critical section; d is the distance from
extreme compression fiber to the centroid of
longitudinal tension reinforcement; 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;
c1 is the maximum value of the rectangular column
width; c2 is the minimum value of the rectangular
column width; fpc is the compressive stress in concrete
slab at the centroid of the cross-section due to
prestressing force; f c′ is the specified compressive
strength of concrete; ρ is the ratio of total area of top
reinforcement bars to (c2 + 3h).d; and h is the overall
thickness of the slab.
By setting these indices of the model equal to those of
typical actual structures (as shown in Table 1), the
performance of the model is expected to be similar to
that of the actual structures. Further details on the
structural indices are referred to in Prawatwong (2012).
Among these indices, the gravity shear ratio (Vg /V0)
appears to be the most important, as many past studies
(Pan and Moehle 1989; Kang and Wallace 2006) have
pointed out that the lateral drift capacity (at which a
connection punching failure occurs) generally drops as
the gravity shear ratio increases. For this reason, to
enhance the lateral drift capacity, the second
specimen—S2—was designed to have a drop panel, by
which the gravity shear ratio was reduced from that of

Table 1. Structural indices of slab-column connections in five representative buildings and those of specimen
without drop panel

Structural indices
Vg b1 c1 f pc
Span Column size b0
Buildings (mm) (mm x mm) V0 d b2 c2 fc' ρ
Office 1 8000 400 × 1000 0.267 21.5 2.07 2.50 0.262 0.014
Office 2 7000 400 × 800 0.296 19.0 1.71 2.00 0.316 0.011
Office 3 8000 400 × 800 0.287 19.0 1.71 2.00 0.262 0.009
University 8000 500 × 800 0.231 18.8 1.44 1.60 0.326 0.007
Hospital 8400 500 × 700 0.366 18.3 1.30 1.40 0.270 0.008
Standard deviation 552 – 0.050 1.25 0.30 0.42 0.031 0.003
Average value 7800 – 0.289 19.3 1.65 1.90 0.287 0.010
Specimen S1 5000 250 × 500 0.280 18.3 1.70 2.00 0.268 0.010

Note: fpc and f c′ unit in MPa.

Advances in Structural Engineering Vol. 15 No. 10 2012 1655

 Seismic Performance of Bonded Post-Tensioned Slab-Column Connections with and without Drop Panels

S1 by 54%. ACI 318 building code was used as the slab spans between adjacent columns on all four
guidelines for designing the drop panel and sides were modeled. Pin connections were attached to
supplementary reinforcing bars for this specimen. the points of contra-flexure under lateral loading. This
model of connection was designed to produce bending
2.2. Design and Construction of Specimens moment and shear of the slab comparable to the
S1 and S2 were designed as 3/5-scale models of prototype in the vicinity of the column where the most
prototype interior slab-column connections as shown in damage was expected. The validation of this model and
Figure 1. As each specimen was developed based on the the assumption for interior slab-column connections are
assumption that inflection points in the interior well explained by Pan and Moehle (1988) and by
connection under earthquake-type loading occur at slab Qaisrani (1993).
mid-span and column mid-story, half the total height of Each of the specimens in the series was identical in
an interior column above and below the slab and half of slab dimension, column dimension, tendon layout, and
prestressing forces. The specimens were of normal
weight concrete, 5700 mm square, 120 mm thick, and
Sub-column 250 × 500 mm column in the center of the panel
connection
(Figure 2). The major variable for S2 was the
3 m. additional drop panel, which decreased gravity shear
3 m.
ratio in the column region. The additional drop panel
in S2 was 1600 mm square and projected below the
3 m. slab 80 mm. Its dimensions conformed to those
required in Section 13.2.5 of ACI 318–08. Figures 3
8.4 m. 8.4 m. 8.4 m. 8.4 m.
and 4 show the reinforcement details of each of the
(a) Full scale prototype specimens. A summary of the main properties of the
specimens is given in Table 2.
Lateral load Moment diagram The PT tendons in the model were ASTM A-416,
Grade 270, 12.7 mm (1/2 in) diameter, seven-wire stress-
1.8 m. relieved strands. Since the flexural strength and general
behavior of the slab can be greatly affected by the
distribution of tendons and supplementary reinforcement
5.0 m.
in the specimens, eight tendons were banded in the
direction of loading (N–S direction) while the other eight
(b) Test specimen
tendons were uniformly distributed in the orthogonal
Figure 1. Relation of specimen to prototype structure direction (E–W direction). In each direction, there were

5700

N Loading

Reaction wall
1600 MTS 500
actuator
250

500
200

120

160

Pivot 840
1600
5700

Pin-ended bar 840

Strong floor
Hinged

Dimension in mm
2500 2500
5700
Elevation
Plan view

Figure 2. Interior slab–column connection specimen with drop panel and its dimensions

1656 Advances in Structural Engineering Vol. 15 No. 10 2012

Unnop Prawatwong, Pennung Warnitchai and Chondro Hidayat Tandian

5700
480 3 × 700 540 3 × 700 480 120 5 × 550 5 × 550
CL

Strain gauges

5 × 550
S11
3 × 350

S10
S9
S1 S2 S3 S4 S5 S6 S7 S8
CL
CL
2 × 500 2 × 500
14.5

900 900
290 290 (b) Bottom bonded reinforcement
(10 mm dia. ribbed bars)
20

(a) Distribution of prestressing strands (12.7 mm dia.)

Anchorage plate

N
12-DB16
9 × 80 Column top view

Prestressing strand
120

φ 12.7 mm (0.5 in)

C 2 + 3hslab
7 × 80

RB 6 @100

250
RB 6 @100
500
Column main bars 12-DB16
(16 mm. dia. ribbed bars)
C 1 + 3hslab
500
(c) Top bonded reinforcement
(10 mm dia. ribbed bars 2000 mm long) (d) Column reinforcement and prestressing in column

Figure 3. Reinforcement for specimen S1 (dimension in mm)

no tendons passing through the column, similar to typical
connections in the representative buildings. Each
tendon was inserted into a flat (20 mm in height)
galvanized duct to prevent bonding to the concrete
before prestressing. To prevent damage due to high
concentrations of stresses at the edges of the slab, an
edge beam with sufficient reinforcing bars was provided
on all sides of the slab. Note that the edge beam is only
40 mm thicker than slab and 150 mm wide, and hence
contribute only about 4.1% of the total flexural rigidity
of the entire slab. Moreover, since the behavior of slab-
column connection is mostly influenced by the
deformation of slab in the vicinity of the column, the
effect of the edge beam on the seismic performance of
slab-column connection is considered to be insignificant.
In addition, straight, instead of curved profile, tendons
were used in the slab. Since the vertical component of
the prestressing force, Vp, contributed only a small
amount to two-way shear strength as documented by
Section 11.12.2 of ACI 318–95 or Section 11.11.2 of
ACI 318–08, it was considered that the minor difference
in their shapes would not significantly affect the
generality of the test results. Distribution of prestressed
tendons and their profile in slab of S1 are shown in
Figure 3. To maintain the ratio of steel area to concrete
area of the slab in the scaled-down model equivalent to
those arrangements in practice, the number of strands per
duct was reduced from 3 (typical) to 1 for S1 while other
supplementary bonded reinforcement in the slab was
scaled down to 10 mm diameter deformed bars (DB10)
with a specified yield strength of 300 MPa.
Supplementary reinforcement bars in peak negative
moment in the vicinity of columns were provided in the
slab of S1 as shown in Figure 3(c). The top
reinforcements corresponded to a negative moment
reinforcement ratio of 0.010 within an assumed
effective width of c + 3h in each direction, where c is
column dimension measured in the transverse direction
of the top reinforcement, and h is the overall thickness
of the slab. These bars were cut off at a distance of 1000 mm
from the column center. In addition, minimum areas of
temperature and shrinkage reinforcement were provided
(ρbottom = 0.0016) in an orthogonal mesh as bottom
reinforcement bars in the slab as shown in Figure 3(b).

