ACI STRUCTURAL JOURNAL TECHNICAL PAPER

Title No. 114-S122

Response of Full-Scale Three-Story Flat-Plate Test

Structure to Cycles of Increasing Lateral Load
by Damon R. Fick, Mete A. Sozen, and Michael E. Kreger

Flat plates were designed and built in the early twentieth century
without sensitivity to earthquake demands. The test reported was
made to study the lateral stiffness, strength, and drift limit in one
direction of a three-story, two-bay, full-scale reinforced concrete
flat-plate building proportioned to resist gravity loading only. The
focus of the experiment was on three issues: 1) the vulnerability of
the slab-column connections; 2) lateral stiffness and drift capacity
of the flat plate; and 3) implications of the observed response for
earthquake resistance of similar structures of which response is
not influenced by nonstructural elements. Measured drifts at the
three levels of loading were estimated to within ±15% using four
types of linear analyses, all based on simple estimates of stiffness
reduction. Results of the investigation indicate similar structures
could survive strong ground motions with peak ground velocities
reaching 600 mm/s (24 in./s).

Keywords: cyclic loading; drift limits; flat plate; full scale; punching shear;
slab-column connections. Fig. 1—Flat-plate test structure after completion of construc-
tion. Reaction wall can be seen in background.
INTRODUCTION
The flat plate, a direct descendant of C. A. P. Turner’s the measured experimental data. The survival of similarly
reinforced concrete flat slab,1 was an immediate success proportioned structures during strong ground motion is also
because of its economy and its flexibility for accommo- discussed.
dating architectural requirements. It was built in all regions
including those with seismic risk. Many of these structures EXPERIMENTAL INVESTIGATION
were designed in the early twentieth century without sensi- Materials
tivity to earthquake demands. The test reported in this paper Compressive strengths of the normalweight-aggregate
was made to study the lateral stiffness, strength, and drift concrete, measured by tests of 150 x 300 mm (6 x 12 in.)
limit of a three-story, two-bay, full-scale reinforced concrete cylinders, in various components of the structure are summa-
flat-plate structure proportioned to resist gravity loading rized in Table 1. Excluding the concrete in the footings,
only (Fig. 1). The objectives of the research were to inves- measured mean strength varied from 22.7 to 28.0 MPa (3290 to
tigate: 1) the vulnerability of the slab-column connections; 4060 psi). The minimum and maximum strengths indicated
2) reductions in lateral stiffness and drift capacity of the flat by a single cylinder test ranged from 21.0 to 30.5 MPa (3050 to
plate; and 3) implications of the observed response for earth- 4420 psi).
quake resistance of similar structures of which response is Arrangements of top and bottom longitudinal reinforce-
not influenced by nonstructural elements. The presented ment are shown in Fig. 2. Figure 3 shows nominal cross-
research and conclusions are limited to flat-plate structures sectional dimensions. Reinforcement ratios in the column
with overhanging slab-edges and without other lateral force- strip (3.05 m [10 ft]) were 0.5% for the negative-moment
resisting elements. The slab-column connections were rein- and 0.25% for the positive-moment reinforcement. In the
forced with moderate levels of flexural reinforcement (0.5%) middle strip, the reinforcement ratio was 0.25% at both critical
and did not have shear reinforcement. sections. The reinforcement in the transverse column and
middle strips was the same. The cross-sectional arrangement
RESEARCH SIGNIFICANCE of column reinforcement is shown in Fig. 4. Column ties
The drift capacity and lateral stiffness of slab-column were spaced at 180 mm (7 in.).
connections is based primarily on reduced-scale and isolated
slab-column specimens. Test results of a full-scale, flat-plate ACI Structural Journal, V. 114, No. 6, November-December 2017.
structure are reported in this paper to evaluate the behavior MS No. S-2016-215, doi: 10.14359/51689502, received June 2, 2016, and reviewed
under Institute publication policies. Copyright © 2017, American Concrete Institute.
of slab-column connections that are part of a four-bay, three- All rights reserved, including the making of copies unless permission is obtained from
story structure. A comparison of analytical procedures to the copyright proprietors. Pertinent discussion including author’s closure, if any, will
be published ten months from this journal’s date if the discussion is received within
estimate the lateral drift of flat plate structures is made with four months of the paper’s print publication.

ACI Structural Journal/November-December 2017 1507

Table 1—Compressive strength of 150 x 300 mm (6 x 12 in.) concrete cylinders, MPa (psi)
Location Mean Maximum Minimum Number of trucks Number of cylinders
Roof slab 27.9 (4040) 29.4 (4270) 25.7 (3730) 4 12
Story 3 columns 27.5 (3990) 29.8 (4320) 24.0 (3480) 2 6
Level 3 slab 28.0 (4060) 30.5 (4420) 25.8 (3740) 4 12
Story 2 columns 22.7 (3290) 23.2 (3360) 22.1 (3200) 2 4
Level 2 slab 25.2 (3650) 28.6 (4150) 21.0 (3050) 4 12
Story 1 columns 26.3 (3810) 28.2 (4090) 26.2 (3800) 1 3
Footings 20.5 (2970) 22.1 (3210) 18.4 (2670) 2 6

Fig. 4—Column reinforcement. (Note: 1 mm = 0.0394 in.)

