EARTHQUAKE ENGINEERING AND STRUCTURAL DYNAMICS Earthquake Engng Struct. Dyn.

2006; 35:1827–1852 Published online 2 August 2006 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/eqe.612

Shake-table experiment on reinforced concrete structure containing masonry infill wall
Alidad Hashemi‡ and Khalid M. Mosalam∗,†,§
Department of Civil and Environmental Engineering, University of California, Berkeley, CA 94720-1710, U.S.A.

SUMMARY A hypothetical 5-storey prototype structure with reinforced concrete (RC) frame and unreinforced masonry (URM) wall is considered. The paper focuses on a shake-table experiment conducted on a substructure of this prototype consisting of the middle bays of its first storey. A test structure is constructed to represent the selected substructure and the relationship between demand parameters of the test structure and those of the prototype structure is established using computational modelling. The dynamic properties of the test structure are determined using a number of preliminary tests before performing the shake-table experiments. Based on these tests and results obtained from computational modelling of the test structure, the test ground motions and the sequence of shakings are determined. The results of the shake-table tests in terms of the global and local responses and the effects of the URM infill wall on the structural behaviour and the dynamic properties of the RC test structure are presented. Finally, the test results are compared to analytical ones obtained from further computational modelling of the test structure subjected to the measured shake-table accelerations. Copyright q 2006 John Wiley & Sons, Ltd.
Received 22 April 2006; Revised 13 June 2006; Accepted 13 June 2006 KEY WORDS:

earthquakes; infilled frame; modelling; reinforced concrete; shake-table; URM wall

INTRODUCTION Complex structures with multiple dissimilar components (hybrid systems) are frequently built in seismically active regions. Examples include reinforced concrete (RC) building frames with unreinforced masonry (URM) infill walls or steel bridge decks supported on RC piers. In order
∗ Correspondence

to: Khalid M. Mosalam, 721 Davis Hall, University of California, Berkeley, CA 94720-1710, U.S.A. E-mail: mosalam@ce.berkeley.edu ‡ Ph.D. Candidate. § Associate Professor. Contract/grant sponsor: National Science Foundation; contract/grant number: CMS0116005

Copyright q

2006 John Wiley & Sons, Ltd.

1828

A. HASHEMI AND K. M. MOSALAM

to develop new modelling techniques and study the behaviour of RC buildings with URM infill walls, a two-phase experimental and analytical study is conducted. Masonry infilled frames have been experimentally investigated for both in-plane and out-of-plane forces by many researchers. Most of these studies are focused on the behaviour of single-frame single-bay URM infilled frames under monotonic or cyclic lateral loading. Earlier studies can be found in Reference [1] among others. Some exceptions where multiple-bays and multiple-stories are tested can also be found in literature [2, 3]. These studies provide evaluations of the importance of infill confinement from bounding frames, the types of failure that can be observed in the infill or the bounding frame, the stiffness and strength of the infilled frames, and the degradation of strength upon load reversals. Since these tests are performed using monotonic, quasi-static, or pseudodynamic loading, it is not clear how well they represent the dynamic properties, e.g. damping characteristics, of a structure with masonry infilled frames subjected to earthquake loading. Limited data are available on dynamic properties of masonry infilled frames since very few shake-table experiments are performed on such structures. Fardis et al. [4] report on a shaketable test performed on single-bay two-storey RC frames with eccentric (non-symmetric in plan) masonry infills subjected to bi-directional ground acceleration. The study focused on the effects of the eccentricity on displacement demands on the corner columns. Zarnic et al. [5] report on two shake-table tests performed on 1 -scale one and two-storey RC frames with strong-block 4 weak-mortar masonry infill walls subjected to one-directional sinusoidal motion at the base of the 1 structure. Dolce et al. [6] report on shake-table tests performed on two-dimensional 3.3 -scale threestorey two-bay RC frames designed for low seismicity regions without infill, with masonry infill and with two different types of energy dissipating and re-centring braces. The study compared the overall response and the dynamic properties of the three frames subjected to a sequence of artificially generated accelerograms with increasing intensity. The mentioned experiments and others are generally performed on small-scale models due to the size limitations of the shake-tables and are focused on different aspects of the problem, e.g. torsional effects due to eccentric infill walls. The current experiment is conducted to study the dynamic performance of a seismically designed symmetric large-scale URM-infilled RC frame subjected to real ground motions. The study focuses on evaluating the effects of the URM infill wall on the surrounding structural elements, namely the RC slab and the RC columns. Moreover, since the structure consists of two frames without infills and an infilled frame, comparing the frame responses and quantifying the distribution of lateral forces between the tested frames before and after infill damage is an important and novel objective of the study. The current study is unique in the aspect of testing a structure with dissimilar frames. Accordingly, the presented shake-table experiment can be viewed as a benchmark for experimentation using mixed-variables (force and displacement) pseudo-dynamic technique with substructuring [7] and for analytical modelling of URM infilled frames. The shake-table experiment is carried out on a reduced-scale one-storey RC moment-resisting frame structure with URM infill wall on the seismic simulator test facility of the University of California, Berkeley. The 3 -scale test structure represents the first storey middle bays of a 4 5-storey RC prototype structure designed based on the requirements of ACI318-02 [8] and NEHRP recommendations [9] in seismic regions. URM walls are assumed in the interior frames, Figure 1. The experimental study serves the purpose of calibrating analytical models being developed using Open System for Earthquake Engineering Simulation (OpenSees) [10]. The objectives of the modelling effort are to enable accurate representation of the in-plane behaviour of URM infill walls, and to refine the modelling techniques of hysteretic strength and stiffness degradation in RC elements and joints of RC moment frames interacting with URM infill walls.
Copyright q 2006 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2006; 35:1827–1852 DOI: 10.1002/eqe

SHAKE-TABLE EXPERIMENT ON REINFORCED CONCRETE STRUCTURE

1829

C2 B0 A0 (a) A1 A2 B1 B2 A3 B3

C3 N (b)

Figure 1. Development of the shake-table test structure: (a) prototype structure; and (b) test structure on the shake-table.

PROTOTYPE STRUCTURE The 5-storey prototype moment-resisting frame structure is designed with its exterior columns (A0, A3, B0, B3, C0, and C3), as the primary lateral load resisting system. Although it is common to have masonry walls on the perimeter of buildings, the URM walls are assumed in the middle frames for two practical reasons. First, having two walls on the perimeter for the test structure would have required about twice as much shear force to damage the walls, thus exceeding the capacity of the shake-table. Second, the failure of the two URM walls would not have been simultaneous due to inherent material and construction variability and that would have caused a significant shift in the centre of rigidity of the test structure compared to its centre of mass and produced large torsional demands with loading patterns out of scope of the study. Due to the limited size of the available shake-table, the prototype structure is scaled to 75% of its original design size. A typical floor plan of the 3 -scaled prototype building is shown in Figure 2. Uniformly 4 distributed superimposed dead and live loads on the scaled prototype structure (DL and LL) are independent of the length scale factor. For each floor, the DL and LL are 160 Pa (110 psf) and 74 Pa (50 psf), respectively, and for the roof, they are 130 Pa (90 psf) and 15 Pa (10 psf), respectively. In the following text, the scaled prototype structure is referred to as the prototype structure. The prototype substructure is selected as the middle bays of the first storey of the prototype building as shown in Figure 2. Although these bays are typically not a part of the primary lateral load resisting system, they are detailed as such because they would have to endure the same displacement demands in an earthquake event and maintain their axial load carrying capacity. The test structure is designed to represent the prototype substructure as described before with the exception that the transverse span is reduced to 1 its prototype length due to the size limitation of the shake-table. 2 In order to determine the boundary conditions and necessary adjustments for this representation, an analytical model of the prototype building is constructed using OpenSees, Figure 3. Beams and columns are modelled using nonlinearBeamColumn element in OpenSees, which is based on force formulation, and considers the spread of plasticity along the length of the element. Sections are defined using fibre discretization with distinct fibres for longitudinal reinforcement. Concrete material is modelled using Concrete01, which is a uniaxial concrete material object with degraded linear unloading/reloading stiffness in compression and no tensile strength. Confining effect due
Copyright q 2006 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2006; 35:1827–1852 DOI: 10.1002/eqe

1830

A. HASHEMI AND K. M. MOSALAM

0 4110 C 3660 340x265 230x305 TYP. TYP.