Advances in Structural Engineering Vol. 15 No. 10 2012 1657

 Seismic Performance of Bonded Post-Tensioned Slab-Column Connections with and without Drop Panels

150
150
480 540 480 120 400 400
CL

Strain gauges Strain gauges

S11
S10
S9

500
S1 S2 S3 S4 S5 S6 S7 S8 B1
CL
80 CL
900 530 470 900
145

470 530

50
350 350 (b) Bottom bonded reinforcement
1600 (10 mm dia. ribbed bars )

RB 9 @95

250
2RB 9 @95
20

(a) Distribution of prestressing strands (12.7 mm dia.) 500

Column main bars 8-DB25
300 300 (25 mm. ribbed bars)
N
160

Anchorage plate

1550
Bottom layer 140
150

8-DB25
100 320
150

1550
Wdrop + 3hslab

200
Prestressing strand
12.7 mm (0.5 in)
320

D1

Top layer
150

1500
Wdrop + 3hslab
(c) Top bonded reinforcement (d) Drop panel reinforcement (e) Column reinforcement and prestressing in column
(10 mm dia. ribbed bars 2800 mm long) (10 mm dia. ribbed bars)

Figure 4. Reinforcement for specimen S2 (dimension in mm)

All bar arrangements were in such a way that the top and
bottom bars in the direction of loading were placed at
the outmost layer. A nominal clear concrete cover of
10 mm was specified for both top and bottom
reinforcement. Since the critical section perimeter-to-
depth ratio (b0/d) and side ratio (b1/b2) of the designed
specimen depended on the column dimension and the
effective depth, d., the bar arrangements resulted in the
actual b0/d, and b1/b2 of 18.3 and 1.7, respectively.
These actual structural indices of S1 are listed in the last
row of Table 1.
The drop panel in S2 was reinforced with DB10
(10 mm diameter) deformed bars at the bottom face to
counter tensile stresses caused by positive bending which
might be induced by significant unbalanced moment due
to lateral cyclic loading during the test. The reinforcement
ratio of the drop panel, ρs, drop = 0.003, was the same in
each direction. To prevent anchorage failure of the
bottom reinforcing bars in the drop panel, the vertical legs
with development length of 40 times bar diameter were
extended into the main slab. Figure 4(d) shows the layout
of the reinforcing bars within drop panel.
S2 also contained the supplementary top reinforcement
bars at the top of its slab according to Section 18.9.3 of
ACI 318-95 building code similar to those of S1. The
code requires that an amount of bonded top reinforcement
bars equal to 0.00075 hl, where l is length of span in
direction parallel to that of the reinforcement being
determined, should be provided over the effective width
of the slab near the supporting column or drop panel in
both directions. The supplementary top reinforcement
bars, which were the same number of top bars as in S1,
were distributed along an assumed effective support
width of Wdrop + 3h in the middle of the column in both
directions [Figure 4(c)], where Wdrop is the width of the
drop panel, and h is the total thickness of the slab outside
the drop panel. The assumed effective support width was
according to Section 13.1.2 of ACI 318–08 building code.
The supplementary top reinforcement ratio within the
drop panel corresponded roughly to 0.002 in each

1658 Advances in Structural Engineering Vol. 15 No. 10 2012

Unnop Prawatwong, Pennung Warnitchai and Chondro Hidayat Tandian

Table 2. Summary of parameters and properties of model slabs

Property Slab S1 Slab S2

(1) (2) (3)
1. Dimensions, in mm 5700 × 5700 5700 × 5700
2. Span, in mm 5000 5000
3. Thickness of slab, in mm 120 120
4. Drop panel size, in mm None 1600 × 1600 × 80
5. Cross-sectional area near critical section, sq mm 6.84 × 105 8.12 × 105
6. Column size, in mm 250 × 500 250 × 500
7. Column height, in mm 1800 1800
8. Number of tendons in each direction 8 8
Aps / strand 98.7 sq mm 98.7 sq mm
9. Distribution of tendons in column strip
N–S direction 100% 100%
E–W direction 50% 50%
10. Effective depth of strands, dp, in mm
N–S direction 80 160*
E–W direction 60 140*
11. Concrete strength, in MPa
fci′ at time of stressing 20.0 (4 days) 31.8 (6 days)
fc′ at time of testing 41.1 (43 days) 45.9 (43 days)
12. Steel strength of strands, in MPa
fpu, tensile strength 1902 1947
fpy, yield strength at 1% elongation 1780 1763
13. Steel strength of DB10, in MPa
fu, tensile strength 587 491
fy, yield strength 503 324
14. Average compressive stress fpc, in MPa
P/A in N–S direction 1.72 1.45
P/A in E–W direction 1.72 1.45
15. Gravity shear ratio 0.28 0.13

Note: Effective depth was measured within drop panel

direction. The top bars were extended into the slab around
the drop panel and cut off at a distance of 600 mm from
the edge of the drop panel to provide development length
as recommended in Section 21.7.5 of ACI 318–08. In
addition, two continuous bottom bars were placed over
the column in each direction according to Section 18.12.7
and 13.3.8.5 of the code to prevent progressive collapse
in the event of a connection shear failure.
In each of the specimens, sufficient transverse and
longitudinal reinforcing bars were provided so the
column would remain elastic without experiencing either
flexural or shear failure during the test. In addition,
vertical prestressing forces of 588 kN were applied to the
column by four unbonded tendons to simulate the effects
of gravity loads from the upper floors. The details of
reinforcement for the columns in S1 and S2 are given in
Figures 3(d) and 4(e), respectively.
Each of the specimens was cast with ready-mix
concrete. A number of standard 150 × 300 mm cylinders
were cast from each batch and used to determine concrete
compressive strength. As shown in Table 2, the actual fc′
of S2 was found to be 5 MPa higher than the control
specimen S1 on the test day. This difference was expected
to have a negligible effect on the punching shear strength
of the connection, which is considered proportionally to
the square root of concrete strength in ACI 318–08 code.
The model slabs were prestressed with eight tendons in
each direction when the concrete slab gained sufficient
strength. The tendons were prestressed individually using
a 200 kN hydraulic jack. An effective force of 147.2 kN,
corresponding to a stress of 0.80 fpu, was applied to each
tendon to produce the average prestress levels shown in
Table 2. At the time of stressing, the prestressing force
was gauged by measuring both the jacking pressure
and the elongation of the steel. Since the slabs were
relatively short, the loss of prestressing force due to a
small amount of slip at the wedge-type anchor grips was
significant and needed to be minimized. Therefore, the
loss was carefully checked and compensated such that the
final prestressing stress of 0.80 fpu in each tendon was
attained. After the tendons were prestressed and the end
recesses were filled, all galvanized ducts were grouted to
provide an effective bond between the tendons and
the ducts.