Fig. 2—Arrangement of longitudinal reinforcement. (Note:

1 m = 3.28 ft.)

Fig. 5—Dimensions of flat-plate test structure. (Note: 1 mm

= 0.0394 in.; 1 m + 3.28 ft.)
test structure had a height of 9.1 m (30 ft) above the footings,
Fig. 3—Typical slab cross section. (Note: 1 mm = 0.0394 in.) each story measuring 3.0 m (10 ft). The footings, 1.4 m
(4 ft 6 in.) square in plan with a thickness of 760 mm (2 ft 6 in.),
Deformed No. 13 (No. 4) reinforcing bars cut from were fixed on the test floor of the laboratory.
Grade 420 (Grade 60) steel had mean yield stresses of Before application of lateral loads, barrels containing
470 MPa (68 ksi) for the top bars and 460 MPa (67 ksi) water were placed on the slab to result in a uniform load of
for the bottom bars, determined by three tests in each case. 64.9 kg/m2 (13.3 lb/ft2) resulting in a total weight of 2260 kN
Measured mean tensile strength was 690 MPa (100 ksi) for (509 kip) for the test structure. The structure was loaded in
the top and 680 MPa (99 ksi) for the bottom bars. Deformed the north-south direction with lateral loads varying linearly
No. 22 (No. 7) column reinforcement had a yield stress of with height (Fig. 6). Rams reacted against the strong wall of
460 MPa (66 ksi) and tensile strength of 670 MPa (97 ksi) the laboratory (Fig. 7). The load was transferred to the slab
based on results of two tests. at each level on east and west edges at two points located at
the middle of the 6.1 m (20 ft) spans.
Specimen
A perspective view as well as relevant dimensions of the EXPERIMENTAL RESULTS AND DISCUSSION
test structure are shown in Fig. 5. Slabs at all levels had plan Response of test structure to cycles of load reversals
dimensions of 9.1 x 15.2 m (30 x 50 ft) and thickness of The relationships between applied load (expressed in terms
180 mm (7 in.). Support was provided by six 460 mm (18 in.) of total base shear) and lateral deflection (drift) measured at
square columns spaced at 6.1 m (20 ft) in both directions. The each level are shown in Fig. 8 through 10.

1508 ACI Structural Journal/November-December 2017

Fig. 6—Distribution of lateral load over height of structure. Top actuators were controlled by displacement; lower-level actu-
ators were force-controlled.

Fig. 7—Plan view of test structure and reaction wall.

Fig. 8—Measured relationship between base shear and roof Fig. 9—Measured relationship between base shear and drift
drift. (Note: 1 kN = 0.225 kip; 1 mm = 0.0394 in.) at Level 3. (Note: 1 kN = 0.225 kip; 1 mm = 0.0394 in.)

ACI Structural Journal/November-December 2017 1509

 During the first cycle to a mean drift ratio (ratio of roof Column cracks
drift to height of structure) of 0.2%, observed response was One of the most useful observations made during the
essentially linear. As drift demand increased in successive testing of the flat plate structure was that of the distribu-
load cycles, nonlinear response increased with increasing tion of flexural cracks in the columns (Fig. 11). The flexural
drift. Load increase with drift approached a very low rate crack distributions revealed the strong difference from the
that could be represented as “plastic” after a mean drift ratio expected distribution of column moments in a reinforced
(MDR) of 2% was exceeded. concrete frame with girders typically having a depth in
The mean stiffness, defined by a line joining the upper and inches equal to the length of the span in feet (that is, depth
lower peaks of the loading cycle, was reduced at maximum to span = 1/12) and those observed in the test structure. The
drift to approximately one-eighth the initial stiffness. It may slab depth for the test structure was L/34, where L is the
be inferred that the mode-1 period of the structure subjected center-to-center span of the columns.
to strong ground motion would increase to approximately In the first cycle of loading to a roof-drift ratio of 0.2%, no
three times its initial mode-1 period if the dynamic response flexural cracks were observed in the columns. In the second
demand reaches the same maximum drift attained in this test. cycle of loading to a roof-drift ratio of 0.4%, flexural cracks
were observed in both first- and second-story columns.
Cracks were recorded over the lower 910 mm (36 in.) of
the first-story columns and over the lower 300 mm (12 in.)
of the second-story columns. No cracks were observed in
the third- story columns. In the third cycle of loading to a
roof-drift ratio of 1.5%, flexural cracks were observed over
the lower 1.9 m (74 in.) of the first-story columns. No cracks
were observed in first-story columns near the slab-column
joint at Level 2. The distribution of cracks over the height
of the columns indicated that the point of contraflexure was
close to the upper joint. Columns were responding almost
as cantilevers. The rotation restraint at the upper joint of
each column was flexible compared with that at the base
connection.
In the second story, cracks in the columns were distributed
within 810 mm (32 in.) of the lower and 640 mm (25 in.) of
Fig. 10—Measured relationship between base shear and the upper joints suggesting that the point of contraflexure
drift at Level 2. (Note: 1 kN = 0.225 kip; 1 mm = 0.0394 in.) was above mid-height of the second-story columns and that
the moments at both ends of columns exceeded the cracking

Fig. 11—Ranges of flexural cracking observed on columns in Load Cycles 2, 3, and 4. (Note: 1 mm = 0.0394 in.; 1 m = 3.28 ft.)

