UNIVERSITY OF HAWAII
DEPARTMENT OF CIVIL AND ENVIRONMENTAL ENGINEERING
Performance
Evaluation of Slab-Column
Connections under Cyclic Testing
BACKGROUND INFORMATION:
Name Ian
Robertson
University of Hawaii
Phone (808)
956-6536
Department of Civil and Environmental Engineering
Email ian@wiliki.eng.hawaii.edu
2540 Dole Street, Holmes Hall 383
Honolulu,
HI 96822
Name Ahmad
Durrani
Rice University
Department of Civil and
Environmental Engineering
Email durrani@rice.edu
6100 Main Street
Houston,
TX 77005-1892
SUMMARY
General
This research investigated the effect of a number of variables on the response of slab-column connections subjected to earthquake-type loading.
The bulk of existing research on flat-slab connections had been performed on individual interior or exterior connections subjected to some combination of moment and shear. This research program incorporated the indeterminacy present in actual structures by modeling a single story of a two-bay frame. Each specimen consisted of one interior connection and two exterior connections. The specimens were each subjected to the same lateral displacement history to simulate the effect of earthquake-type loading on the connections.
A total of nine specimens were tested. Of these, seven were multiple connection subassemblies, while the remaining two were individual interior and exterior connections. The variables considered during this project included the slab edge condition at the exterior connection, the use of slab shear stirrups, and the level of gravity load applied to the slab during the test.
General Testing Program
Nine large-scale specimens were tested as part of this research program. Seven of these specimens were multiple-connection subassemblies representing a half scale model of a single story of a two-bay prototype frame, as shown in figure (1). The specimen configuration is based on the assumption that under lateral loading, the points of contraflexure in a multistory frame remain stationary at mid-height of the columns. Each of the seven multiple-connection subassemblies, therefore, consisted of one interior and two exterior slab-column connections, as shown in figures (2) and (3). In addition, two specimens representing the individual interior and exterior connections were also tested to determine the effect of load redistribution on the behavior of connections in indeterminate systems.
Two of the multiple-connection subassemblies were identical in all respects and represented the control specimens. The remaining five subassemblies were used to study the effect of a number of variables. The four variables considered were, 1) the effect of a torsionally stiff edge beam at the exterior connections, 2) the effect of slab shear stirrups, 3) the effect of slab overhang beyond the exterior connections, and 4) the effect of varying levels of slab gravity load.
|
Designation |
Description (MCA = multiple connection assembly) |
|
Control Specimen (MCA) No Shear Reinforcement |
|
|
Control Specimen (MCA) No Shear Reinforcement |
|
|
Stiff Edge Beam at exterior connections (MCA) |
|
|
Closed Hoop Stirrups (MCA) |
|
|
Edge Overhang beyond exterior columns (MCA) |
|
|
Identical to Control Specimen with increased gravity loads (MCA) |
|
|
Identical to Control Specimen with gravity loads between Specimen 1 and 6LL (MCA) |
|
|
Individual Interior Connection (reinforcement similar to Control Specimen) |
|
|
Exterior Connection (reinforcement similar to Control Specimen) |
The test specimens were sized as half scale models of a prototype structure. The prototype structure (figure (1)) was chosen as representative of a typical flat slab residential or office building. It consisted of a 9 inch thick flat plate supported on 20 inch square columns at 20 ft. and 18 ft. centers in orthogonal directions. The story height was set at 10 ft.
For a true half-scale modeling of the chosen prototype, the specimens would have a span of 10 ft. and slab width of 9 ft. Due to the constraints of the testing frame, these dimensions were reduced to 9.5 ft. and 6.5 ft. respectively, as shown in figure (2). Observations from previous research studies have shown that discontinuity in the lateral direction may not have a significant effect on the behavior of the slab-column connections. The columns in the test subassemblies were pinned at the assumed inflection points at mid-height of the story above and below the slab.
The two single-connection specimens represent the interior and exterior connections of the control specimen. The slabs of these specimens were pinned at mid-span as shown in figure (3). The point of contraflexure is assumed to be at midspan, so no lateral or rotational restraint was provided at the slab edge. Steel channels were attached above and below the slab edge to distribute the single vertical reaction along the slab edge.
Prototype Design and Subassembly Design
For the prototype building shown in figure (1), the design gravity loading on each floor
consisted of the slab self weight plus 20 psf superimposed dead load and 50 psf
superimposed live load, which is typical of an office or apartment building.