Advances in Structural Engineering Vol. 15 No. 10 2012 1659

 Seismic Performance of Bonded Post-Tensioned Slab-Column Connections with and without Drop Panels
2.3. Test Set-Up and Instrumentation
The specimens were set up with loading and boundary
conditions as shown schematically in Figure 5. Each
specimen was subjected to quasi-static reverse cyclic
loads that simulated earthquake loadings. A reversed
horizontal loading was applied to the top of the column
using a 500 kN capacity servo-controlled hydraulic
actuator, which was mounted to a rigid reaction wall.
The hydraulic actuator was pinned at its ends to allow
rotation during testing. The lower portion of the column
was connected to a concrete strong floor of the structural
laboratory and restrained from horizontal movement in
any direction but allowed to rotate in the direction of
loading. Ten vertical steel edge links that acted as roller
supports were connected to the strong floor on the north
and south sides of the column. The pin-type connections
on both ends of the vertical links would allow rotation
and horizontal translation with no vertical deflection. As
Pan and Moehle (1988) point out, the use of hinge
support at the column base and pin-ended links at the
slab edges resulted in an unstable condition of the
specimens against torsion around the column vertical
axis, so additional torsional restraining systems were
installed in both of the slabs’ north and south sides to
prevent this rigid-body twisting of the specimens. This
consisted of two wire ropes on each side, which
diagonally crossed from either of the slab corners to the
channel firmly anchored to the strong floor.
A load cell and a displacement transducer were
installed at top of the column to monitor detailed overall
force-displacement relations of the connections
throughout the loading history. The lateral forces at the
top column were measured using the load cell, while the
displacements were measured by displacement
transducer No 1 (DT1). In addition, DT4 and DT5,
shown in Figure 5, were installed to monitor horizontal
displacement of the column support and horizontal
displacement of the steel reaction wall relative to the
ground; these displacements were expected to be
negligibly low compared with the column top
displacement. The strain in some post-tensioning tendons
and supplementary reinforcing steel were monitored by
Supporting
Hydraulic
frame
Steel actuator Load cell
reaction wall DT1
Sand bags
DT2
DT5 Strong floor
DT4
Torsional restraining system Hinge support
Figure 5. Test setup showing simulated gravity load
1660
electrical resistance type strain gauges installed on the
bars prior to casting the specimens on locations as shown
in Figures 3 and 4. Further details of instrumentation can
be found in Pongpornsup (2003) and Tandian (2006).
2.4. Loading Protocol and Data Collection
Prior to lateral loading of S1, a superimposed dead load
was introduced to the PT slab by a large number of
sandbags placed on top and hung from the bottom side
of the slab to simulate the gravity shear ratio (Vg/V0) of
0.28 on a critical perimeter located at a distance d/2
around the column. This gravity shear ratio
approximated the average value of those found in slab-
column frame buildings in Thailand as given in Table 1.
The gravity shear at the critical section (Vg) was
determined by an elastic finite element analysis of the
slab-column model under gravity loads. The same slab
gravity loading was applied to S2. But since the slab in
the connection region was thicker due to the presence of
the drop panel, a gravity shear ratio of 0.13 was obtained
for this case. Note that the values of the gravity shear
ratio, Vg /V0, of each specimen in the experimental
program were calculated based on an assumed effective
depth d of 0.8 times the slab or drop panel thickness
(0.8h) and actual concrete strength at the time of testing.
The seismic movement was simulated by applying
lateral displacement to desired drift levels at the top of
the column through the hydraulic actuator as shown in
Figure 5. This loading device was manually operated to
have better control of the lateral displacement
increment. The lateral displacement protocol and the
direction of the positive drift are given in Figure 6. The
lateral displacement protocol was a modified version of
ACI 374.1–05. Drift is defined as the ratio of the lateral
displacement of the column at the lateral loading point
to the total column height (1800 mm). At each target
drift, two complete cyclic displacement loops were
conducted to evaluate the loss of strength and stiffness
of the specimens during the repeated cycles.
8
Column top displacement
6 N (+)
3 4 1 2
4
Target drift (%)
1% drift = 18 mm
2
12
0
4
−2 3
DT3 −4
−6
−8
0 2 4 6 8 10 12 14 16 18 20 22 24 26
Cycle number
Figure 6. Pattern of lateral cyclic displacement
Advances in Structural Engineering Vol. 15 No. 10 2012
Unnop Prawatwong, Pennung Warnitchai and Chondro Hidayat Tandian

(a) S1
250 Linear elastic
Hysteretic response
200 Envelop curve
1 First crack
150 2

Lateral force (kN)

2 Peak load and maximum drift
100
3 Punching failure of slab
1
50
3 3
0
−50 1

−100
−150 2

−200
−250
−7 −6 −5 −4 −3 −2 −1 0 1 2 3 4 5 6 7
Figure 7. General lateral cyclic test picture of bonded PT slab- Specimen drift (%)
column interior connection
(b) S2
250 Hysteretic response Linear elastic
During testing, all data from the instruments were 200 1
Envelop curve
First crack
2
3 5 4

continuously recorded by a data acquisition system. A 150 2

Main flexural crack extended
though the entire width of slab 1

Lateral force (kN)

general picture of a test slab installation in position is 100 3 Peak load
4 maximum drift
shown in Figure 7. To investigate the damage pattern of 50
5 Punching failure of slab
PT slabs near the connections, every visible crack on the 0
slab and column was traced. A grid of squares −50
−100
measuring 100 mm was marked on the slab surfaces to 1
−150
facilitate accurate location of cracks and the capture of 2
−200
crack locations by manual and digital cameras during −250
4 5 3

the tests. After failure of a specimen, the slab S2 in the −7 −6 −5 −4 −3 −2 −1 0 1 2 3 4 5 6 7

series was separated from the column by sawing along a Specimen drift (%)
perimeter located 75 mm from the column faces to Figure 8. (a) Lateral force-drift result of specimen without drop
investigate the failure mechanism. panel; (b) Lateral force-drift result of specimen with drop panel

3. EXPERIMENTAL RESULTS AND

DISCUSSION
Observations and analysis of experimental results
obtained from tests of both specimens are presented and
discussed in the following sections. The discussion
includes the seismic performance of S1 and S2, which
highlight the global and local behaviors of the test
connections under the lateral cyclic loading applied in
this study. Subsequently, further analyses of the lateral
force-drift response in terms of eccentric shear stresses
and drift capacity are discussed. Comparisons with
previous similar tests by others are made. Also, the
applicability of ACI 318–08 Code on slab-column
connection design is evaluated.
3.1. Lateral Force-Drift Relationship
The hysteretic response of each specimen was plotted
using the data recorded at the point of the application of
the actuator. Due to issues such as pin slip at the bottom
base of the column and unexpected movement of the
actuator, however, the recorded hysteretic response for
S2 had to be corrected. The magnitude of the correction
was within 5% of the maximum lateral displacement.
The corrected hysteretic responses of S1 and S2 were
plotted in Figures 8(a) and 8(b), respectively. The
envelop curve of cyclic force-drift response was also
depicted.
The figures show that both specimens displayed long
and narrow hysteresis loops, demonstrating a limited
ability to dissipate energy of bonded PT interior slab-
column connections. In one cycle of lateral drift, each
test specimen behaved similar to a linear elastic
structure with viscous damping. As the drift level
became higher, in general, specimen stiffness degraded
more and the hysteresis loops were wider. No
significant pinching was observed from the hysteresis
loops of either specimen. Both S1 and S2 experienced
punching failure. Punching failure in each of the
specimens was characterized by a sudden drop of lateral
load-carrying capacity to less than 50% of its maximum
force attained during the tests.
S1, without a drop panel, could only withstand 2.0%
drift [Figure 8(a)]. After the maximum lateral load of
107 kN was attained, this specimen suddenly failed in
brittle punching shear and completely lost its lateral
strength and stiffness while no peak load saturation was
perceived in advance. It is likely that the higher gravity
shear ratio in the first test was a key factor that caused
early failure in punching shear in this specimen. S2

Advances in Structural Engineering Vol. 15 No. 10 2012 1661

 Seismic Performance of Bonded Post-Tensioned Slab-Column Connections with and without Drop Panels

(with drop panel) attained 80% higher lateral load-
carrying capacity than the control specimen (S1). As can
be seen in Figure 8(b), S2 exhibited a saturation of peak
load for a drift of 2% to 6%, indicating that flexural
yielding took place long before punching failure. Until
the end of the test, the specimen with the additional drop
panel showed much higher drift capacity at about 6% at
punching failure than the one without the drop panel.
3.2. Stiffness Degradation
Stiffness degradation is commonly used to describe the
stiffness change when a test specimen is subjected to
lateral cyclic loading. The ratio of the average
maximum lateral load to the corresponding column top
displacement in a load cycle is defined as secant
stiffness, as illustrated in Figure 9(a). The secant
stiffness in Figure 9(b) for S1 and S2 was computed
from lateral force-drift results in Figures 8(a) and 8(b)
respectively. For S2, information on the lateral load
stiffness was obtained by analyzing the response of the
specimen under negative drift since some peak load
points in positive loading directions were harmed by
unexpected movement of the actuator as mentioned
earlier.
As observed from the stiffness-versus-drift plots in
Figure 9(b), the initial secant stiffness of S2 is up to
100% greater than that of S1, indicating a significant
effect of the drop panel on the initial rigidity of the
connection. This was expected since increased depth of
the PT slab with the drop panel increased the moment of
inertia of the slab, which increased the stiffness of the
connection. The decrease of the stiffness on the curves
in Figure 9(b) is indicative of stiffness degradation with
an increase in drift level. There is, however, no
significant stiffness degradation for the repeated cycle
of the same drift level. The stiffness values and slope of
the curves in Figure 9(b) within the drift range from
0.25 to 2.0% indicate that the PT connection with drop
panel is stiffer but exhibited a higher rate of stiffness
degradation than S1. The stiffness of S2 degraded
rapidly during 0.25 to 1% drift. The degradation was
mainly associated with the extensive and progressive
flexural cracking and subsequent reduction in the
effective moment of inertia of the section. Within the
drift range from 0.25 to 2.0%, the ratios of the stiffness
of S2 to that of S1 varied in a range from 1.7 (at 2.0%
drift) to 2.0 (at 0.25% drift). The higher stiffness in S2
resulting in higher unbalanced moments were