1510 ACI Structural Journal/November-December 2017

 Fig. 14—Dimensions defining boundaries of shear failure at
top of slab. (Note: 1 mm = 0.0394 in.)
and two-thirds of the story height with the column moment
Fig. 12—North face of interior column 2B after failure. at the lower joint being half that at the upper joint.
The aforementioned inferences provide a platform for
preliminary estimates of column moments in a flat plate
with flexible horizontal elements. The proportions of the test
specimen suggest the stiffness to span ratio should meet the
following criterion

I slab 1  I column 
<  (1)
2 ⋅ Spanslab 5  Heightstory 
where Islab is gross moment of inertia of the full slab width,
in.4; Icolumn is gross moment of inertia of the column, in.4;
spanslab is center-to-center span of the slab, ft; and Heightstory
is story height, ft.
Development of cracks on slab surfaces and changes in
their widths are recorded in 89 figures available in Fick2 and
discussed in Reference 3.
Fig. 13—West face of interior column 2B after failure.
Failure
moment. The distribution of flexural cracks in the third-story As the test structure was being loaded south, at MDR =
columns was nearly the reciprocal of the crack distributions 2.9% or 267 mm (10.5 in.), there was a sudden drop in base
in the first-story columns. Flexural cracks were observed in shear from 685 to 605 kN (154 to 136 kip). Loading was
the upper 1.1 m (45 in.) of the column height. No cracks stopped for approximately 4 minutes and then resumed to
were observed near the joint at Level 3. reach a roof-drift ratio of 3% at a base shear of 654 kN
There was little change in the flexural crack distributions (147 kip).
during Load Cycle 4. In the first-story column, cracking Examination of the joints at Level 3 revealed that the slab
covered the lower 2.0 m (80 in.) of the column height and in had fractured around the interior column of Frame B
the second story, the lower set of cracks extended from 810 (Fig. 12 and 13). The measured story drift ratio (SDR) for
to 1100 mm (32 to 45 in.). No change in crack distribution this level was 3.3% at the sudden drop in base shear. Dimen-
was observed in columns of the third story. sions related to the boundary of the fracture observed on the
The following is inferred from the observed distributions top surface of the slab are recorded in Fig. 14. The vertical
of flexural cracks in the columns: separations of the slab, measured after loading was discon-
1. In preliminary modeling of the response of a flat plate tinued, are shown in Fig. 15. Failure was clearly due to
with similar dimensions to lateral loading, the first-story rotation or moment demand at the slab-column connection.
columns may be considered to respond as cantilevers with It was expected to occur at Level 3 and on Line 2 because,
their base moments obtained as the product of the shear in contrary to what would take place in a frame of typical
the column and the column height. proportions, the flexibility of the slab tends to move the
2. Maximum moments in the intermediate-story columns larger moment from the first joint above base to the second
may be estimated as the product of the shear in the column one. The failure is more likely to occur at the interior joint
and half the story height. The story stiffness will be affected rather than at an exterior joint because the slab, discontinued
by the rotation of the slab-column joints. on one side of the column at exterior connections, does not
3. The moments at the top joints of third-story columns develop as high a moment as the interior joint does.
may be estimated as the product of the shear in the column

ACI Structural Journal/November-December 2017 1511

 selected yield mechanism is not surprising because of the
expected strain hardening in the reinforcement.
The test demonstrated that the flexural capacity of a full-
scale flat-plate structure, given a known distribution of
lateral forces, can be determined conveniently to result in a
safe estimate of base shear strength related to a yield mech-
anism involving plastic hinges at column bases and slab-
column joints with the slab contribution limited to that from
the column strip.