1 4110 340x265 TYP.

2 4110 340x265 TYP .

3

C530x530

B 230x305 TYP . 3660 URM INFILL Transverse direction Experiment substructure C305x305 Longitudinal direction

A

Figure 2. Floor plan of the 3 -scaled prototype structure (dimensions in mm). 4

Figure 3. Computational model. Table I. Concrete model properties.
Property Peak compressive stress (MPa (ksi)) Strain at peak compressive stress Ultimate strain Stress at ultimate strain (MPa (ksi)) Foundation 34.4 (4.98) 0.002 0.006 0 Beam 37.2 (5.39) 0.002 0.006 0 Column cover 38.4 (5.56) 0.002 0.006 0 Column core 45.3 (6.57) 0.004 0.020 6.90 (1.00)

to the prescribed transverse reinforcement is accounted for using confined concrete properties for column core concrete material [11, 12]. Steel reinforcing bars are modelled using Steel01, which is a uniaxial bilinear material object with kinematic hardening. The material properties used for the elements of the model are defined in Tables I and II. The column–footing joints are modelled using the recommendations of FEMA 356 [13] by a tri-linear moment-rotation relationship as
Copyright q 2006 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2006; 35:1827–1852 DOI: 10.1002/eqe

SHAKE-TABLE EXPERIMENT ON REINFORCED CONCRETE STRUCTURE

1831

Table II. Steel model properties.
Property Yield stress (MPa (ksi)) Yield strain (rad) Modulus of elasticity (GPa (ksi)) Kinematic hardening ratio Parameter 458 (66.5) 0.0023 200 (29 000) 0.01

M MP MY

Mcr
cr Y P

Figure 4. Column–footing joint model.

Table III. Column–footing joint model properties (refer to Figure 4).
Mcr (kN m (kip in)) cr (rad) MY (kN m (kip in)) Y (rad) M P (kN m (kip in)) P (rad) 29.9 (265) 0.002 130 (1150) 0.015 158 (1400) 0.030

Strut
f mo Parabola

Straight line
f mu
mo mu

Tension

Compression

Figure 5. URM infill strut model.

shown in Figure 4 and Table III. The masonry infill wall is modelled using equivalent diagonal compression-only struts as shown in Figure 5 and Table IV. For the 102 mm (4 ) thick URM wall, the cross-sectional area of the equivalent strut is estimated using FEMA 356 guidelines as 462 cm2 (71.6 in2 ). This value proved inconsistent with the test results and later modified based on the estimated stiffness and strength of the infill panel during snap-back and shake-table tests
Copyright q 2006 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2006; 35:1827–1852 DOI: 10.1002/eqe

1832

A. HASHEMI AND K. M. MOSALAM

Table IV. Masonry strut properties (refer to Figure 5).
f mo (MPa (ksi))
mo

17.0 (2.46) 0.0028 1.99 (0.29) 0.0041

f mu (MPa (ksi))
mu

900 750

0

50

[kN-m] 100

150

200 4000 3000

600 Axial load [kip] 450 300 150 0 -150 -1000 -300 0 500 1000 Moment [kip-in] 1500 0 [kN] Prototype substructure Test structure 2000 1000

Figure 6. Range of change in axial load of the middle columns.

to 223 cm2 (34.6 in2 ). The concrete slab is modelled using horizontal elastic truss members. The material properties and the area of these truss members are determined such that they have the same stiffness as the concrete slab, which is expected to remain elastic during the test. Using the OpenSees model, a non-linear time history analysis of the prototype structure subjected to different levels of selected ground motion is performed. These levels are discussed in subsequent sections. Analysing the results, the axial load and base shear affecting the prototype substructure are determined for each level of input ground motion. With the same assumptions as the prototype structure, an OpenSees model of the test structure is analysed and the required amount of additional mass is determined for the test structure. This determination is based on matching the computationally determined base shear of the test structure to that of the prototype substructure when subjected to the design-level ground motion. In order to explore the effects of the column axial loads due to the weight of the upper stories in the prototype structure, the range of change in the axial force during each level of ground motion is superposed on the moment–axial interaction diagram for the middle columns. An example of such results for the design-level ground motion is shown in Figure 6. The static axial force on the column due to upper storey dead and live loads is 386 kN (86.8 kips) corresponding to 160 kN m (1420 k in) moment capacity for the column section. During the design level ground motion, the axial load ranges from 220 (49.5 kips) to 429 kN (96.5 kips) and the corresponding moment capacity of the section ranges from 144 (1270 k in) to 164 kN m (1450 k in). Comparing this variation with the moment capacity of the section at zero axial force, 120 kN m (1060 k in),
Copyright q 2006 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2006; 35:1827–1852 DOI: 10.1002/eqe

SHAKE-TABLE EXPERIMENT ON REINFORCED CONCRETE STRUCTURE

1833

150 Maximum base shear [kips]

0

5

[mm] 10

15

20 Hysteretic energy [kip-in] 600 500

250 200 150 15 100 50 0 (b) 10 5 0 20 40 60 80 Time [sec] 100 0 120 Prototype substructure Test structure 25 20 [kN-m]

100 300 50 200 Prototype substructure 100 Test structure 0 0.2 0.4 0.6 0.8 Corresponding displacement [in] [kN] 400

0 (a)

0

Figure 7. Comparison between prototype substructure and test structure: (a) max. base shear versus corresponding displacement; and (b) cumulative hysteretic energy.

it is concluded that the axial load on the column has significant effect on its flexural capacity (up to 37% increase) of the section. To accommodate this observation for the test structure, the static column axial load due to weight of the upper stories is applied to the columns in the form of unbonded concentric prestressing. The resulting range of change in the axial load in the columns corresponding to the design level ground motion for the test structure with the axial prestressing is also obtained using OpenSees model and shown to be very close to that of the prototype as shown in Figure 6. Comparison between the response of the prototype substructure and the test structure when subjected to different levels of selected input ground motion is performed using non-linear time history analyses. The plots for maximum base shear versus its corresponding first storey displacement for both the prototype substructure and the test structure for different levels of the input ground motion are shown in Figure 7(a). The cumulative hysteretic energy plots for these consecutive runs are compared for the two structures in Figure 7(b). These results show reasonable agreements, e.g. less than 13% mismatch for the cumulative hysteretic energy at the end of all runs, between the response of the prototype substructure and the test structure. As mentioned earlier, by providing additional mass and prestressing in the test structure, an acceptable match between the responses of the two structures is achieved. However, the shortcomings are due to the effects of the overturning moment and higher modes in the prototype structure. In general, the overturning moments have considerable impact on the axial force of the columns on the perimeter of the structure. These effects are less significant when one considers the middle columns. Preliminary analysis of the prototype structure shows that the variation of axial load in the first storey columns due to the overturning moments is about 175% for perimeter column (e.g. column A0) and only about 15% for middle columns (e.g. column A1). To explore the effects of higher modes on the prototype structure, a modal response spectrum analysis [14] for the prototype structure subjected to the NEHRP design spectra is performed. The resulting modal base shears and first storey displacements are compared and it is concluded that the effects of higher modes is limited to about 10% of the first mode effects. Accordingly, the effects of overturning moments and higher modes are neglected when relating the test structure to the prototype.
Copyright q 2006 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2006; 35:1827–1852 DOI: 10.1002/eqe