(b)
16
S1, first cycle
(a) 14 S1, repeated cycle
Lateral stiffness (kN/mm)

Lateral load S2, first cycle

12 S2, repeated cycle
B 10
Secant stiffness
K = Pi /ui 8
ESi Pi
6
A C
4
Energy dissipated Displacement
S1
within cycle i = EDi 2 S2
ui
0
0 1 2 3 4 5 6 7
D Specimen drift (%)
(d) 500
(c)
400
Unbalance momente (kN-m)

14
S1, without drop panel S2, without drop panel 300
Equivalent viscous
damping ratio (%)

12
First cycle First cycle
10 200
Repeated cycle Repeated cycle
8 100
6 0
4 −100
Recommended by
2 −200 S1,Vg /V0 = 0.28
newmark (1982) S1
(without drop panel)
0 −300 S2
0 1 2 3 0 1 2 3 4 5 6 7 S2,Vg /V0 = 0.13
−400
Specimen drift (%) Specimen drift (%) (with drop panel)
−500
−7 −6 −5 −4 −3 −2 −1 0 1 2 3 4 5 6 7
Range of typical values of prestress concrete structure
for linear elastic analysis with classical damping Specimen drift (%)
recommended by newmark (1982) Note: unbalance moment induced by lateral force = lateral force × 1.80

Figure 9. (a) Definition of secant stiffness and energy dissipation capacity; (b) Secant stiffness degradation of both specimens;
(c) Equivalent viscous damping ratios for various displacement cycles; (d) Unbalanced moment at center of connection versus specimen drift

1662 Advances in Structural Engineering Vol. 15 No. 10 2012

Unnop Prawatwong, Pennung Warnitchai and Chondro Hidayat Tandian
transferred from the connection to the columns as shown
in Figure 9(d).
3.3. Energy Absorption Characteristics
Equivalent viscous damping (EVD) ratio is commonly
used to measure a structure’s energy absorption. As
described in Clough and Penzien (1993), the EVD ratio
can be estimated from hysteretic response under fully
reversed cyclic loading as shown below.
1 E Di
ξeq = (1)
4π ESi
where EDi is energy dissipation capacity within loop or
cycle i of measured hysteretic response, and Esi is the
recoverable elastic strain energy stored at maximum
displacement in that cycle. In this study, the value of EDi
was obtained from the area enclosed by the force-
displacement curve under fully reversed cyclic loading
ABCD within loop or cycle i [Figure 9(a)], while
Esi = kui2/2 was calculated from the secant stiffness k
determined by experiment as defined in Figure 9(a). In
the case of S2, the values of EDi, ESi, and ξeq under fully
reversed cyclic loading were estimated using data from
the negative direction from the corrected hysteretic
response in Figure 8(b). Also, the last few cycles
resulting in a drop in lateral load resistance of more than
20% of the peak load were excluded from the
calculation.
The energy absorption of S1 and S2 are depicted in
Figure 9(c), in which the values of equivalent viscous
damping for virgin and repeated cycles were plotted
versus specimen drift. A band of typical values of the
equivalent viscous damping ratio for linear elastic
analysis with classical damping of prestressed concrete
structure at or just below the yield point as
recommended by Newmark (1982) was also overlaid for
reference. As can be seen in Figure 9(c), the values of
equivalent viscous damping ratio from S1 and S2 vary
between 4 to 8%, and 2 to 6%, respectively, at drift
range up to 2.0%. Within the respective drift range, the
damping values of S2 seem to be lower than those of S1,
suggesting that S2 experienced less damage. Figure 9(c)
also shows that the damping ratios obtained from each
first cycle of all specimens at drift range less than 2%
correlate well with the recommended typical values.
However, there are significant degradations of energy
absorption in subsequent cycles when the level of drift
cycle is not increased. This is because of a significant
decrease of energy dissipation capacity but no
significant change of the recoverable elastic strain
energy stored at maximum displacement in the repeated
cycle.
Advances in Structural Engineering Vol. 15 No. 10 2012
3.4. Crack Pattern Development and Damage
Observations
Care was taken to observe the progression of cracking
on the upper slab surface due to shear and moment
transfer near the connection region of the specimens
throughout the tests. S1 was damaged by abrupt
punching. In contrast, S2 with drop panel performed
very well with only minor damage under lateral cyclic
loading until reaching high drift levels. S2 before failure
was very ductile, with distributed flexural cracks
observed at the edges of both the drop panel and the
column.
Figure 10(a) illustrates the progression of cracking
patterns that occurred during the experiment on S1. For
this specimen most crack damage developed within a
limited distance between lines that are 750–800 mm
outside opposite faces of the column. This visible
limited distance [Figure 10(a)] can be approximately
expressed in terms of the member dimension as
c2 + 12h, where c2 is the column width and h is the
overall thickness of the slab. First visible cracking
developed earlier for S1 than for S2 and the width of
initial cracks was very small. They initiated at the lateral
drift of 0.25% as indicated by point 1 in Figure 8(a).
The first visible cracks were transverse cracks along the
faces of the column, longitudinal cracks running in the
direction of loading along the faces of the column, and
diagonal cracks at column side faces. Diagonal cracks
originating near the corners of the column followed
afterward at 0.50% drift. As the drift level increased, all
cracks lengthened and grew in number around the
column within the limited distance. Similar damage also
developed on the opposite side of the connection when
the load was reversed to negative direction during the
test.
While the specimen S1 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 [Figure 8(a)].
After that when the column was being pulled in the
negative direction, 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 10(a). A
photograph of the failure of S1, which had no drop
panel, at the end of test is shown in Figure 11(a). For
this case, however, the punching surface on the slab was
far from the critical section at d/2 from column faces as
assumed by current ACI building design code.
Throughout the test, no cracks were found on the
column of S1. It was also observed that the widths of
cracks marked on the slab surface during the testing of
S1 were small up to failure.
1663
 Seismic Performance of Bonded Post-Tensioned Slab-Column Connections with and without Drop Panels

(a)
Loading direction
Specimen
drift (+)

Specimen
drift (−)

Specimen drift = +0.25% Specimen drift = +0.50% Specimen drift = +0.75% Specimen drift = +1.00%

Failure
surface Loading direction
Specimen
drift (+)

Diagonal
crack
Transverse Specimen
crack drift (−)
Longitudinal
crack c2+12h

Specimen drift = +1.50% Specimen drift = +2.00% Punching shear failure at Punching shear failure at
positive drift = +1.7% negative drift = −0.80%

(b)
Drift (+)

Loading
direction

Drift (−)

Specimen drift = +0.50% Specimen drift = +1.00%

Drift (+)

Loading
direction

Drift (−)

Specimen drift = +2.00% Specimen drift = +3.00%

Drift (+)

Loading
Type 3 direction

Type 1
Drift (−)
Primary negative Edge of
moment yield line Type 2
drop panel
Specimen drift = +4.00% Specimen drift = +5.00%

Drift (+)

Loading
direction

Drift (−)

Specimen drift = +6.00% Final punching shear failure after +6.0%

Types of cracks:1 = transverse slab crack, 2 = longitudinal slab-column interface crack, 3 = slab diagonal crack

Figure 10. (a) Damage progression on specimen S1; (b) Damage progression on specimen S2