Root cause of joint failure

The load and deformation capacities of the test structure
were limited by a joint failure. Even though the location
Fig. 15—Vertical section at Level 3 column 2B after failure of the failure could have been identified by estimating the
in shear. (Note: 1 mm = 0.0394 in.) magnitudes of the moment transfer and the story drifts, to
determine the moment that triggered the failure is not within
the realm of mechanics because of the random nature of
crack trajectory around the columns. Each slab-column joint
is threatened by three demands: 1) flexure; 2) shear related
to gravity load; and 3) shear resulting from gravity loads
combined with moment transfer.
The slab reinforcement ratio in the column strip was 0.5%.
Flexure was not a threat by itself. The yield moment would
have been developed without damage to the concrete. The
vertical reaction on the interior columns was 130 kN (30 kip)
at each level. The total cross-sectional area determined at
a section d/2 distant from the face of the column (d is the
effective depth) was 0.372 m2 (576 in.2). The corresponding
mean unit shear stress is barely over 340 kPa (50 psi). Based
on what is known about the shear resistance of slab-column
Fig. 16—Controlling yield mechanism used in calculating joints, shear caused by the vertical reaction was not a threat
limiting base shear. either. It follows that the failure must have been driven by
shear stresses related to the gravity load combined with
After completion of loading, examination of the slab at
those resulting from the moment transfer.
Level 3 revealed that a similar vertical-separation fracture
A brief look at Fig. 14 would suggest that to determine
had occurred during loading to the north in the frame along
combined shear stresses based on a section at d/2 from the
Line A but it had not had a perceptible effect on the load-
column face is not quite correct. A reasonable option is to
deflection relationship of the structure.3
look for a solution based on observation.
Morrison et al.4 reported results of small-scale tests of
Base shear strength limit based on flexural
slab-column connections. Morrison et al. also developed a
yielding
computational model comprising bars with linear response
Devoid of nonstructural elements and with material prop-
properties running in the two orthogonal directions primarily
erties and member dimensions known reasonably well,
to sense the torsional shear stresses. Success was obtained as
determining the limiting base shear strength for the known
long as the drift involved was low. Rotation demand at the
load distribution was a simple task. The only arbitrary deci-
joint, reflected in the drift measurement, appeared to be a
sion involved was the effective width of the slab. Rather
dominant factor influencing the strength of the joint.
than assuming that the entire slab width contributed to the
Pan and Moehle5,6 reported their tests of slab-column
flexural resistance, it was assumed to limit the source of the
connections noting the reduction in the drift capacity with
slab flexural resistance to the column strip width of 3.05 m
increase in the shear stress related to gravity loading. Figure
(10 ft). The yield mechanism is shown in Fig. 16. The
17 is a plot of the gravity shear ratio, γ, versus story drift
average yield moment of the columns was 315 kN-m
ratio from 19 investigations4-22 shown in Table 2. A reason-
(232 kip-ft) using a yield stress fy = 460 MPa (66 ksi), and
able lower bound to the observed data is provided by the
for the top and bottom reinforcement in the slab, it was
expression
153 kN-m (113 kip-ft) using a yield stress of 470 MPa
(68 ksi) and 85 kN-m (63 kip-ft) with a yield stress of
SDR(%) = 4(1 – 2γ) (2)
460 MPa (67 ksi), respectively. The limiting base shear was
determined to be 658 kN (148 kip). The measured base shear
where γ is ratio of gravity shear stress to estimated shear
strength was higher by 9%. It was 716 kN (161 kip). That the
capacity, vg/vc; vg is ratio of column reaction to the effective
structure would resist more than the load determined by the
section at distance of d/2 from column face; and vc = 4√fc′.

1512 ACI Structural Journal/November-December 2017

Fig. 17—Plot of gravity shear ratio versus story drift ratio from 19 investigations.4-22
Equation (2) was derived from the collection of data C, and D) used are two-dimensional (2-D) and linear with
shown in Fig. 17 as a safe lower-bound estimate for gravity plausible estimates of stiffness reduction. These models are
shear ratios not exceeding 0.5 and made simple enough to be assembled assuming that the contribution of slab stiffness
suitable for design.3 The result from the test of the flat-plate to lateral resistance is limited to a slab section measuring
structure also suggests the small-scale specimens provide a 3.05 m x 180 mm (10 ft x 7 in.), which is the dimension of
reliable basis for determining the reduction in drift capacity the column strip. The fourth model (A) is a linear three-di-
of a full-scale structure. mensional (3-D) model. It also includes plausible estimates
of stiffness reduction. The lateral load distributions for the
COMPARISON OF PREDICTIONS AND four models over the height of the structure are based on
EXPERIMENTAL RESULTS those applied during the test runs. The applied loads are 3P
The four loading cycles selected for the test of the flat at the roof level, 2P at Level 3, and P at Level 2, adding to a
plate targeted the four ranges of response of a flat plate base-shear force of 6P.
subjected to a strong ground motion. The first cycle with a Model A is a 3-D finite-element model23 constructed using
maximum mean drift ratio (MDR or ratio of roof drift to 230 x 230 mm (9 x 9 in.) shell elements for slabs and frame
height of structure) of approximately 0.2% was selected elements for columns. The eight shell elements that bound
because it was estimated that within that drift range the the 460 x 460 mm (18 x 18 in.) column perimeter were
initial stiffness of the structure would not change. The connected to the column joint by modeling 25 x 25 mm
maximum MDR of the second cycle (approximately 0.4%) (1 x 1 in.) frame objects within the plane of these shells.
was selected to investigate whether the initial stiffness would Body constraints were added to enforce rigid-body behavior
be sustained at that level of drift. The third cycle to MDR = 1.5% of the shell members through the column section.
was selected to obtain a basis for estimating the mode-1 Model B is the standard linear frame model with a 3.05 m
period during a range of response that would be considered (10 ft) slab width.
tolerable for a flat-plate structure. In the fourth cycle, the Model C is, in effect, based on an ancient time-honored
intent was to push the structure to its limit under static structural-engineering procedure that treats each story as a
loading. As described earlier, the flat plate suffered a joint linear spring. This method was established before the use
failure at MDR of 2.9% but was observed to maintain its of computers and is a simple calculation that can be used to
overall integrity to MDR = 3%. check the results of commercial software. The story drift is
In this section, mean stiffness of the structure is esti- calculated by dividing the story shear by the story stiffness
mated for each of the four load cycles and compared with Vstory/Kstory. The springs are determined in accordance with
the experimental results. The analyses are made using four Eq. (3).
different models that are within the range of procedures in
the realm of design. The idealized models used for the four  
analyses are shown in Fig. 18. Three of the four models (B, 24 ⋅ E  1 
K story = 2   (3)
h  2 1 1 
 ∑ K + +
c ∑ K sa ∑ K sb 