1834

A. HASHEMI AND K. M. MOSALAM

CONFIGURATION AND INSTRUMENTATION OF THE TEST STRUCTURE The overall dimensions of the test structure are 4.88 m × 4.42 m (16 -0 × 14 -6 ) in plan and 3.43 m (11 -3 ) in height, Figure 2. Member sizes and reinforcement details for different elements of the test structure are summarized in Table V. The reinforcement is specified as ASTM A615 [15] Grade 60. The yield stress of the 19 mm diameter (#6) bars is 458 MPa (66.5 ksi) per mill certification. The specified 28-day compressive strength of the standard concrete cylinder per ASTM C 837-99 [16] is 31 MPa (4.5 ksi). The masonry wall is made of clay bricks with modular size of 102 mm × 203 mm × 68 mm (4 × 8 × 2 2 ) and ASTM C270 [17] Type N mortar. The 3 measured average 28-day compressive strength of the standard masonry prism according to ASTM C 1314 [18] is 17 MPa (2.46 ksi). The average shear strength for 102 mm (4 ) thick masonry panels constructed and measured in accordance with ASTM E519 [19] is 1.81 MPa (263 psi). Uniformly distributed mass is added to the slab in the form of stacked lead ingots bolted to the slab using 10 mm ( 3 ) diameter high strength rods. Static tests confirmed that the friction 8 forces between the slab and the lead ingots are large enough to accommodate up to 4.0g lateral acceleration at the slab level. To measure the floor acceleration in three directions, 11 accelerometers are installed on the floor level (Figure 8) as follows: in the longitudinal direction, one at each of the six columns and one at the middle of the beam in frame B; in the transverse direction, three on the diagonal of the slab on opposite corners and at the centre; and in the vertical direction, one at the middle of the

Table V. Member sized and reinforcement details for the test structure.
Structural element Concrete slab Dimensions 95 mm (3 3 ) thick 4 Main reinforcement M10 (#3) top and bottom @ 305 mm (12 ) o.c. each way 8–19 mm diam. (#6), 32 mm diam. (1 1 ) prestressing 4 rod 3–19 mm diameter (#6) top and bottom None Transverse reinforcement

Columns

305 mm × 305 mm (12 × 12 )

M10@95 mm (#3@3 3 ) over 610 mm 4 (24 ) from the face of the joints and M10@152 mm (#3@6 ) elsewhere

Longitudinal beams (single span)

267 mm × 343 mm (10 1 × 13 1 ) 2 2

M10@70 mm (#3@2 3 ) over the 4 711 mm (28 ) from the face of the beam-column joint and M10@203 mm (#3@8 ) elsewhere M10@305 mm (#3@12 ) M10@102 mm (#3@4 ) None

Short direction beams 305 mm × 229 mm (double span) (12 × 9 ) Footing Masonry wall
Copyright q

2–19 mm diameter (#6) top and bottom 4–22 mm diameter (#7) top and bottom None

356 mm × 457 mm (14 × 18 ) 102 mm (4 ) thick

2006 John Wiley & Sons, Ltd.

Earthquake Engng Struct. Dyn. 2006; 35:1827–1852 DOI: 10.1002/eqe

SHAKE-TABLE EXPERIMENT ON REINFORCED CONCRETE STRUCTURE

1835

A

B

C

Shake-table outline

A
6'-0" [1829mm]
A10

B
6'-0" [1829 mm]

C
A12

1

1
A14

A11 D08

20'-0" [6096 mm] Square

19'-0" [5791 mm]

13'-6" [4115 mm]

A01 D03 A02 A03

Accelerometer
A04 A06 A05

Displacement transducer

A13 D07

2

2
A07

A08 A09 D04 D05 D06

D02

D01

FOUNDATION PLAN

FLOOR PLAN

Figure 8. Global instrumentation plan for the test structure.

beam in frame B. Moreover, three accelerometers at the base of the URM wall on the footing (one in each direction) and eight accelerometers built-in the actuators beneath the shake-table (one per horizontal and vertical actuator) are monitored. To measure global displacements of the shake-table and the test structure with respect to the stationary ground, eight displacement transducers (Figure 8) are used. Five of them measure the displacement of the first floor (three in the longitudinal direction and two in the transverse direction) and the remaining three measure the displacements of the shake-table itself (two in the longitudinal direction and one in the transverse direction). To measure local displacements and rotations, 75 displacement transducers are used: nine are installed in the plane of the slab, 14 are installed in the plane of the URM wall measuring the diagonal deformations, wall sliding and opening with respect to the bounding frame, four are measuring the diagonal displacements of the frames and the rest are installed on four of the columns (12 per column) to measure the rotations and average curvatures along the length of the column. Finally, 78 strain gauges out of more than 150 strain gauges installed on the reinforcing bars throughout the test structure are used during each run of the shake-table. As some of the strain gauges are damaged during the different test runs, alternative gauges are selected and monitored. The prestressing rods are also gauged and monitored during all test runs. The layouts of these local measurements, e.g. Figures 10, 17, and 20, are presented with results in the relevant sections.

SYSTEM PROPERTIES Pull (snap-back) tests are performed on the test structure before and after the wall construction to determine the stiffness, natural frequency and damping ratio of the structural system before starting the shake-table experiment. These tests are conducted for both in-plane and
Copyright q 2006 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2006; 35:1827–1852 DOI: 10.1002/eqe

1836

A. HASHEMI AND K. M. MOSALAM

Table VI. Snap-back test results (refer to Figure 1(c) for orientation of the North (N) direction).
In-plane (North–South direction) Conditions of the test structure at time of the pull (snap-back) test Before building the wall columns not prestressed no additional mass After building the wall columns prestressed no additional mass After building the wall columns prestressed with additional mass Natural period (s) 0.135 0.055 0.134 Damping ratio (%) 4.30 5.70 6.85 Stiffness (kN/mm (kips/in)) 19.8 (113.3) 74.5 (425.5) 75.5 (431.0) Out-of-plane (East–West direction) Natural period (s) 0.134 0.122 0.232 Damping ratio (%) 4.40 4.30 4.25 Stiffness (kN/mm (kips/in)) 23.5 (134.0) 29.3 (167.1) 29.4 (168.0)

Table VII. Ground motion specifications.
Ground motion Northridge, CA, 1994 Duzce, Turkey, 1999 Station Tarzana Lamont Direction 090 N PGA (g) 1.570 0.762 PGV (mm/s (in/s)) 920 (36.23) 329 (12.97) PGD (mm (in)) 130 (5.13) 19 (0.75)

Table VIII. Scale factors for different levels of input ground motions.
Level Northridge, CA, 1994 (TAR) Duzce, Turkey, 1999 (DUZ) 1 0.05 — 2 0.17 — 3 0.23 — 4 0.39 — 6 0.59 — 7 — 1.50 8 — 2.00

out-of-plane directions of the test structure, separately. The results of these tests are summarized in Table VI.

GROUND MOTIONS Two different ground motions as described in Table VII are used in the experiment. In this table, PGA, PGV, and PGD refer to peak ground acceleration, velocity, and displacement, respectively. These ground motions are intended to be unidirectional in the direction parallel to the URM infill wall of the test structure (longitudinal direction). Each ground motion is scaled to generate different levels of intensity as listed in Table VIII. The scaling is based on the average spectral acceleration of selected ground motions and the NEHRP design spectrum for a site with mapped spectral response acceleration at 1 s, S1 = 0.65g and at short periods, Ss = 1.60g and site class D with 5% damping over the range that the period of structure is expected to vary during the experiment. For Northridge Tarzana and Duzce ground motions, this range is estimated from the
Copyright q 2006 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2006; 35:1827–1852 DOI: 10.1002/eqe

SHAKE-TABLE EXPERIMENT ON REINFORCED CONCRETE STRUCTURE

1837

2.5

Begining of the test (TAR 1) TAR 4

Duzce Northridge Design Spectrum
1 0 Acceleration [g]

2 Spectral acceleration [g]

Northridge, Tarzana
1.57g

TAR 6

After removal of the wall

1.5

-1 0 5 10 Duzce 1 0 15

1

0.5

DUZ 8 DUZ 7

-1 0 5 10 Time [sec] 15

0

0

0.2

0.4 0.6 Period [sec]