1664 Advances in Structural Engineering Vol. 15 No. 10 2012

Unnop Prawatwong, Pennung Warnitchai and Chondro Hidayat Tandian
(a)
Loading
W
direction
(b)
S2
W
(c)
Loading direction S2
Shear band
200 12°
E
550
Figure 11. (a) Punching shear failure of specimen S1 at 2.0% drift;
(b) Punching shear failure of specimen S2 after 6.0% drift;
(c) Cut at plane perpendicular to moment axis in specimen S2 after
punching failure
The progression of cracking patterns on the upper slab
surface of S2 is illustrated in Figure 10(b). Since the
additional drop panel in S2 gave it much more flexural
stiffness than the flat slab-column connection in S1, the
presence of the drop panel had a significant influence on
the cracking patterns. For this specimen, primary flexural
cracks under negative bending generally appeared more
around the edges of the drop panel than at the slab-
column interface. Flexural cracks in the form of
transverse cracks (transverse to the loading direction)
formed in the slab adjacent to the edges of the drop panel
at 0.5% drift and subsequently extended through the
entire width of the slab slowly as the specimen drift
increased to 2% drift. By the end of 2% drift, fewer but
longer cracks were observed on the slab at the column
side faces of S2 compared to S1, and all were hairline in
nature and needed close inspection. After the drift level of
2.5%, the primary flexural crack under negative moment
at the edges of the drop panel widened while the hairline
diagonal cracks outside the drop panel zone started to
grow more extensively. As the drift increased, more
Advances in Structural Engineering Vol. 15 No. 10 2012
cracks appeared both inside and outside the drop panel
zone, and the primary transverse cracks along the edge of
the drop panel widened. These conform well to the
increase of energy absorption shown in Figure 9(c).
Within the drift range from 3.5 to 5.0%, large flexural
deformation and widened transverse cracks along the
edge of the drop panel were obvious. The approximate
crack width at 5.0% drift was 2 mm and can be seen by
bare eyes at 3 meters. However, in the unloading stage,
these widened transverse cracks were able to close and
appeared like thin cracks. This is thought to be the effect
of sufficiently high prestressing force in the slab.
After drift increase to a level above 5.0%, new diagonal
and transverse cracks rapidly developed inside the drop
panel zone as depicted in Figure 10(b). Overall, there were
more severe cracks near the connection inside the drop
panel of S2 than S1. Finally, the specimen suddenly failed
as a result of punching failure of the slab within the drop
panel. A cut at a plane parallel to the loading direction at
75 mm from the column face of S2 after punching failure
is shown in Figure 11(c). It appears that a shear band
initiated within the drop panel zone and extended from the
bottom of the slab at the column face to the slab top
surface at an angle of about 12 degrees to the horizontal.
3.5. Variation of Strains and Curvatures
Strain gauges provided quality information about the
state of stress and strain at discrete locations in
prestressed tendons and supplementary reinforcements.
Strains in tendons and supplementary reinforcements
were recorded at several locations (Figures 3 and 4) to
study shear and moment transfer mechanisms in the
bonded PT slab-column connections. The amount of
strain data in these tests was quite large; therefore, only
results related directly to the objectives of the
experimental program are presented in this section.
Based on the assumption that a perfect bond exists
between prestressed steel and the surrounding concrete
slab, the tensile strains in bonded tendons at the strain
gauge location may be also used as an indicator for
negative bending curvature in the slab by considering
the following expression.
tensile strain
Curvature, ϕ = (2)
depth of neutral axis
For S1, the maximum increase in strains in tendons
during the test was 880 × 10–6 at Gauge S9. With the initial
strain of about 7300 × 10−6 and this increase, the strains in
the tendons did not reach their yield strain (10000 × 10–6).
Thus, all tendons in Specimen S1 remained elastic. Since
1665
 Seismic Performance of Bonded Post-Tensioned Slab-Column Connections with and without Drop Panels

these tendons were straight lines with the same offset from (a) c1
the top slab surface, the distance from the neutral axis of 1200
Drift (%)
bending slab to the strain gauge was likely to be the same 1000 −2.00 −0.75
−1.50

Strain (micro strain)

S5
(approximately) for every gauge. Hence, higher tensile (+) 800
−1.25 −0.50
600 −1.00 −0.25
strain means higher negative curvature. In this way, the
400 S6
tensile strains at Gauges S1, S2, S3, S4, S5, S6, S7, and S8 200
S7
S3
in Specimen S1 were indicative of the bending curvature 0
S1 S2
profile along a section parallel to the loading direction, −200 S8
S4
while the tensile strains at S5, S9, S10, and S11 were −400
indicative of the bending curvature profile along another −600
−250 −200 −150 −100 −50 0 50 100 150 200 250
section that was orthogonal to the loading direction.
Distance from center of column (cm)
Eqn 2 also implies that the compressive (–) strains in
tendons can also be used as an indicator of positive (b) Wdrop
bending curvature. However, since the neutral axis is c1
likely to be closer to the extreme compressive fiber than 1400
Drift (%) S5
the extreme tensile fiber (especially after cracking), the 1200
1000 −3.50 −1.25 S6
S7

Strain (micro strain)

neutral axis distance to tendon in this positive bending −3.00 −1.00
800 −2.50 −0.75
case is likely to be shorter than in the case of negative 600 −2.00 −0.50
−1.50 −0.25
400
bending. Hence, a lower compressive strain does not 200 S1 S8
S2
mean a lower bending curvature, when comparing 0
tensile strains. −200
−400
−600
3.5.1. Distribution of strains and curvatures S4
−800
along the direction of loading −250 −200 −150 −100 −50 0 50 100 150 200 250
Distance from center of column (cm)
Figure 12(a) shows the results of strains measurements
of Specimen S1 in a tendon placed along the direction of (c)
2500
loading to determine the slab zone with high negative Drift (−) Drift (+)
bending curvature. A similar plot for Specimen S2 is
shown in Figures 12(b). The measured strains were 2000 B1 D1
S2
plotted at actual negative drift during the first cycle of The strains in tendons
each lateral drift level up to the drift level of 2.0% for S1 are controled by flexural
Strain (micro strain)

1500 yielding at negative

and 3.5% for S2 since some strain gauges in S2 failed to D1
moment yield line
capture the strain data after this drift level (3.5%). B1
The typical strain profiles in Figures 12(a) and 12(b) 1000
S9
show that the increased tensile strains in bonded tendons
at different locations were non-uniform. As the lateral drift 500
S5 Strain in tendon by S5
S7 Strain in tendon by S7
level of the specimen increased, the tensile strain along the Strain in tendon by S9
length of the bonded tendon generally increased due to the Strain by B1
Strain by D1
increase of negative bending curvature of the slab near 0
0 −1 −2 −3 −4 −5 −6 −7
the connection region. The strain profiles also show that
for S1 the high negative bending curvature concentrates −500
near the column face, but for S2 the high negative bending Specimen drift (%)
curvature spreads out over the drop panel zone and extend
Figure 12. (a) Strain variations in tendon in direction of loading at
far outside the drop panel. The largest tensile strain in the
different drift ratio for specimen S1; (b) Strain variations in tendon
tendon along the direction of loading was detected from
in direction of loading at different drift ratio for specimen S2; (c)
Gauge S5 near the slab-column interface for S1
Strain variations at some selected gauges across vertical crack at
[Figures 12(a)]. By the additional drop panel, the strain
edges of drop panel for specimen S2
reading from Gauge S5 slightly decreased from
approximately 800 × 10–6 in S1 to approximately 700 × 10–6
in S2 at the drift level of 2.0% while the distance from the curvature of the PT slab near the critical region of the
neutral axis was significantly increased. If Eqn 2 is connection at the drift level of 2.0%. This correlates well
applied, this means that the presence of the drop panel in with the minor damage observed within the drop panel
S2 helped to significantly decrease the negative bending region of S2 from the drift range of 0.25 to 2.0%.