ACI Structural Journal/November-December 2017 1513

Table 2—Experimental data for slab-column connections
Peak drift Peak drift
Investigator Label γ ratio, % Failure Investigator Label γ ratio, % Failure
DNY 1 0.20 3.0 F I.I 0.08 5.0 F
DNY 2 0.30 2.0 P Luo and Durrani 199516 INT 1 0.43 N/A P
Durrani et al. 19957
DNY 3 0.24 2.0 F INT 2 0.50 N/A P
DNY 4 0.28 2.6 F-P MG-2A 0.58 1.2 P
Elgabry and Ghali 1987 8
1 0.46 N/A P MG-7 0.29 3.1 F-P
Megally and Ghali 200017
1 0.04 4.8 F MG-8 0.42 2.3 F-P
2 0.04 4.0 F MG-9 0.36 2.2 F-P
Farhey et al. 19939
3 0.26 3.6 P S1 0.03 4.7 F
4 0.30 2.4 P S2 0.03 2.8 F
SM 0.5 0.31 6.0 F Morrison et al. 19834
S3 0.03 4.2 F
Ghali et al. 197610 SM 1.0 0.33 2.7 F-P S4 0.07 4.5 F
SM 1.5 0.30 2.7 F-P S5 0.15 4.8 F
A12 0.29 N/A P AP 1 0.37 1.6 F-P
A13L 0.29 N/A P AP 2 0.36 1.5 F-P
Pan and Moehle 19895
B16 0.29 N/A P AP 3 0.18 3.7 F-P
Hanson and Hanson 196811
B7 0.04 3.8 F-P AP 4 0.19 3.5 F-P
C17 0.24 N/A F-P 1 0.35 1.5 P
C8 0.05 5.8 F 2 0.35 1.5 P
Pan and Moehle 19926
S1 0.33 3.8 P 3 0.22 3.1 F-P
S2 0.45 2.0 P 4 0.22 3.2 P
Hawkins 197412
S3 0.45 2.0 P 1 0.21 2.8 F
Robertson 199018
S4 0.40 2.6 P 2C 0.22 3.5 F-P
Hwang and Moehle 200013 4 Int. Joints 0.24 4.0 N/A 3SE 0.19 3.5 F
1 0.25 3.7 P 5SO 0.21 3.5 F
Islam and Park 1976 14
2 0.23 3.3 P 6LL 0.54 0.9 P
Robertson et al. 200219
3C 0.23 4.0 F-P 7L 0.40 1.5 P
ND1C 0.23 3.0 to 5.0 F-P 8I 0.18 3.5 F-P
ND4LL 0.28 3.0 F-P 1C 0.17 3.5 P

Robertson and Johnson ND5XL 0.47 1.5 P S6 0.86 1.1 P

Symonds et al. 197620
200615 ND6HR 0.29 3.0 P S7 0.81 1.0 P
NC7LR 0.26 3.0 F-P Wey and Durrani 1992 21
SC 0 0.25 3.5 P
ND8BU 0.26 3.0 F-P Zee and Moehle 198422 INT 0.21 3.3 F-P

Notes: F denotes flexural failure; P denotes punching shear failure; N/A is not available.

where Kstory is stiffness of linear spring representing the stiff- Model D was inspired by the observations made in the
ness of the story; ΣKc is sum of Ic/h, where Ic is the gross tests of the flat plate reported in this paper. It is simple and
moment of inertia of the column and h is the story height straightforward and reduces the task of estimating the lateral
(=3Kc for the geometry of the test specimen); ΣKsa and ΣKsb drift to one of elementary arithmetic. Similar to Model C, the
are sums of Is/L, where Is is the gross moment of inertia of second and third stories of the frame are modeled as linear
the slab considered to contribute to lateral stiffness, L is the springs. The first story is modeled as a cantilever with a stiff-
center-to-center span length, subscripts a and b identify slab ness of 3EI/h3 multiplied by the number of columns. The
stiffnesses above and below the story (for the first story, the effective stiffnesses of the second story determined using
third term in the denominator within the parentheses is zero; Eq. (3). The effective stiffness of the third story is further
for the third story, the coefficient 24 is divided by 2; ΣKsa = reduced with respect to that of Model C by dividing the coef-
ΣKsb = 2Ks for the geometry of the test specimen); and E is ficient 24 by 4. As in Model C, gross moments of inertia
Young’s modulus for concrete. are used and the stiffness reduction is applied to the story