0.8

1

Figure 9. Response spectra (5% damping) for the selected ground motions.

period of the undamaged infilled structure to the period of the structure after removal of the infill (between 0.15 and 0.28 s). The scaled spectra of the two selected ground motions and the design spectrum are shown in Figure 9. In this figure the natural period of the test structure corresponding to important milestones of the experiment are marked. The test structure is subjected to a sequence of ground motions starting with Northridge Tarzana level 1–6 (denoted by TAR 1 through TAR 6) and Duzce levels 7, 8 and finally 7 again (denoted by DUZ 7, DUZ 8 and DUZ 7-2). Level TAR 1 is selected as a small amplitude motion to check the performance of the shake-table and data acquisition system. Levels TAR 2 and TAR 3 are selected as intermediate intensity levels and levels TAR 4 and TAR 6¶ correspond to 10/50 (design) and 2/50 (MCE) spectra, respectively. Note that, e.g. 10/50 means 10% probability of being exceeded in 50 years. Levels DUZ 7 and DUZ 8 are selected to achieve higher demands on the test structure up to the limits of the shake-table. √ The ground motion records are compressed in time by 3/4 factor to account for the reduced length scale of the test structure. In this way the frequency content of the compressed record at the natural frequency of the undamaged reduced-scale test structure is the same as that of the un-compressed record at the natural frequency of the test structure without scaling.

SYSTEM IDENTIFICATION Figure 10 shows the methodology used to find the average damping coefficient and average stiffness of the test structure during each run and the distribution of forces in different elements of the test structure. For this purpose, the test structure is idealized as a single degree of freedom (SDOF) system represented by the average floor displacement. The total floor acceleration u t is calculated ¨ by taking the weighted average of the measured accelerations on the floor. Dividing the floor
¶ During

the last stages of planning for the sequence of ground motion, level TAR 5 with PGA of 0.49g is deemed redundant and is not used for the actual experiment. 2006 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2006; 35:1827–1852 DOI: 10.1002/eqe

Copyright q

1838
mu t m mu
mu mu

A. HASHEMI AND K. M. MOSALAM

F I

m

u

ut

FS

FI

ˆ cu
h H H

System Identification:
u ut ug u F I u dt , u u dt 2 ˆ ˆ (k u c u) Re gression k , c

6

d1
top 1 top 2
top

Fwall
, M top

FS
i 1

Vcol

M top Vtop

Vcol Vtop

h

Vcol

M top h

M bot
Vtop Fwall M bot

Fwall h 2

bot 1

bot 2
bot

, M bot

Measurements for column i

Figure 10. System identification and member shear force calculation in the test structure.

into three strips consistent with the tributary area of each frame and designating the mass and the measured acceleration associated with the ith strip as m i and u i , respectively, the total floor ¨t t and the total inertia force acting on the structure F are determined as follows: acceleration u ¨ I
3 3

ut = ¨ FI =

i=1 3

mi ui ¨t mi ui ¨t

mi
i=1

(1)

(2)

i=1

The acceleration of the floor relative to the shake-table u is calculated by subtracting the ground ¨ acceleration from u t . The corresponding relative velocity and displacements are determined by ¨ integrating the relative acceleration in time. The dynamic equilibrium equation for the idealized SDOF system can be written as FI + FD + FS = 0 at each instant of time. With the assumption of viscous damping, FD = cu is the damping force and FS is the restoring force of the test structure. ˙ Assuming a constant damping coefficient c, and a constant average stiffness over the duration of ˆ each run, the least-square estimates of damping coefficient c and average stiffness k are obtained ˆ using regression function FI = − (ku + cu) in vector space of (u, u) considering all the data points ˙ ˙ in the duration of each run. Using the estimated value of the damping coefficient, the restoring
Copyright q 2006 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2006; 35:1827–1852 DOI: 10.1002/eqe

SHAKE-TABLE EXPERIMENT ON REINFORCED CONCRETE STRUCTURE

1839

ˆ˙ force in the structure is calculated from the dynamic equilibrium, i.e. FS = −FI − cu referred to as the total restoring force of the test structure. The portion of the total restoring force carried by each column can be calculated using the data from the strain gauges located at the top and bottom of each column. By making the Bernoulli assumption of plane section remains plane after bending at those locations and knowing the geometric dimensions, section curvature is calculated as in Equation (3), where d1 is the distance between the two strain-gauges in the section. Using strains, curvature and constitutive relationships for the reinforcing bars and concrete, the bending moments at each end of the column segment between the two pairs of strain-gauges (top and bottom) are determined by section analysis. From equilibrium, column shear is determined from Equation (4):
=(
1

2 )/d1

(3) (4)

Vcol = (Mtop + Mbot )/ h

where Mtop , Mbot , and h are defined in Figure 10. For the middle frame columns where there is contact between the URM wall and the column, the equation for column shear force above the contact length is rewritten as in Equation (5), where h is the contact length between the URM infill and the RC column segment bounded by the two instrumented sections and FWall is the horizontal component of the portion of the force in the URM infill wall that is transferred to the column within the portion of contact length H namely h as shown in Figure 10. Since the strain-gauges are located at a section well above the column–footing joint, the values of both and FWall are relatively small and the second term of Equation (5) is neglected in the calculations of the middle column shear force. Vcol = (Mtop + Mbot )/ h − FWall ( /2) (5)

Finally, the shear force in the URM infill wall is calculated as the total restoring force minus the sum of shear forces in all six columns comprising the test structure RC framing, Figure 10.

TEST RESULTS AND DISCUSSION The intact test structure with the URM wall is subjected to the sequence of ground motions as described before. The global response of the structure in terms of the overall drift, stiffness and damping ratio as well as the local response of the different elements of the test structure are quantified and discussed. Global response The total base shear versus displacement plots for selected test levels are presented in Figure 11. Total base shear is defined in the following as the sum of restoring and damping forces in the structure, which from dynamic equilibrium is the inertial forces FI as described before. In the discussion of the results presented in Figure 11, the stiffness is estimated by the tangent stiffness of the loading branch evaluated at different times during the experiments. During levels TAR 1 and TAR 2, there is no considerable change in the stiffness of the structure. The plot for level TAR 3 (Figure 11(a)) shows slight reduction in stiffness (about 9% from 75.5 kN/mm (431 kips/in)
Copyright q 2006 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2006; 35:1827–1852 DOI: 10.1002/eqe

1840
[mm]
200 150 -60 -40 -20 0 20

A. HASHEMI AND K. M. MOSALAM

[mm]
40 60 800 600 200 150 -60 -40 -20 0 20 40 60 800 600

Base shear [kips]

50 0 -50

K=391 kips/in

Base shear [kips]

100

400 200

100 50 0 -50 -100 -150 -200 -3 -2 -1 0

K=365 kips/in

400 200 0 -200 -400 -600 -800

[kN]

0 -200 -400 -600 -800 -2 -1 0 1 2 3

-100 -150 -200 -3

1

2

3

(a)

Displacement [in.] [mm]
200 150 -60 -40 -20 0 20 40 60 800 600

(b)
-60 -40

Displacement [in.] [mm]
200 150 -20 0 20 40 60 800 600 400 200 0

Base shear [kips]

50 0 -50 Final stiffness

Base shear [kips]

100

Initial stiffness K=364 kips/in

400 200

100 50 0 -50 -100 -150 -200 -3

Initial stiffness K=278 kips/in

K=160 kips/in

[kN]

0 -200 -400 -600 -800 -2 -1 0 1 2 3

K=62 kips/in

-200 -400

K=289 kips/in -100
-150 -200 -3

K=159 kips/in
-2 -1 0 1 2

-600 -800 3

(c)
-60 -40

Displacement [in.] [mm]
200 150 -20 0 20 40 60 800 600

(d)
-60 -40

Displacement [in.] [mm]
200 150 -20 0 20 40 60 800 600 400 200 0 -200 -400 -600 -800 -2 -1 0 1 2 3

Initial stiffness K=116 kips/in Static friction K=281 kips/in

Final stiffness K=51 kips/in Static friction K=177 kips/in

Base shear [kips]

50 0 -50 -100 -150 -200 -3

Base shear [kips]

100

400 200

100 50 0 -50 -100 -150 -200 -3

[kN]

0 -200 -400

Final stiffness K=63 kips/in
-2 -1 0 1 2 3

-600 -800

(e)

Displacement [in.]