1666 Advances in Structural Engineering Vol. 15 No. 10 2012

Unnop Prawatwong, Pennung Warnitchai and Chondro Hidayat Tandian

Note that the strain profiles along the direction of 1.2 S9

loading [Figures 12(a), 12(b)] obviously presented a
S5
significant difference between the strain behavior of Drift (%)
1.0
bonded and unbonded tendons in PT flat plate slab- − 2.00
− 1.50
column connections. In unbonded systems, similar tests − 1.25

Measured strain/S5 strain

by Smith and Burns (1974) and Qaisrani (1993) showed
that the change in tendon force (or stress and strain)
prior to failure is relatively small since the strain in a
tendon is averaged between the end anchorages and is
approximately equal throughout the tendon’s length.
Their results also showed that the increase in tendon
stress (and strain) becomes significant only after a major
failure – the column was punched through and the slab
dropped, forcing the tendons to change shape.
Figure 12(c) shows the variations of strains in the
tendon (Gauges S5, S7, and S9) and in bottom
reinforcement bars (Gauges B1 and D1) with the drift
level. Gauges B1 and D1 are located near the edge of drop
panel and the column face, respectively (Figure 4). Strain
saturation at these two locations (B1 and D1) can be
observed when the lateral drift ratio exceeds 2%. Similar
strain saturation is also observed for Gauges S5, S7, and
S9 when the lateral drift ratio exceeds 3.5%. These strain
saturations in tendon and bottom reinforcement bars
suggest that there might be flexural yielding at certain
locations which, judging from the crack pattern, are most
likely the edges of the drop panel and/or the column faces.
After yielding, a further increase in lateral drift could
result in localized plastic rotations at these flexural
yielding locations without any significant increase in
bending curvature at other locations. Since the measured
strains depend on the bending curvature at gauge
locations, the strain saturations could be caused by these
flexural yieldings. The strain results further suggest that
flexural yielding by positive bending moment occurred
first at around 2% drift and was later followed by flexural
yielding by negative bending moment at about 3.5% drift,
where a complete plastic mechanism was attained. This is
consistent with the observation in force-deformation
relationships where S2 appeared to have the first
significant yield at 2.0% drift but attained maximum
lateral strength at drift around 3.5% as shown in
Figure 8(b).
3.5.2. Distribution of strains and curvatures
across the width of PT slab
To investigate the distribution of negative bending
curvatures across the width for a typical bonded PT
interior slab-column connection, the recorded strains at
Gauges S5, S9, S10, and S11 (Figure 3) mounted on the
tendons across the width of S1 were plotted at each drift
level in terms of the ratio of the measured strain to the
strain recorded by Gauge S5 as shown in Figure 13.
− 1.00
0.8
S10 − 0.75
− 0.50
− 0.25
0.6
(c2 + 3h)/2 S11
0.4
c2/ 2
0.2
Column face
Lb* is distance from column face
0.0
0 1 2 3 4 5 6 7 8 9 10
Lb*/ hslab
c2/2 + 6hslab
CL
Figure 13. Strain variations in tendon across the width of slab at
different drift ratio for specimen S1
Results indicate that the magnitudes of tensile strain (and
negative bending curvature) at each level of specimen
drift were highest at tendons placed within a region of
370 mm (or in terms of slab dimension 3h) from the face
of the column and gradually decreased for tendons
located further away from this region. The tendon at
approximately 720 mm (6h) from the face of the column
still, however, experienced a relatively high level of
tensile strain (about 60 to 70% of that in the tendons near
the column face). This strain distribution across the
width of PT slab does not vary much with the change in
drift level. Based on the strain data in Figure 13, it can be
concluded that the actual width of high negative bending
curvatures in the PT slab of S1 was considerably greater
than the effective width of c2 + 3h as recommended for
design by Section 13.5.3 of ACI 318–08. Results
obtained from damage observation as shown in Figure
10(a) produced similar conclusions for S1.
4. COMPARISON WITH OTHER TEST
RESULTS AND ACI CODE PREDICTIONS
When a slab-column connection is subjected to lateral
drifts due to earthquakes, punching failures can occur if
punching shear strength of the connection is reached or if
a limiting drift ratio is reached for a given gravity shear
ratio. Two basic failure mechanisms, stress-induced
failure and deformation-induced failure, are recognized
by ACI 318–08 building code. For the former, Section
11.11.7.2 of the code requires that the maximum stress
must be assessed at the critical section since a connection

Advances in Structural Engineering Vol. 15 No. 10 2012 1667

 Seismic Performance of Bonded Post-Tensioned Slab-Column Connections with and without Drop Panels

will fail when the shear stress exceeds the limit (shear
strength). Code formulas for computing the maximum
stress and the strength are available. For the latter, the
maximum deformation capacity is widely known to
depend on the gravity shear ratio, and a code formula for
this relationship is given in Section 21.13.6 of the code.
As many engineers use these code formulas for the design
of slab-column connections, it is important to check
whether or not these formulas agree with the test results.
But since a design code formula normally has a built-in
safety factor, it may not be reasonable to directly compare
the test results with predictions by code formulas. Thus,
the results of test specimens are also compared to results
of similar specimens reported by other investigators.
4.1. Eccentric Shear Stress
In the following, the ACI model for the design of slab-
column connections as shown in Figure 14 is used to
calculate the eccentric shear stress due to a gravity shear
Vu and an unbalanced moment Mu along the critical
section at d/2 from the column face. For the specimen
with drop panel (S2), two critical sections as shown in
Figure 14(b) must be investigated, where d1 is the
average effective depth of the slab within thickened
drop panel region and d2 is the average effective depth
of the slab outside drop panel.
The maximum shear stresses at the critical sections are
expressed by the well-known equations shown below.
c
(b)
(a)
c1 + d
D c A Column
Critical
c2 + d section
Column
Critical
d/2
Drop panlel section 1
C cCD cAB B
c c
(c) c
Vu vu
γ v (AB ) = +
Ac
Vu γ v M u cAB
v u ( AB ) = + (3)
Ac Jc
where Ac = b0d = the area of critical perimeter; cAB is the
distance from the centroidal axis of the critical section to
line AB [Figure 14(a)]; Jc is a property of the critical
perimeter analogous to the polar moment of inertia; and
γ v is the fraction of the unbalanced moment transferred
by eccentricity of shear stress and is given by
γ v = 1 – 1 / (1 + (2 / 3) b1 / b2 ) (4)
Eqn 3 shows that the maximum stress is the sum of
uniformly distributed gravity shear and non-uniform
lateral-force-induced shear. The gravity shear Vu in each
specimen is computed from a linear finite analysis,
which may have some estimated error, but the error is
not so important since the gravity shear stress is low
when compared with the lateral-force-induced shear
(Table 4). The lateral-force-induced (or moment-
induced) shear is estimated from γ v Mu. For each
specimen, the unbalanced moment Mu can be accurately
determined by multiplying the peak lateral force by the
column height (1800 mm) of the specimen.
To apply Eqns 3 and 4 to PT slab-column connections,
however, some value must be assumed for the effective
depth d. It is appropriate to take d as the greater of two
values, the actual effective depth (d) or 0.80h, in
accordance with the concepts of ACI 318–02. For PT
slabs in both specimens, all effective depths were less
than 0.8h and therefore the latter value (0.8h) was used.
Table 3 presents the properties of the critical sections in
PT slabs for both specimens in this study. Based on the
d1/2
peak unbalanced moment (Mu) and the gravity load (Vu)
d2/2
on the test specimens, the maximum shear stresses on the
critical sections of both PT specimens were obtained and
Critical
listed in Columns 6 and 7 of Table 4.
section 2 In Table 4, S2 has two lists of data; the first row lists
γ v M u c AB values at the critical section inside the drop panel, while
the second row lists values at the critical section outside
Jc

( )
D Mu A the drop panel. The critical shear section around the
= 1− 1 / 1+ (2 / 3)
v

γv b1 / b 2
column shows much higher shear stress than the critical
C B Critical shear section outside the edge of the drop panel. This is
section
c CL Column consistent with the observed physical damage of the
Specimen without drop panel Specimen with drop panel specimen as shown in Figure 11(c), where punching
b1 = c1 + d2 = c2 + d Critical section shear failure occurred around the column inside the edge
1: b1 = c1 + d1, b2 = c2 + d1
Critical section 2: b1 = b2 = Wdrop + d2
of the drop panel.
In Figure 15(a), the maximum shear stresses vu of S1
Figure 14. Critical sections at interior column for linear varying and S2 in Table 4 are plotted and compared with shear
shear stress according to ACI building code: (a) Slab without drop stress limits expressed by Eqn (11–34) of ACI 318–08
panel; (b) Slab with drop panel; (c) Stress distribution along critical and by previous works from other investigators. Both
section shear stress vu and fpc are normalized by the square root