1514 ACI Structural Journal/November-December 2017

Fig. 18—Idealized linear elastic analytical models.
stiffness instead of individually to the columns and slabs. Table 3—Calculated results
The story drift is calculated by dividing the story shear by
Stiffness reduction
the story stiffness. Base
(1/16)
The four aforementioned models were used to determine shear, Calculated roof
Model kN Column Slab displacement, mm
the drifts at the three levels of the test structure at base shears
of 231, 356, 609, and 716 kN (52, 80, 137, and 161 kip). The 231 16 16 11.2
results are listed in Table 3 and plotted with the measured Model A 356 8 6 37.1
load displacement response of the test structure in Fig. 19. (3-D finite
element) 609 5 2 141
Slab and column stiffness reductions were made not to
716 3 1 297
produce an exact model (column and slab stiffnesses would
be expected to be different at different levels) but to provide 231 16 16 12.7
examples of reasonable decisions that could be made by a Model B (2-D 356 8 8 38.6
designer having an interest in estimating the responses of linear frame) 609 4 4 132
the structure being proportioned at different magnitudes of
716 3 2 267
ground-motion demand.
Assuming that ±20% is acceptable in estimating mean stiff- 231 16 16 20.8
ness (it would result in a period estimate error range below Model C (2-D 356 14 13 39.1
10%), it is observed that for Cycle 1 (a maximum MDR < linear spring) 609 8 6 141
0.2%), Models A, B, and C produce tolerable results using
716 6 4 246
gross-section properties of the structure. A stiffness reduc-
tion of seven-eighths was used for both columns and slabs 231 14 17.8
in the simplified Model D. To repeat the same at MDR = Model D 356 10 38.6
0.4% requires a stiffness reduction by one-half for Model B (Simplified) 609 5 132
and five-eighths for Model D. The stiffness reductions were
716 3 259
between three-eighths and seven-eighths for Models A and C.
At the drift ratio of 1.5%—which is a reasonable limit for Notes: 1 mm = 0.0394 in.; 1 kN = 0.225 kip.
a flat plate not having details to maintain its integrity under

ACI Structural Journal/November-December 2017 1515

Fig. 19—Base shear versus measured and calculated roof displacements. (Note: 1 kN = 0.225 kip; 1 mm = 0.0394 in.)
cyclic loading (in an earthquake)—Model A requires the As suggested by Westergaard,24 the spectral drift of a
stiffness be reduced between one-eighth and five-sixteenths structure of low or medium height is related to its first-mode
and be reduced between one-quarter and one-half for period, T, and maximum ground velocity, Vg. Equation (4)
Models B, C, and D. represents a reasonable upper bound to ground motions on
At the limiting MDR of 3%, all models provide an accept- stiff ground for structures having a base shear strength coef-
able result using stiffness reductions between one-sixteenth ficient not below 0.1 and with a reasonably uniform stiffness
and three-eighths for the slab and column members (Table 3). over its height. The coefficient 5/4 is the participation factor
Considering Model D demands a fraction of the labor that projects the spectral displacement Sd to the roof of the
required for the other three models and accepting the fact structure.
that the exact analysis of an approximate model is seldom From Sozen25
good enough to serve as the approximate analysis of an
exact model, Model D is the method recommended for Vg T
determining the lateral stiffness of flat plates with propor- Sd = (5)
2
tions similar to those of the test structure. It does require
judgment that comes from experience but the results are not The initial period of the test structure was measured using
sensitive to the chosen reduction factors within the range of both forced and free vibration tests26 and determined to be
those listed in Table 3. 0.5 seconds. Models A and B described earlier resulted in
calculated periods of 0.48 and 0.47 seconds, respectively,
Projection to earthquake environment using gross section properties. The test structure maintained
The dynamic response of a 3-D reinforced concrete its integrity to a MDR of approximately 3%. Given that,
structure with changing stiffnesses to a multi-dimensional the corresponding maximum ground velocity it is likely to
ground-motion demand of an earthquake that has not survive is 610 mm/s (24 in./s).
occurred is not a problem that lends itself to analysis with It is generally assumed that a ground velocity of 250 mm/s
expectation of accurate results. Nevertheless, the following (10 in./s) represents moderate earthquake demand and
generalizations may be acceptable based on the assumption 500 mm/s (20 in./s) refers to a strong ground motion. It may
that the MDR response of a flat-plate structure is the critical be concluded that the test structure is likely to survive a
issue for its survival in an earthquake. strong ground motion.
A structure of reasonably uniform stiffness and mass
distribution over its height would be expected to have a SUMMARY AND CONCLUSIONS
mean drift ratio expressed by Eq. (4) A full-scale, three-story reinforced concrete structure
(Fig. 1) was built and tested to failure under cycles of increasing
5S d lateral load (Fig. 8 through Fig. 10). The experimental study
MDR = (4) was focused on: 1) vulnerability of the slab-column connec-
4h
tions; 2) changes in stiffness related to increasing drift; and
where Sd is spectral displacement response; and h is height 3) limiting drift of a flat plate not proportioned and detailed
from base to roof. for earthquake resistance.