(f)

Displacement [in.]

Figure 11. Total base shear versus lateral displacement for different test levels: (a) TAR 3; (b) TAR 4; (c) TAR 6; (d) DUZ 7; (e) DUZ 8; and ( f ) DUZ 7-2.

Copyright q

2006 John Wiley & Sons, Ltd.

Earthquake Engng Struct. Dyn. 2006; 35:1827–1852 DOI: 10.1002/eqe

[kN]

[kN]

[kN]

SHAKE-TABLE EXPERIMENT ON REINFORCED CONCRETE STRUCTURE

1841

of the pull test results) to 68.4 kN/mm (391 kips/in) but there is no visible sign of damage in the test structure. The response of the test structure in level TAR 4 (design level) as depicted in Figure 11(b) shows some drop in the stiffness (about 15% reduction), but the overall behaviour remains almost linear. Close observation of the wall after the completion of TAR 4 reveals small visible cracks at the wall–column interfaces. Figure 11(c), corresponding to level TAR 6, shows the first significant signs of damage. The stiffness of 63.7 kN/mm (364 kips/in) in the initial motion shifts to 50.6 kN/mm (289 kips/in) at the peak of the ground motion (21% shift). Observations after the test suggests that some cracks are developing especially along the column-wall interface and some small vertical splitting cracks are observed in the mortar head joints at the wall corners. The maximum total base shear reached in this stage is 605 kN (136 kips) corresponding to 144% of the total gravity load. The response of the test structure during level DUZ 7, Figure 11(d), shows the most significant change in the structural behaviour of the system. The stiffness at the beginning of this level is 48.7 kN/mm (278 kips/in). Significant wall cracks with clear pattern and load path definition are formed during this test level. The force–displacement behaviour of the assembly at this point can be described as a bi-linear relationship. For small displacements (less than about 6 mm ( 1 )), the 4 cracks on the wall open and close without engaging the URM infill wall resulting in an observed lateral stiffness of about 10.9 kN/mm (62 kips/in). This stiffness can be interpreted as the stiffness of the cracked RC frame before full contact with the URM wall is reached. Once the cracks close, the URM infill wall picks up the load causing further damage in the wall and stiffness increase to 28.0 kN/mm (160 kips/in). The peak total base shear observed during all stages of the test, namely 756 kN (170 kips) corresponding to 180% of total gravity load, at floor displacement of 19.1 mm (0.75 ) is produced at this stage of the experiment and right before a major horizontal crack in the URM infill wall is developed. Figure 11(e) shows the gradual disintegration of the URM infill wall as the test structure is cycled back and forth in level DUZ 8. The measured stiffness of the test structure at the beginning of DUZ 8 is 49.2 kN/mm (281 kips/in) for small displacements. Comparison between this stiffness and that of the previous run suggests that at small force demands, the force transferred through the wall is not enough to overcome the static friction between the cracked surfaces. Accordingly, at such small forces, the wall, although cracked, acts as a whole increasing the apparent stiffness of the structural system. Once the force demands at the crack surfaces exceed the static friction (at about 110 kN (25 kips) corresponding to 27% of total gravity load), the cracked portions start to move with respect to each other and the stiffness reduces to that of RC frames including cracked URM infill wall, i.e. 20.3 kN/mm (116 kips/in) as shown in Figure 11(e). Note that the stiffness of the test structure before building the wall (Table VI) is 19.8 kN/mm (113 kips/in). From this point on, the URM infill wall is considered as structurally insignificant. As the test structure goes through large displacements, the RC frame starts to accumulate damage mostly concentrated at the bases of the columns. The stiffness of the test structure at the end of this level is reduced to 11.0 kN/mm (63 kips/in) suggesting significant damage in the RC frame in addition to the URM infill wall collapse. Finally, Figure 11(f ) shows the results obtained from the repeat run of level DUZ 7, i.e. DUZ 7-2. Beyond the static friction at the beginning of the motion where the stiffness is high at a value of 31.0 kN/mm (177 kips/in), the stiffness of the test structure is 8.9 kN/mm (51 kips/in), i.e. about 20% reduction from that at the end of DUZ 8. Figure 12(a) shows the change in the ‘effective’ stiffness of the test structure for all levels of testing plotted along with the measured initial stiffness of the test structure with and without the
Copyright q 2006 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2006; 35:1827–1852 DOI: 10.1002/eqe

1842

A. HASHEMI AND K. M. MOSALAM

0.5
400

70 60

Results using ground motion signals Results using white-noise signals Wall removal

Effective stiffness [kips/in]

350 300 250

0.4 Period [sec]
DUZ 8

50 40

200 150 100 50 0 TAR 1

Test results Initial stiffness with wall Initial stifness without wall

[kN/mm]

0.3
Cracking of the wall

30 20 10 0

0.2

0.1 0

TAR 2

TAR 3

TAR 4

TAR 6

DUZ 7 DUZ 8 DUZ 7-2

(a)

Test progress
(b)

TAR 1

TAR3

TAR 6 DUZ 7 Test level

DUZ 7-2

Figure 12. Variation of dynamic properties of the test structure: (a) effective stiffness; and (b) natural period.

URM infill wall. This effective stiffness denotes the average tangent stiffness of the test structure after the demand force exceeds the initial static friction. It can be observed that the existence of the URM wall considerably increases the stiffness of the structural system. As the wall undergoes damage, the stiffness reduces with most rapid reduction taking place during levels TAR 6, DUZ 7 and DUZ 8 suggesting significant disintegration of the wall in these levels. Note that the stiffness of the test structure reduces to a level less than the stiffness of the elastic structure without the wall, which is due to the damage at column–footing and beam–column joints in the RC frame structure. Corresponding to the change in the stiffness of the structure, there is significant change in the natural period of the test structure. In order to identify this change throughout the test runs, a low amplitude white noise signal with approximately constant small amplitude of 0.07g over frequency range of 1-10 Hz is applied before each run and the resulting acceleration of the floor is analysed in frequency domain. The period corresponding to the peak amplitude of the frequency response is taken as the natural period of the test structure before each run. Alternatively, the transfer function of the test structure obtained from the ratio of the Fourier transform of the measured floor and base accelerations is examined. The effective natural frequency of the structure during each level of the test is defined as the frequency corresponding to the peak in this transfer function. Figure 12(b) demonstrates the variations in natural period of the test structure as the shake-table experiment progressed, which are determined using both the white noise and ground motion signals. Because the white noise tests are performed with very low amplitudes, existing cracks in the structure remain closed and the structure appears stiffer. Thus, the resulting natural periods from whitenoise tests represent a lower bound for the natural period of the test structure during the actual ground motion. Significant elongation in the natural period of the test structure is observed, i.e. from 0.147 s during TAR 1 to 0.392 s during DUZ 7-2 with 167% elongation. Another indication of the change in the dynamic properties of the test structure due to the induced damage during the shake-table experiment is the change in the damping ratio. The average damping coefficient c during each shake-table run is obtained with the procedure described earlier, Figure 10. The damping ratio is subsequently obtained as = c/ccr where ccr = 2m n is the critical damping
Copyright q 2006 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2006; 35:1827–1852 DOI: 10.1002/eqe

SHAKE-TABLE EXPERIMENT ON REINFORCED CONCRETE STRUCTURE

1843

14 12 Regression Energy equivalent

Damping ratio [%]