The lateral earthquake loading was based on the NEHRP design recommendations
(Ref. 2.3) for a Category 2 or moderate earthquake. The frame was analyzed for
this lateral loading to find the moments and shears at the connections at the
second floor level. These moments and shears were then added to those obtained
from a gravity load analysis to get the design slab moments and shears. The
slab was then designed using the ACI 318-85 Building Code. The final slab
thickness of 9 in. was selected such that the maximum ultimate shear stress on
the slab critical perimeter was 
(psi units).
The subassemblies were designed following the
procedure used in the design of the prototype building to obtain slab moments
for the column and middle strips, and unbalanced moments at each connection.
The ultimate shear stresses at all connections were again close to 
as for the
prototype. The flexural reinforcement in the model slab consisted of No. 3
deformed bars positioned in accordance with ACI 318-85 code procedures.
Additional reinforcement was placed in a slab strip 
wide and centered on the column,
to resist portion of the unbalanced moment 
, as prescribed by the code, where 
is the column
dimension perpendicular to the loading direction, 
is the slab thickness and 
is the portion of
unbalanced moment carried in flexure.
Additional bottom reinforcement was added through
the columns in accordance with ACI Committee 352 recommendations to prevent
progressive collapse. All longitudinal slab reinforcement was continuous
through the length of the model. Steel ratios, 
for flexural reinforcement in a slab
width 
, 
outside this width,
and 
perpendicular
to the loading direction, are given in table (1).
At the exterior connections, the slab edge was reinforced for torsion as required by the code. This was achieved by adding No.2 smooth bar closed hoop stirrups at 2.5 in. on center along the slab edge. These stirrups enclosed six transverse No. 3 slab bars to form a "beam" within the slab depth. All longitudinal slab reinforcement was anchored by means of a 90-degree bend into this "beam". The specimen reinforcement layout is shown in figure (4) for the control specimen, 2C. The columns were designed to remain elastic during the test.
Material Properties
Twelve 6 in. diameter concrete cylinders and two standard test beams (for modulus of rupture tests) were made while casting each test specimen. Three of the cylinders were kept in a curing room at 100% humidity and 75 degrees Fahrenheight. These cylinders were tested at 28 days. The other nine cylinders and test beams were kept with the test specimen and cured in the same manner as the test specimen. Three of these cylinders were tested for compressive strength at 28 days. The remaining six cylinders were kept with the specimen. Three were tested for compressive strength at the time of testing the specimen while the others were tested for split cylinder tensile strength. The test beams were also tested for modulus of rupture at the time of testing the specimen.
The steel used as main reinforcement in both slab
and columns was grade 60 Type 2 deformed bars with a specified minimum yield
strength of 60 ksi. The reinforcement was all obtained from the same supplier,
but delivered in two different loads. Coupons of each bar size were taken from
each delivery and tested for tensile strength. The results are summarized in table (2). The observed test yield strengths were
used in the analysis of the test results.
|
Specimen |
fc' |
fr |
d ave |
d flex |
r within c2+3h |
r outside c2+3h |
||
|
(psi) |
(psi) |
(in) |
(in) |
r top |
r bottom |
r top |
r bottom |
|
|
1 |
5275 |
689 |
3.6 |
3.94 |
0.83% |
0.36% |
0.31% |
0.31% |
|
2C |
4460 |
663 |
3.6 |
3.94 |
0.83% |
0.36% |
0.31% |
0.31% |
|
3SE |
5890 |
650 |
3.6 |
3.94 |
0.83% |
0.36% |
0.31% |
0.31% |
|
4S |
5730 |
641 |
3.6 |
3.94 |
0.83% |
0.36% |
0.31% |
0.31% |
|
5SO |
5308 |
544 |
3.6 |
3.94 |
0.83% |
0.36% |
0.31% |
0.31% |
|
6LL |
4440 |
542 |
3.6 |
3.94 |
0.83% |
0.36% |
0.31% |
0.31% |
|
7L |
4050 |
567 |
3.6 |
3.94 |
0.83% |
0.36% |
0.31% |
0.31% |
|
8I |
5320 |
541 |
3.6 |
3.94 |
0.83% |
0.36% |
0.31% |
0.31% |
|
9E |
5320 |
541 |
3.6 |
3.94 |
0.83% |
0.36% |
0.31% |
0.31% |
|
Batch |
Spec No. |
Bar Size |
No Tested |
Area As |
Mean fy |
Mean fu |
Elastic Modulus Es |
|
(in2) |
(Ksi) |
(Ksi) |
(Ksi) |
||||
|
1 |
1- 5 |
#3 Type 2 |
5 |
0.1124 |
72.55 |
116.1 |
27,400 |
|
2 |
6 -9 |
#3 Type 2 |
5 |
0.1168 |
76.13 |
116.9 |
29,460 |
|
1 & 2 |
All |
#2 Smooth |
6 |
0.0475 |
46.67 |
59.1 |
29,500 |
Test Setup
The specimens were all tested in a large steel reaction frame as shown in Figure (6). The test setup was designed so that all column reactions, both shear and axial, could be monitored during the test. The tops of the columns were connected by load cells to a load distribution beam. This beam was supported independently of the specimen and prevented any lateral out-of-plane motion. The lateral loading was applied to the distribution beam by a servo-controlled actuator. No axial load was applied to the columns as column axial load does not have a significant effect on the connection behavior. Connection failure invariably occurs in the slab around the column and not in the joint itself.