1668 Advances in Structural Engineering Vol. 15 No. 10 2012

Unnop Prawatwong, Pennung Warnitchai and Chondro Hidayat Tandian

Table 3. Properties of critical sections

S1 S2 S2
(Inner section) (Outer section)
(1) (2) (3) (4)
Effective depth, d (mm) 96 160 96
Length of critical section, b0 (mm) 1884 2140 6784
Area, Ac (mm2) 180864 342400 651264
Distance to centroid, cAB (mm) 298 330 848
Properties of critical section, Jc (mm4) 9370 × 106 22400 × 106 312000 × 106
Fraction of Mu resisted by shear, γ v 0.467 0.458 0.400

Table 4. Ultimate shear stressesa

Shear stress, vu (MPa)

Vu Mu From From
Specimen (kN) (kN.m) direct shear moment transfer Total vu / fc′
(1) (2) (3) (4) (5) (6) (7)
S1 118 192.6 0.65 2.86 3.51 0.55
S2*b 120 347.0 0.35 2.35 2.70 0.40
S2*c 120 347.0 0.18 0.38 0.56 0.08

Notes: aNo load factors were used in calculations

bShear stress at critical section inside drop panel
cShear stress at critical section outside drop panel

1.0 0.6 S1 (without drop panel)

0.9 S2 (with drop panel)
Best-fit line v c = 0 .46 fc' +0.32 fpc
0.8 0.5 “Stress-induced failure” predicted
by the best-fit equation in figure 16
0.7 (for specimen S2)
0.4
0.6 S1
v v Shear stress limit by ACI
0.5 0.3 (for specimen S2)
fc' S2 fc'
0.4 F (saturation)
v c = 0.29 fc' + 0.3 pc 0.2
Flexural yielding
0.3
 
 Eqn 11–34,ACI 318–08 Flexural yielding
0.2
Reference 1 0.1 Illustration of flexural yielding at both
0.1 sides of the drop panel in specimen
PT bonded (this study)
S2 after −3.5%
0 0.0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0 −1 −2 −3 −4 −5 −6 −7
fpc / fc' Specimen drift (%)
1
Note: ACI-ASCE committee 423 (1974)

Figure 15. (a) Test data of specimen S1 and S2 versus ACI 318-08 equation; (b) Variation of maximum shear stress on critical section of
the test specimens with specimen drift

of the slab compressive strength in SI units (MPa). The for connections transferring shear only, and all tested PT
test data from previous works, represented by the black specimens were unbonded flat plate connections that
dots, were summarized by ACI-ASCE Committee 423 failed in shear. These test data show clearly that though
(1974). They were mostly obtained from tests conducted the normalized shear stress at failure generally increases

Advances in Structural Engineering Vol. 15 No. 10 2012 1669

 Seismic Performance of Bonded Post-Tensioned Slab-Column Connections with and without Drop Panels
with the increase in fpc as predicted by Eqn (11–34) of
ACI, the stress is significantly higher than the code-
specified limit, indicating a built-in safety margin of the
code equation. To determine the true stress limit, an
'
empirical best-fit equation (vu = 0.46 fc + 0.32 f pc )
was derived from these test data as depicted in Figure
15(a). This best-fit equation therefore represents the
most likely value of shear stress at failure in slab-
column connections. When the shear stress at the critical
section exceeds the value predicted by the best-fit
equation, a shear failure may be expected to occur on
slab of the connection. In this paper, such failure is
referred to as a “stress-induced failure”.
Figure 15(a) shows that the shear stresses at failure in
S1 and S2 are both higher than the corresponding code-
specified limits, suggesting that the code equation can
be conservatively used for the design of bonded PT slab-
column connections against stress-induced failure. For
S1, where a sudden punching shear occurred at a low drift
level while the stress was still increasing [Figure 15(b)], the
stress at failure is very close to the value predicted by
the best-fit equation. The result of S1 therefore
conforms well with those of available experimental data
(ACI-ASCE Committee 423 1974) where failures
belong to stress-induced type.
In S2, however, the stress at failure was higher than
the code-specified limit but significantly lower than the
value predicted by the best-fit equation. Moreover, the
shear stress became ‘saturated’ at the drift level of
around 2% [Figure 15(b)], but the punching shear
failure occurred much later at the drift level of 6%. This
shear stress saturation was apparently caused by the
flexural yielding mechanism in the slab as described in
Section 3.5. S2 obviously showed ductile behavior as
opposed to the brittle behavior found in S1.
The punching shear failure at the high drift of 6% in
S2 was believed to happen as a result of slab shear
strength degradation by flexural cracking. As the drift
level increased, crack density and crack widths in the
slab critical region increased as explained in Section 3.4,
significantly reducing slab shear strength. When the
degraded shear strength dropped below the saturated
shear stress, punching shear failure developed. This
failure mechanism is deformation-induced failure.
4.2. Connection Drift Capacity
For seismic design, ACI code provides guidance for the
evaluation of design story drift limit, which is
empirically expressed as a bilinear function of gravity
shear ratio. This limit is based primarily on tests of
isolated, RC slab-column connections subjected to
quasi-static loading. The designer must ensure that the
1670
maximum story drift does not exceed the limit. But if it
is not possible to keep the expected story drift ratio
below the limit, sufficient shear reinforcement should be
provided in the slab-column connection region. This
limit is shown in the following expression:
 3.5 − 5(Vg / φVc ) Vg / φVc ≤ 0.6