10 8 6 4 2 0

TAR 1 TAR 2 TAR 3 TAR 4 TAR 6 DUZ 7 DUZ 8 DUZ 7-2

Test levels

Figure 13. Variation of the equivalent damping ratio.

of the test structure where m and n are the total mass and natural frequency, respectively, of the test structure. Alternatively, the equivalent damping ratio can be estimated as 1/4 times the ratio of the dissipated energy to the maximum strain energy in each cycle [14]. The resulting variation of the damping ratio for different levels of shaking calculated using both methods is shown in Figure 13. It should be emphasized that while the regression method, assuming an average linear response, only limits the damping to the velocity related term in the equation of motion, the equivalent energy method includes both the viscous damping and the inelastic deformations as sources of dissipating energy. Both methods indicate smaller damping ratios (4–6% on average) for the intact structure (levels TAR 1–TAR 4) and large damping ratios (11–13%) where significant damage occurs in the test structure (levels TAR 6–DUZ 7-2). Local response URM infill. The main failure mode of the URM infill wall takes place in level DUZ 7 and is characterized by large cracks at 60◦ from the horizontal axis starting from the top corners of the wall and connecting with a long horizontal crack at the lower third of the wall to a series of 45◦ cracks propagating into the opposite bottom corners along each of the wall diagonals. At the same time early signs of corner crushing are observed at the top corners. The markings in Figure 14(a) show the observed crack pattern in the wall after DUZ 7. The sharp angle of cracking at the top corner is attributed to the weaker bond between the upper most mortar bed joint and the RC beam relative to the bond between the side vertical mortar joints and the RC columns. Partial collapse of the top corners and sides of the wall follows the formation of the crack pattern in level DUZ 8. The markings in Figure 14(b) show the final crack pattern on the wall at the end of level DUZ 8. Finally, in level DUZ 7-2 the loose portions of the wall collapsed leading to the damage state shown in Figure 14(c). The shear forces carried by the URM infill wall and by each RC column are calculated as described earlier, Figure 10. The shear force carried by the RC frame is the sum of all the shear forces of the six columns. Figure 15 shows the portion of the total base shear which is carried
Copyright q 2006 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2006; 35:1827–1852 DOI: 10.1002/eqe

1844

A. HASHEMI AND K. M. MOSALAM

(a)

(b)

(c)

Figure 14. Observed damage of the test structure: (a) cracking after Duzce 7; (b) partial collapse after Duzce 8; and (c) final state after Duzce 7-2.

180 160 140 Base shear [kips] 120 100 URM wall RC Frame

800 700 600 500 [kN] 400 Base shear [kips]

180 160 140 120 100 URM wall RC Frame

800 700 600 500 400 [kN]

80 60 40 20 0
TAR 1 TAR 2 TAR 3 TAR 4 TAR 6 DUZ 7 DUZ 8 DUZ 7-2

80 60 40 20 0
TAR 1 TAR 2 TAR 3 TAR 4 TAR 6 DUZ 7 DUZ 8 DUZ 7-2

300 200 100 0

300 200 100 0

(a)

Test Levels

(b)

Test Levels

Figure 15. Effect of progression of damage on load sharing between URM infill and RC frames: (a) at peak base shear; and (b) at peak floor displacement.

by the URM infill wall compared to that carried by the three RC frames: at the peak of the total base shear (Figure 15(a)) and at the peak of the lateral floor displacement for all testing levels (Figure 15(b)). The plots confirm that (before level TAR6) the undamaged URM infill wall governs the behaviour of the test structure. As the wall experiences damage, the RC frames pick up a larger portion of the load. At level DUZ 7-2, the wall is completely disintegrated and can be considered structurally insignificant as the load is carried almost entirely by the three RC frames. The distinction between the force distributions for peak base shear and peak floor displacement is made to emphasize the different state of the test structure at these different peak points particularly for levels DUZ 7 and DUZ 8. In these levels, the point of peak base shear takes place at the time where the wall is still resisting large portion of the load at its incipient failure. However, the peak floor displacement takes place after the damage in the wall has occurred. As an example, the time histories of the shear forces in the wall and that carried by the RC frames and the corresponding floor displacements are presented in Figure 16. From this figure, the maximum base shear takes place at 8.8 s while the maximum floor displacement occurs at 17.3 s following wall cracking at 15.4 s. An important aspect of the response of the URM infill wall is the shear force versus shear deformation in the plane of the wall. The shear deformation is determined using the diagonal
Copyright q 2006 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2006; 35:1827–1852 DOI: 10.1002/eqe

SHAKE-TABLE EXPERIMENT ON REINFORCED CONCRETE STRUCTURE

1845

130 100 50 0 -50 -100 -130

URM Wall RC Frames 250 [kN] 0 -250 7 8 9 10 Maximum base shear 11 -500 12 Shear force [kips] 500 130 100 50 0 -50 -100 -130 14 Wall cracking 25 15 0 -15 -25 7 8 9 10 Time [sec] 11 12 [mm] Displacement [in] 1.25 1 0.5 0 -0.5 -1 -1.25 14

URM Wall RC Frames 500 250 0 -250 15 16 17 18 -500 19 Maximum floor displacement 25 15 0 -15 -25 15 16 17 Time [sec] 18 19 [kN] [mm]

Shear force [kips] Displacement [in]

1.25 1 0.5 0 -0.5 -1 -1.25

(a)

(b)

Figure 16. Time histories of shear forces and displacements for level DUZ 7: (a) partial time history from 7 to 12 s; and (b) partial time history from 14 to 19 s.

H

∆1

∆2

∆2 − ∆1 2

L2 + H 2 LH

L

Figure 17. Shear deformation of the URM infill wall.

measurements in the plane of the wall as illustrated in Figure 17. The shear force in the wall, FW , is estimated as discussed earlier, Figure 10. The shear force versus shear deformation plot for the URM infill wall for levels TAR 6 and DUZ 7 are shown in Figure 18. Before cracking, i.e. level TAR 6, a linear shear force versus shear deformation relationship is obtained. After cracking, i.e. level DUZ 7, there is a rapid degradation of the shear stiffness of the URM infill wall due to significant increase in the shear deformation. RC slab. The 95 mm (3 3 ) thick RC slab is supported on boundary beams from all sides and spans 4 between the bare and infilled frames. In its plane, the RC slab acts as a diaphragm distributing the inertia force to the lateral resisting elements of the test structure by both deforming in shear and in-plane bending. The inertia force is generated by the acceleration of the mass at the slab level in the test structure consisting of the tributary mass of the test structure and the added mass of the lead ingots. In the out-of-plane direction, the slab spans between the three frames. Since the aspect ratio of the slab is 2.25, the out-of-plane behaviour of the slab can be characterized as one-way action. This is confirmed by observing the gravity induced crack patterns on the slab after loading the lead ingots. These cracks run more or less parallel to the long edges of the slab. While the vertical forces on the slab hardly change during the shake-table test, the lateral inertia forces acting on the slab and the resisting reactions from the frames beneath change dramatically
Copyright q 2006 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2006; 35:1827–1852 DOI: 10.1002/eqe

1846
125 100 50 FW [kip] 0 -50 -100 -125 -8.5 -7 (a)

A. HASHEMI AND K. M. MOSALAM

-5

-3

-1 0 1

3

5

500 400 300 200 100 0 -100 -200 -300 -400 -500 7 8.5 x10-3

125 100 50 FW [kip] [kN] 0 -50 -100 -125 -8.5 -7 (b)

500 400 300 200 100 0 -100 -200 -300 -400 -500 -5 -3 -1 0 1 3 5 7 8.5 x10-3

Figure 18. Shear force versus shear deformation of the URM infill wall (thickness = 102 mm (4 )): (a) level TAR 6; and (b) level DUZ 7.