The bottom of the center column rested on a load cell measuring the vertical reaction while the column base was restrained against lateral movement. The exterior columns were supported on rollers at the base with lateral restraint provided by a load cell. This arrangement allowed for shear measurement at the base of the exterior columns. The shear at the base of the interior column is then obtained from horizontal equilibrium. Equilibrium of vertical loads and reactions, and overturning moment equilibrium were used to obtain the vertical reactions under the exterior columns.
This test setup maintained the tops and bases of the columns equidistant. This arrangement closely modeled a first floor situation or the situation where one floor experiences greater inelastic action compared to the adjacent floors.
Once installed, the specimen was instrumented to
monitor and obtain the initial loads and stresses for the lateral loading test.
All of the specimens in this research program were subjected to slab loading
simulating the full dead load and 30 percent of the live load of the prototype
structure. This load was applied to the slab surface by hanging forty 450 pound
weights from cables anchored on the top surface of the slab. In the specimens
with increased gravity load, additional lead weights were added to the top of
the slab. The equivalent uniformly distributed slab loads for these specimens
are listed in table (3). Also listed are the gravity
shear stress, 
,
and the shear stress based on the eccentric shear stress model of the ACI code
including the unbalanced moment at exterior connections.
All instrumentation was monitored during application of the slab gravity loading to determine the initial condition prior to application of lateral load. The slab weights were applied individually in a symmetric sequence so as to avoid concentrated stresses in the test specimen. Once all the loads were in place, the instrumentation was monitored to obtain the initial loads and stresses for the lateral loading test.
|
Spec No. |
Superimposed Slab Load (psf) |
Measured Gravity Shear at Specimen Interior Connection |
|||
|
Prototype |
Model |
Vg (kips) |
vg (psi) |
vg |
|
|
2C |
40 |
140 |
11.9 |
60.9 |
0.88 |
|
7L |
120 |
285 |
20.4 |
104.2 |
1.56 |
|
6LL |
220 |
420 |
27.2 |
138.7 |
2.03 |
FEMA 273/274 Comparison
Slab-column connection test data collected to date has been compared with FEMA 273/274 provisions. Preliminary observations indicate that the FEMA connection stiffness predictions over-estimate the actual stiffness. The researchers suggest a modified stiffness be taken into account which is based on assuming the cracked stiffness as the initial stiffness. The viewer may observe the FEMA plots for each of the specimens on their corresponding web-pages.
For details on the properties and test results for each specimen, click on the specimen below.
Stiffness degradation plots were generated along with the hysteretic plots. Please refer to the web pages of the individual specimens for the downloadable EXCEL 97 version files.
|
Designation |
Description (MCA = multiple connection assembly) |
|
Control Specimen (MCA) No Shear Reinforcement |
|
|
Control Specimen (MCA) No Shear Reinforcement |
|
|
Stiff Edge Beam at exterior connections (MCA) |
|
|
Closed Hoop Stirrups (MCA) |
|
|
Edge Overhang beyond exterior columns (MCA) |
|
|
Identical to Control Specimen with increased gravity loads (MCA) |
|
|
Identical to Control Specimen with gravity loads between Specimen 1 and 6LL (MCA) |
|
|
Individual Interior Connection (reinforcement similar to Control Specimen) |
|
|
Exterior Connection (reinforcement similar to Control Specimen) |
Figure #4
Figure #6