1516 ACI Structural Journal/November-December 2017

 1. At a maximum mean drift ratio of 1.5%, the response of Inelastic Range of Response,” PhD dissertation, Purdue University, West
Lafayette, IN, 2008, 189 pp. [Link]
the structure was without fault. 3. Fick, D. R., “Cyclic Lateral-Load Test to Failure of a Full-Scale
2. Displaced to a mean drift ratio of 3% north, there was Three-Story Flat-Plate Reinforced Concrete Structure,” Proceedings of the
a failure, in combined torsion and shear, of one of the inte- 9th U.S. National, 10th Canadian Conference on Earthquake Engineering,
Toronto, ON, Canada, July 25-29, 2010.
rior column connections at Level 3. Initiation of the failure 4. Morrison, D. G.; Hirasawa, I.; and Sozen, M. A., “Lateral-Load Tests
was not detectable in the observed relationship between base of R/C Slab-Column Connections,” Journal of Structural Engineering,
shear and drift. Maximum base shear reached in the north ASCE, V. 109, No. 11, 1983, pp. 2698-2714.
5. Pan, A., and Moehle, J. P., “Lateral Displacement Ductility of
loading was 716 kN (161 kip) at a mean drift ratio of 3%. Reinforced Concrete Flat Plates,” ACI Structural Journal, V. 86, No. 3,
3. As the structure was displaced to a mean drift ratio of May-June 1989, pp. 250-258.
2.9% south, there was a sudden reduction of 18% in base 6. Pan, A. D., and Moehle, J. P., “Experimental Study of Slab-Column
Connections,” ACI Structural Journal, V. 89, No. 6, Nov.-Dec. 1992,
shear. Loading was continued to reach a base shear of pp. 626-638.
685 kN (154 kip) at a mean drift ratio of 3%. 7. Durrani, A. J.; Du, Y.; and Luo, Y. H., “Seismic Resistance of Nonduc-
4. The measured drifts at the three levels in all four cycles tile Slab-Column Connections in Existing Flat-Slab Buildings,” ACI Struc-
tural Journal, V. 92, No. 4, July-Aug. 1995, pp. 479-487.
could be estimated to within ±15% using linear analyses 8. Elgabry, A. A., and Ghali, A., “Tests on Concrete Slab-Column
of types ranging from one involving a simple application Connections with Stud-Shear Reinforcement Subjected to Shear-
of elementary mechanics to one involving a detailed 3-D Moment Transfer,” ACI Structural Journal, V. 84, No. 5, Sept.-Oct. 1987,
pp. 433-442.
modeling of the structure, all based on reasonable estimates 9. Farhey, D. N.; Adin, M. A.; and Yankelevsky, D. Z., “RC Flat Slab-
of stiffness reduction with displacement. Column Subassemblages under Lateral Loading,” Journal of Structural
5. The maximum MDR measured in the test structure used Engineering, ASCE, V. 119, No. 6, 1993, pp. 1903-1916.
10. Ghali, A.; Elmasri, M. Z.; and Dilger, W., “Punching of Flat Plates
with established relationships between drift, spectral displace- under Static and Dynamic Horizontal Forces,” ACI Journal Proceedings, V.
ment, and peak ground velocity suggest flat plates of similar 73, No. 10, Oct. 1976, pp. 566-572.
proportions could survive a broadband ground motion having 11. Hanson, J. M., and Hanson, N. W., “Shear and Moment Transfer
between Concrete Slabs and Columns,” Journal of the PCA Research and
a peak ground velocity of 600 mm/s (24 in./s). Development Laboratories, V. 10, No. 1, 1968, pp. 2-16.
6. The drifts corresponding to the observed joint failures 12. Hawkins, N. M., “Shear Strength of Slabs with Moments Transferred
were consistent with the trend observed in tests of small- to Columns,” Shear in Reinforced Concrete, SP-42, American Concrete
Institute, Farmington Hills, MI, 1974, pp. 785-816.
scale tests of slab-column connections. 13. Hwang, S.-J., and Moehle, J. P., “Vertical and Lateral Load Tests
of Nine-Panel Flat-Plate Frame,” ACI Structural Journal, V. 97, No. 1,
AUTHOR BIOS Jan.-Feb. 2000, pp. 193-203.
ACI member Damon R. Fick is an Assistant Professor in the Civil Engi- 14. Islam, S., and Park, R., “Tests on Slab-Column Connections with
neering Department at Montana State University, Bozeman, MT. He is Shear and Unbalanced Flexure,” Journal of the Structural Division, ASCE,
Secretary of ACI Committee 352, Joints and Connections in Monolithic V. 102, No. 3, 1976, pp. 549-568.
Concrete Structures, and a member of ACI Subcommittee 318-B, Anchorage 15. Robertson, I., and Johnson, G., “Cyclic Lateral Loading of Nonduc-