as the test structure is subjected to different levels of shaking. Assuming uniform distributions of the inertia forces along the slab width, i.e. normal to the shaking direction, the maximum shear in the slab, VS,max , can be estimated as stated in Equation (6), where Vb denotes the total inertia force at each instant of time and VA and VC are the total shear forces resisted by the two bare frames on axes A and C, Figure 1, respectively. VS,max = max{VA , (Vb /2) − VA , VC , (Vb /2) − VC } (6)

In order to investigate the change in the demand force in the slab, we consider two extreme cases. (1) Assuming that the infilled frame (frame B) is infinitely stiff compared to the two bare frames (frames A and C), then frame B attracts all the inertia forces in the slab. It is implicit in this assumption that the slab is rigid in its plane relative to the frames. In this case, VB = Vb and VA = VC = 0, and the maximum shear in the slab is Vb /2. (2) Assuming that the URM infill wall is completely disintegrated and the three frames, A, B and C, have the same lateral stiffness. In this case it is reasonable to assume that Vb is evenly distributed between the three frames and the shear force in each frame would be Vb /3. In this case, the maximum shear in the slab is Vb /3. Examining the above two extreme cases suggests that the URM infill wall increases the shear demands on the diaphragm. Figure 19(a) shows the ratio of the maximum shear in the slab to the maximum total base shear in the test structure for different levels of shaking. As expected, this ratio reduces as the infill wall is damaged during high levels of shaking. The time history results from the experiment confirm this as well. Figure 19(b) shows the time history plot of the maximum shear in the slab (absolute value) as well as the two limits of Vb /2 and Vb /3 for level DUZ 7. As discussed previously, severe damage in the URM infill wall occurred during this level at t = 15.4 s. Comparing the slab shear forces for intervals of time at the beginning (t = 7.0 s) and at the end (t = 17.0 s) of the motion shows that the shear demand in the slab changes from close to Vb /2 before the URM infill wall damage to being close to Vb /3 after the URM infill wall damage. In order to measure the slab deformations, a set of displacement transducers are installed as shown in Figure 20(a). The displacement transducers are mounted on aluminium plates secured to the top of the prestressing bars extending from the centre of each column. The transducers are arranged to form an imaginary statically determinate truss. In order to find the displacement
Copyright q 2006 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2006; 35:1827–1852 DOI: 10.1002/eqe

[kN]

SHAKE-TABLE EXPERIMENT ON REINFORCED CONCRETE STRUCTURE

1847

0.55 Slab shear to base shear ratio Case: VA=VC=0 → VS=Vb/2 0.5 B TAR 1 TAR 2 TAR 3 0.45 0.4 0.35 0.3 Case: VA=VB=VC → VS=Vb/3 TAR 4 TAR 6 Vb B DUZ 7 DUZ 8 VS VS VC VB VA

110 100 Shear force [kips] 80 60 40 20 0

Slab shear envelope 1/2 Base shear 400 1/3 Base shear Wall cracking 300 200 100 0 20 [kN] [kN]

5

10 Time [sec]

15

(a)

Test level

(b)

Figure 19. Variation of shear demand in the slab due to damage in the URM infill wall: (a) slab-shear demand variation at peak base shear; and (b) partial time histories for DUZ 7.

80
ˆ P 1

Maximum slab shear [kips]

y

ˆ u qT V

DUZ 8 DUZ 7 DUZ7 TAR 4
(1
S =

350 300 250 200

70 60 50 40
TAR 3
V

00

V

ˆ q

0)

+

6.

78

TAR 6

67

DUZ 8

30 20 10 0 0

150 100 50

x

TAR 2 TAR 1 TAR 2

1

2

3

4

5

6 x10-4

0

(a)

(b)

(c)

Figure 20. Shear deformation of the RC slab: (a) displacement transducers; (b) nodes and virtual force system in equilibrium; and (c) slab shear stiffness.

of a specific node, we invoke the principle of virtual force (PVF) by applying a virtual force ˆ P in the direction of the required displacement u and solve the determinate truss to find the corresponding virtual internal forces satisfying equilibrium, Figure 20(b). Arranging the measured ˆ deformations and the corresponding virtual internal forces into vectors V and q, respectively, u is ˆ ˆ ˆ determined from P T u = qT V with P = 1. Once the relative node displacements u are determined, the shear deformation, , of the slab can be calculated as the differential displacement along the opposite edges of the slab divided by the distance between the two measurements. It should be noted that using PVF yields more accurate results than the approximate method used for the shear deformation of the URM wall, Figure 17. The approximation is suitable for the URM infill wall but not for the RC slab where the shear deformation is an order of magnitude smaller than that of the URM wall. Figure 20(c) shows a plot of the maximum shear in the slab versus its corresponding shear deformation. Although it is observed during the experiment that with increase in the intensity of
Copyright q 2006 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2006; 35:1827–1852 DOI: 10.1002/eqe

1848

A. HASHEMI AND K. M. MOSALAM

C1

C2

B1

B2 Denotes long. reinf. yielding Denotes trans. reinf. yielding

A1

A2

Figure 21. Damage in the RC frames after level DUZ 7.

shaking, the gravity-induced cracks become more visible, Figure 20(c) suggests that there is no significant drop in the shear stiffness of the slab. The slope of the fitted line through the data points is about one-eighth of the elastic shear rigidity of the uncracked section of the slab (GA = 5783 MN (1.3 × 106 kips)) indicating that although the shear rigidity of the slab remains more or less unchanged, its value is significantly reduced due to initial gravity cracks. The deformed shapes (amplification factor = 250) plotted on the graph are obtained using PVF at times corresponding to the maximum slab shear deformations at levels TAR 2, DUZ 7 and DUZ 8. These deformations include the effects of both shear forces and bending moments in the plane of the RC slab. RC frames. As mentioned earlier, up to level TAR 6 and the initial part of DUZ 7, the stiff URM infill wall carries most of the base shear protecting the three RC frames. At incipient failure of the wall and after the major cracking occurs, there are large shear demands at the bases of the columns where there are contacts with the wall as well as large bending moment demands on the frames as the test structure undergoes larger displacements. Figure 21 shows the locations in the test structure where reinforcing bar yielding is recorded by the strain gauges during level DUZ 7 after partial failure of the wall. It can be observed that all the yielding occurred in the middle infilled frame and the bare frames are practically undamaged. Figure 21 also includes a photograph of the column–footing joint taken after level DUZ 7 and indicating the onset of damage at this joint location. As discussed before, during level DUZ 8 and the repeat of level DUZ 7, i.e. DUZ 7-2, the effect of the URM infill wall on the behaviour of the test structure diminishes and the RC frames take the significant portion of the earthquake-induced forces. The high demands on the RC frames during these levels cause some damage in the form of local spalling and cracking in the beam– column and column–footing joints. Figure 22 shows photographs of the RC frame joints taken after DUZ 7-2.

MODEL VALIDATION The results of the experiment are compared with the simulated results of the OpenSees model of the test structure to validate the prototype modelling and the established relationship between
Copyright q 2006 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2006; 35:1827–1852 DOI: 10.1002/eqe

SHAKE-TABLE EXPERIMENT ON REINFORCED CONCRETE STRUCTURE

1849

(b) (a) C1 C2 (d) (c) B1 (a) B2 (b)

A1

A2

Denotes long. reinf. yielding Denotes trans. reinf. yielding (c) (d)

Figure 22. Damage in the RC frames at the completion of level DUZ 7-2: (a) bare frame, column base; (b) bare frame, beam–column joint; (c) infilled frame, column base; and (d) infilled frame, beam–column joint.
1.25 1 0.75 0.5 0.25 0 -0.25 -0.5 -0.75 -1 -1.25 0 (a) 180 125 75 0 -75 -125 -180 Simulation base shear [kips]

Ground acceleration [g]

TAR 2 10 20 30

TAR 3 40 50

TAR 4 60 Time [sec] 70

TAR 6 80 90

DUZ 7 10 0 110 115

600 300 0 -300 -600 0 10 TAR 2 20 30 40 TAR 3 50 60 TAR 4 70 80 TAR 6 90 100 110 115

180 125 75 0 -75 -125 -180

Experiment base shear [kips]