Drift limit (%) =  (5)
 0.5 Vg / φVc > 0.6
According to the code, the gravity shear ratio is
defined as Vg / φVc, where Vg is the vertical gravity shear
at the critical perimeter due to gravity load, Vc is the
concrete contribution to the shear capacity, and φ is the
strength-reduction factor. For RC and PT connections,
the concrete contribution to the shear capacity Vc is
generally calculated from the governing equation found
in Section 11.11.2.1 and Section 11.11.2.2 of the code,
respectively. To compare the above drift limit to the test
data, the value of φ is taken as 1.0.
Figure 16 shows a plot of the gravity shear ratio and
drift capacity at punching of both specimens from the
current study, along with other test results of slab-
column specimens without shear reinforcement. Most of
the test results of RC slab-column specimens were
collected and compiled by Pan and Moehle (1989),
while those of unbonded PT slab-column interior
connections were tested by Trongtham and Hawkins
(1977) and Qaisrani (1993) and summarized by Kang
and Wallace (2006). ACI 318-08 design drift limit for
slab-column connections is also plotted in Figure 16 for
reference.
Contrary to the results of unbonded PT slab-column
connections, where the drift capacity is more than twice
7.0
RC slab-column connection
S2
6.0 PT unbonded
(Trongtham and Hawkins 1977)
Drift capacity (%)
5.0 PT unbonded (Qaisrani 1993)
4.0 PT bonded (This study)
ACI 318–08 drift limit
3.0
2.0
S1
1.0
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Gravity shear ratio (Vg /Vc)
Note : Vc = 0.33 f c’ b0 d for RC connections ;
( )
Vc = 0.29 fc’ +0.33 fpc b0 d for P T connections.
Figure 16. Gravity shear ratio versus drift capacity at punching for
RC and PT slab-column connections
Advances in Structural Engineering Vol. 15 No. 10 2012
Unnop Prawatwong, Pennung Warnitchai and Chondro Hidayat Tandian
ACI 318-08 drift limit in every case, the results of the
two bonded PT slab-column connections do not show
that they always possess higher drift capacity. Instead,
the results are more or less consistent with those of RC
specimens, suggesting that ACI 318–08 design drift
limit could also be used for bonded PT slab-column
connections. Nevertheless, since the number of test
specimens is only two, more tests on bonded PT
connections are definitely required before a clear
conclusion can be drawn on this issue.
5. CONCLUSIONS
(1) S1 showed non-ductile behavior under reversed
cyclic loading. A sudden stress-induced punching
shear failure developed at a low lateral drift level
of 2%. The shear stress level at failure matched
well with the average stress limit of unbonded PT
slab-column connections, suggesting that ACI’s
formula for stress limits can also be applied when
designing bonded PT slab-column connections.
(2) Adding a drop panel to a bonded PT slab-
column connection greatly improves its overall
seismic performance. S2 exhibited dramatic
increases in lateral deformation capacity, lateral
strength, and lateral stiffness compared to those
of S1.
(3) Ductile behavior in S2 was clearly demonstrated
by its lateral force-drift relationship. This was
evidently caused by flexural yielding in the slab,
leading to shear stress saturation in the slab’s
critical region at a level below shear strength.
Stress-induced failure was therefore inhibited.
S2 finally failed by a deformation-induced
punching shear inside the drop panel at a very
high drift level of 6%. The results also indicate
that ACI318-08’s design drift limit can be used
when designing bonded PT slab-column
connections.
ACKNOWLEDGMENTS
This experimental work was conducted with funding
provided by the Thailand Research Fund (TRF) and the
Siam City Cement Public Company Limited (SCC). The
prestressed strands, ducts, and anchors used in the tests
were donated by the Concrete Products and Aggregate
Company (CPAC) Limited. Heartfelt gratitude is
conveyed to TRF, SCC, and CPAC for supporting this
research program.
REFERENCES
ACI-ASCE Committee 423 (1974). “Tentative recommendations for
prestressed concrete flat plates”, ACI Journal, Vol. 67, No. 2,
pp. 61–71.
Advances in Structural Engineering Vol. 15 No. 10 2012
ACI Committee 318 (1995). Building Code Requirements for
Structural Concrete (ACI 318–95) and Commentary, American
Concrete Institute, Farmington Hills, MI, USA.
ACI Committee 318 (2002). Building Code Requirements for
Structural Concrete (ACI 318–02) and Commentary, American
Concrete Institute, Farmington Hills, MI, USA.
ACI Committee 318 (2008). Building Code Requirements for
Structural Concrete (ACI 318–08) and Commentary, American
Concrete Institute, Farmington Hills, MI, USA.
ACI Committee 374 (2005). Acceptance Criteria for Moment
Frames Based on Structural Testing and Commentary
(ACI 374.1–05), American Concrete Institute, Farmington Hills,
MI, USA.
ATC (1996). Seismic Evaluation and Retrofit of Concrete
Buildings, Report No. 96–01, Seismic Safety Commission,
California, USA.
Clough, R.W. and Penzien, J. (1993). Dynamics of Structures, 2nd
Edition, McGraw Hill Inc., New York, USA.
Dovich, L.M. and Wight, J.K. (2005). “Effective slab width model
for seismic analysis of flat slab frames”, ACI Structural Journal,
Vol. 102, No. 6, pp. 868–875.
Hawkins, N.M. and Mitchell, D. (1979). “Progressive collapse of
flat-plate structure”, ACI Structural Journal, Vol. 76, No. 7,
pp. 775–809.
Hawkins, N.M. (1981). “Lateral load resistance of unbonded post-
tensioned flat plate construction”, PCI Journal, Vol. 26, No. 1,
pp. 94–116.
Hueste, M.D. and Wight, J.K. (1999). “A nonlinear punching shear
failure model for interior slab-column connections”, Journal of
Structural Engineering, ASCE, Vol. 125, No. 9, pp. 997–1008.
Kang, T.H.K. and Wallace, J.W. (2006). “Punching of reinforced
and post-tensioned concrete slab-column connections”, ACI
Structural Journal, Vol. 103, No. 4, pp. 531–540.
Luo, Y.H. and Durrani, A.J. (1995). “Equivalent beam model for
flat-slab buildings – part I: interior connections”, ACI Structural
Journal, Vol. 92, No. 1, pp. 115–124.
Megally, S. and Ghali, A. (2000). “Punching shear design of
earthquake-resistant slab-column connections”, ACI Structural
Journal, Vol. 97, No. 5, pp. 720–730.
Moehle, J.P. (1996). Seismic Design Considerations for Flat-Plate
Construction, ACI SP-162, Mete A. Sozen Symposium,
Farmington Hills, USA.
Mitchell, D., Adams, J., Devall, R.H., Lo, R.C. and Weichert, D.
(1986). “Lessons from the 1985 Mexican earthquake”, Canadian
Journal of Civil Engineering, Vol. 13, No. 5, pp. 535–557.
Newmark, N.M. and Hall, W.J. (1982). Earthquake Spectra and
Design, Earthquake Engineering Research Institute, Berkeley,
California, USA.
Pan, A.P. and Moehle, J.P. (1988). Reinforced Concrete Flat Plates
under Lateral Loading: An Experimental Study Including Biaxial
Effects, Report UCB/EERC-88/16, Earthquake Engineering
Research Center, University of California - Berkeley, California,
USA.
1671
 Seismic Performance of Bonded Post-Tensioned Slab-Column Connections with and without Drop Panels

Pan, A.P. and Moehle, J.P. (1989). “Lateral displacement ductility of b2
reinforced concrete flat plates”, ACI Structural Journal, Vol. 86,
No. 3, pp. 250–258. c1
Pan, A.P. and Moehle, J.P. (1992). “An experimental study of slab
column connections”, ACI Structural Journal, Vol. 89, No. 6, pp. c2
626–638.
Pongpornsup, S. (2003). Seismic Performance of Post-Tensioned d effective depth of slab (mm)
Interior Flat Slab-Column Connections, Master Thesis, Asian f′c
Institute of Technology, Thailand. fpc
Prawatwong, U. (2012). Seismic Performance of Bonded Post-
Tensioned Interior Slab-Column Connections with and without fpu
Drop Panel, PhD Thesis, Asian Institute of Technology,
Thailand. fpy
Qaisrani, A.N. (1993). Interior Post-Tensioned Flat-Plate fy
Connections Subjected to Vertical and Biaxial Lateral Loading, fu
PhD Thesis, Department of Civil Engineering, University of h slab thickness (mm)
California-Berkeley, California, USA. Jc
Robertson, I.N. and Durrani, A.J. (1992). “Gravity load effect on
seismic behavior of interior slab-column connections”, ACI
Structural Journal, Vol. 89, No. 1, pp. 37–45. Mu
Smith, S.W. and Burns, N.H. (1974). “Post-tensioned flat plate to
column connection behavior”, Journal of the Prestressed
Concrete Institute, Vol. 19, No. 3, pp. 74–91. vc
Trongtham, N. and Hawkins, N.M. (1977). Moment Transfer to
Columns in Unbonded Post-Tensioned Prestressed Concrete vu
Slabs, Report SM77–3, Department of Civil Engineering,
University of Washington-Seattle, Seattle, Washington, USA. V0
Tandian, C.H. (2006). Seismic Performance of Bonded Post-
Tensioned Interior Flat Slab-Column Connections with Drop Vc
Panel, Master Thesis, Asian Institute of Technology, Thailand.
Wey, E.H. and Durrani, A.J. (1992). “Seismic response of interior Vg
slab-column connections with shear capitals”, ACI Structural
Journal, Vol. 89, No. 6, pp. 682–691. Vp
NOTATION Vu
Aps area of prestressed steel (mm2)
As area of nonprestressed longitudinal tension (mm2) γv
b effective width of slab (mm)
b0 perimeter of critical section for shear in slabs (mm) ρ ratio of tension of the nonprestressed reinforcement
b1 (c1+ d) is the dimension of the critical section b0
measured in the direction of lateral load (mm) φ strength-reduction factor for shear
(c2 + d) is the dimension of the critical section b0
measured perpendicular to b1 (mm)
column dimension in the direction of the lateral
load (mm)
column dimension transverse to the lateral load
(mm)
concrete cylinder compressive strength (MPa)
average compressive stress in concrete due to
effective prestress force only (MPa)
specified tensile strength of prestressed steel
(MPa)
yield strength of prestressed steel (MPa)
yield strength of flexural reinforcement (MPa)
tensile strength of flexural reinforcement (MPa)
property of the critical section analogous to the
polar moment of inertia of the critical section
according to ACI 318–08 (mm4)
factored unbalanced moment transferred acting on
the centroid of the critical section of the
connection (kN.m)
nominal shear strength provided by concrete
(MPa)
ultimate shear stress at the critical section due to
applied forces (MPa)
slab punching strength in the absence of moment
transfer (kN)
pure shear capacity of a slab-column connection
with no shear reinforcement (kN)
direct gravity shear force acting on slab critical
section (kN)
vertical component of effective prestress force at
section (kN)
factored shear force at the centroid of the critical
section (kN)
fraction of Mu transferred by eccentricity of shear
stress according to ACI 318–08
As to bd

1672 Advances in Structural Engineering Vol. 15 No. 10 2012