and Reinforcement (Structural Concrete Building Code). His research tile Slab-Column Connections,” ACI Structural Journal, V. 103, No. 3,
interests include the behavior and design of reinforced concrete structures. May-June 2006, pp. 356-364.
16. Luo, Y. H., and Durrani, A. J., “Equivalent Beam Model for Flat-Slab
ACI Honorary Member Mete A. Sozen is the Kettlehut Distinguished Buildings—Part I: Interior Connections,” ACI Structural Journal, V. 92,
Professor of Structural Engineering at Purdue University, West Lafayette, IN. No. 1, Jan.-Feb. 1995, pp. 115-124.
17. Megally, S., and Ghali, A., “Punching of Concrete Slabs Due to
Michael E. Kreger, FACI, is the Garry Neil Drummond Endowed Chair Column Moment Transfer,” Journal of Structural Engineering, ASCE,
in Civil Engineering and the Director of the Large-Scale Structures Labo- V. 126, No. 2, 2000, pp. 180-189.
ratory at the University of Alabama, Tuscaloosa, AL. He is Chair of ACI 18. Robertson, I. N., “Seismic Response of Connections in Indeterminate
Committee 133, Disaster Reconnaissance, and a member of ACI Committee Flat-Slab Subassemblies,” PhD thesis, Rice University, Houston, TX, 1990,
318, Structural Concrete Building Code; Joint ACI-ASCE Committee 352, 258 pp.
Joints and Connections in Monolithic Concrete Structures; and ACI 19. Robertson, I. N.; Kawai, T.; Lee, J.; and Enomoto, B., “Cyclic Testing
Technical Activities Committee’s Design Standards Committee. of Slab-Column Connections with Shear Reinforcement,” ACI Structural
Journal, V. 99, No. 5, Sept.-Oct. 2002, pp. 605-613.
20. Symonds, D. W.; Mitchell, D.; and Hawkins, N. M., Slab-Column
ACKNOWLEDGMENTS Connections Subjected to High Intensity Shears and Transferring Reversed
The tests were carried out in Bowen Laboratory for Large-Scale Civil Moments, Division of Structures and Mechanics, Department of Civil Engi-
Engineering Research of Purdue University with the sponsorship of the neering, University of Washington, Seattle, WA, 1976.
U.S. Army Experiment Station, Champaign, IL. The writers benefited 21. Wey, E. H., and Durrani, A. J., “Seismic Response of Interior Slab-
from the advice and understanding of J. Hayes and J. Wilcoski of the U.S. Column Connections with Shear Capitals,” ACI Structural Journal, V. 89,
Army Experiment Station, Champaign, IL, during the design, construction, No. 6, Nov.-Dec. 1992, pp. 682-691.
and testing of the flat plate. Hunt Construction and J. and K. Hockema 22. Zee, H. L., and Moehle, J. P., “Behavior of Interior and Exterior
are thanked for their generous contributions to the construction of the test Flat Plate Connections Subjected to Inelastic Load Reversals,” Report No.
structure. The support of H. Tidrick in the laboratory and S. Shriver in the UCB/EERC-84/07, University of California, Berkeley, Berkeley, CA, 1984,
departmental office was of great help. The advice of M. Schroeder of Shore 148 pp.
Western Manufacturing contributed to the success of the experiment. Grad- 23. Computers and Structures Inc., SAP2000, 2014, Berkeley, CA.
uate students who performed above and beyond the call of duty for the 24. Westergaard, H. M., “Measuring Earthquake Intensity in Pounds per
experimental work were F. Consuegra, T. Krahn, M. Murray, J. Rautenberg, Square Foot,” Engineering News-Record, V. 110, No. 16, 1933, 504 pp.
and R. Watve. 25. Sozen, M. A., “The Velocity of Displacement,” Seismic Assessment
and Rehabilitation of Existing Buildings, 2003, Springer Netherlands,
Dordrecht, the Netherlands, pp. 11-28.
REFERENCES 26. Consuegra, F., and Irfanoglu, A., “Variation of Small Amplitude
1. Sozen, M. A., and Siess, C. P., “Investigation of Multiple-Panel
Vibration Dynamic Properties with Displacement in Reinforced Concrete
Reinforced Concrete Floor Slabs. Design Methods—Their Evolution
Structures,” Experimental Mechanics, V. 52, No. 7, 2011, pp. 817-828.
and Comparison,” ACI Journal Proceedings, V. 60, No. 8, Aug. 1963,
pp. 999-1028.
2. Fick, D. R., “Experimental Investigation of a Full-Scale Flat-Plate
Reinforced Concrete Structure Subjected to Cyclic Lateral Loading in the

ACI Structural Journal/November-December 2017 1517

 NOTES:

1518 ACI Structural Journal/November-December 2017