DUZ 7 600 300 0 -300 -600 [kN]

0

10

20

30

40

50

(b)

60 Time [sec]

70

80

90

100

110 115

Figure 23. Comparison between the experimental and simulated results: (a) concatenated input ground (shake-table) motion for the simulation; and (b) time histories of base shear.
Copyright q 2006 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2006; 35:1827–1852 DOI: 10.1002/eqe

[kN]

1850
1.5 1 0.5 0 -0.5 -1 -1.5 180 125 75 0 -75 -125 -180

A. HASHEMI AND K. M. MOSALAM

Roof disp.[in]

Simulation Experiment

Base shear [kips]

30 15 0 -15 -30 100 102 104

0 [mm] 150

10

[mm] 20

30

800 600

Base shear [kips]

100

600 300 0 -300 -600 96 98 100 Time [sec] 102 104

400 200 Experiment Simulation 0 1.5

[kN ]

50

0 0 (b)

(a)

0.25 0.5 0.75 1 1.25 Maximum displacement [in]

Figure 24. Comparison between the experimental and simulated results: (a) partial time histories for DUZ 7; and (b) peak response.

the prototype substructure and the test structure as described at the beginning of the paper. The measured accelerations at the base of the test structure during levels TAR 2 thru DUZ 7 are concatenated as a single input ground motion for the simulation, Figure 23(a). For this ground shaking, the simulated results are compared with the experimental results. Time history plots of total base shears are shown in Figure 23(b) with detailed partial plots for total base shears and floor displacements (level DUZ 7) in Figure 24(a). For levels TAR 1 through DUZ 7, the relationships of peak floor displacement versus the corresponding total base shear for the experimental and simulated results are compared in Figure 24(b). Considering the simple model used with a single strut for the URM infill wall, the OpenSees simulation results are in good agreement with the experimental results. However, the computational model overestimates the peak floor displacement by 21% and underestimates the corresponding total base shear by 16% at level DUZ 7 leaving the door open for future refinement of the computational model to obtain closer agreement with the experimental results.

CONCLUDING REMARKS The URM infill wall has a significant role in the strength and ductility of the test structure and should be considered in both analysis and design. Globally, it makes the test structure stiffer by a factor of 3.8, shortens the natural period of the test structure by 50%, increases the damping coefficient depending on the level of shaking from about 4 to 5–12% and increases the dissipated energy in the system. Such changes significantly affect the level of demand forces on the structure and generally reduce the displacement demands. Locally, the URM infill wall changes the load path and the distribution of forces between different elements of the test structure by increasing the demand forces on its adjacent elements, e.g. the top and bottom of the RC columns and the RC slab. Quantitatively, the URM infill wall causes about 30% increase in the demand forces on the diaphragm and collector elements in the test structure.
Copyright q 2006 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2006; 35:1827–1852 DOI: 10.1002/eqe

[kN]

96

98

SHAKE-TABLE EXPERIMENT ON REINFORCED CONCRETE STRUCTURE

1851

The benefits and limitations of the analytical model of the test structure is discussed and verified by comparing the simulation results with the shake-table test results. Using this model, the relationship between the prototype structure and the test structure is established. The test results also serve as a basis for future modelling efforts to capture the behaviour of the infilled frame more accurately especially after initial cracking and through the transition phase. Efforts are underway to introduce models that represent the cyclic behaviour of the masonry infills both for the in-plane and the out-of-plane directions. Such models can be used to analyse the behaviour of the prototype building more reliably.

ACKNOWLEDGEMENTS

The authors gratefully acknowledge the financial support provided by the National Science Foundation under NSF Contract No. CMS0116005. The reinforcing bars for the project were generously donated by Mr T. Tietz, Western Regional Manager of the Concrete Reinforcing Steel Institute (CRSI). The help of Mr T. Elkhoraibi during the experiment is greatly appreciated. Special thanks are due to EERC laboratory personnel especially Mr W. Neighbour for his assistance.

REFERENCES 1. Moghaddam HA, Dowling PJ. The state of the art in infilled frames. ESEE Research Report No. 87-2, Civil Engineering Department, Imperial College of Science and Technology, London, 1987. 2. Mosalam KM, White RN, Gergely P. Static response of infilled frames using quasi-static experimentation. Journal of Structural Engineering (ASCE) 1997; 123(11):1462–1469. 3. Mosalam KM, White RN, Ayala G. Response of infilled frames using pseudo-dynamic experimentation. Earthquake Engineering and Structural Dynamics 1998; 27(6):589–608. 4. Fardis MN, Bousias SN, Franchioni G, Panagiotakos TB. Seismic response and design of RC structures with plan-eccentric masonry infills. Earthquake Engineering and Structural Dynamics 1999; 28(2):173–191. 5. Zarnic R, Gosti S, Crewe AJ, Taylor CA. Shaking table tests of 1:4 reduced-scale models of masonry infilled reinforced concrete frame buildings. Earthquake Engineering and Structural Dynamics 2001; 30(6):819–834. 6. Dolce M, Cardone D, Ponzo FC, Valente C. Shaking table tests on reinforced concrete frames without and with passive control systems. Earthquake Engineering and Structural Dynamics 2005; 34(14):1687–1717. 7. Elkhoraibi T, Mosalam KM. Pseudo-dynamic experiment on one-storey RC structure with and without masonry infill. Proceedings of the 100th Anniversary Earthquake Conference Commemorating the 1906 San Francisco Earthquake, 18–22 April 2006. 8. ACI Committee 318. Building Code Requirements for Structural Concrete (ACI 318-02) and Commentary (ACI 318R-02). American Concrete Institute, 2002. 9. Building Seismic Safety Council. NEHRP Recommended Provisions for Seismic Regulations for New Buildings and Other Structures, Part 1: Provisions (FEMA368) and Part 2: Commentary (FEMA369). Washington, DC, 2000. 10. McKenna F, Fenves GL. OpenSees Manual. Pacific Earthquake Engineering Research Center, 2001, http:// opensees.berkeley.edu 11. Mander JB, Priestley MJN, Park R. Theoretical stress–strain behaviour of confined concrete. Journal of Structural Engineering (ASCE) 1988; 114(8):1804–1826. 12. Mander JB, Priestley MJN, Park R. Observed stress–strain behaviour of confined concrete. Journal of Structural Engineering (ASCE) 1988; 114(8):1827–1849. 13. American Society of Civil Engineers. Prestandard and Commentary for the Seismic Rehabilitation of Buildings, (FEMA 356). Reston, Virginia, 2000. 14. Chopra AK. Dynamics of Structures: Theory and Applications to Earthquake Engineering (2nd edn). Prentice-Hall: Upper Saddle River, NJ, 2001. 15. ASTM A 615/A 615M. Standard Specification for Deformed and Plain Carbon-Steel Bars for Concrete Reinforcement. ASTM: West Conshohocken, PA, 2001. Copyright q 2006 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2006; 35:1827–1852 DOI: 10.1002/eqe

1852

A. HASHEMI AND K. M. MOSALAM

16. ASTM C 873-99. Standard Test Method for Compressive Strength of Concrete Cylinders Cast-in-Place in Cylindrical Molds. ASTM: West Conshohocken, PA, 1999. 17. ASTM C 270. Standard Specification for Mortar for Unit Masonry. ASTM: West Conshohocken, PA, 2003. 18. ASTM C 1314. Standard Test Method for Compressive Strength of Masonry Prisms. ASTM: West Conshohocken, PA, 2003. 19. ASTM E519. Standard Test Method for Diagonal Tension (Shear) in Masonry Assemblages. ASTM: West Conshohocken, PA, 2002.

Copyright q

2006 John Wiley & Sons, Ltd.

Earthquake Engng Struct. Dyn. 2006; 35:1827–1852 DOI: 10.1002/eqe

Sign up to vote on this title
UsefulNot useful