You are on page 1of 12

Experimental Evaluation of Pretopped Precast Diaphragm

Critical Flexure Joint under Seismic Demands


D. Zhang, S.M.ASCE1; R. B. Fleischman, M.ASCE2; C. J. Naito, M.ASCE3; and R. Ren, S.M.ASCE4
Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/05/14. Copyright ASCE. For personal use only; all rights reserved.

Abstract: Precast concrete diaphragm seismic response is examined in experimental research integrating model-based simulation with
physical testing. The experimental substructure is the diaphragm critical flexural region of a prototype precast parking structure. This region
is expected to undergo significant inelastic flexural deformation, while potentially nonductile regions remain elastic on the basis of capacity
design rules from an emerging design methodology. The physical test is conducted at half-scale. The test specimen is detailed using dia-
phragm reinforcement intended to meet deformability requirements. Predetermined displacement histories are applied to the test specimen on
the basis of nonlinear transient dynamic analyses of the prototype structure. The loading history is applied by a test fixture capable of
simultaneously providing shear, axial, and moment to the joint. Moment strength, stiffness, rotational deformation capacity, and progressive
damage are examined under a sequence of increasing intensity earthquakes. Design recommendations are provided. DOI: 10.1061/(ASCE)
ST.1943-541X.0000352. © 2011 American Society of Civil Engineers.
CE Database subject headings: Diaphragms; Precast concrete; Seismic design; Tests; Joints.
Author keywords: Diaphragms; Concrete; Precast; Seismic design; Tests.

Introduction experienced in the analytical superstructure during the earthquake


simulations.
The seismic performance of the high flexure region in pretopped The research is part of a larger effort to develop a new seismic
precast concrete diaphragms is examined through analysis- design methodology for precast floor diaphragms [Building
driven integrated physical testing. A 3-story precast parking garage Seismic Safety Council (BSSC) 2009]. The methodology counts
[Fig. 1(a)] is used as the prototype structure. In these tests, model- on the precast diaphragm possessing sufficient inelastic flexural de-
based simulation of an analytical superstructure, a finite-element formation capacity during seismic loading. For high seismic haz-
(FE) representation of the prototype structure [Fig. 1(b)], is used ard, this requirement is to be provided by diaphragm reinforcement
to control the loading of a physical test substructure (shaded details classified as “high deformability elements” (HDE). The spe-
region). The testing program is a multiuniversity effort, with the cific inelastic deformation targets are dependent on design factors
analytical and experimental components originating from separate currently under calibration in parallel work (Zhang 2010). In this
research institutions. The test load histories are derived from study, a prototype structure designed with trial factors is examined
earthquake simulation via nonlinear transient dynamic analyses under anticipated seismic demands, including those that extend the
(NLTDA). response into the inelastic regime, to examine the performance of
The physical test is conducted at half-scale. The test specimen is a flexure-critical pretopped precast diaphragm joint with HDE
composed of the precast units surrounding the critical flexural joint. details, including: (1) the flexural strength and stiffness under
The specimen is placed in a test fixture capable of simultaneously progressive damage; (2) the rotational deformation capacity under
providing shear, axial, and moment to the joint. The seismic load- realistic cyclic loading; and (3) the failure modes including map-
ing is applied as predetermined displacement histories (PDHs), rep- ping damage levels to seismic hazard.
resenting the interface degrees of freedom (DOF) between the
superstructure and substructure [see arrows in Fig. 1(b)]. The PDHs
subject the critical flexural joint to the sliding and opening histories Background

1
Graduate Student, Dept. of Civil Engineering and Engineering In precast diaphragms, joints between the precast floor units serve
Mechanics, Univ. of Arizona, Tucson AZ, 85712-0072. E-mail: zdc@ as critical cross sections, as they represent planes of weakness.
email.arizona.edu Deformation will tend to concentrate at these joints, particularly
2 for inelastic diaphragm action (Fleischman and Wan 2007). For
Associate Professor, Dept. of Civil Engineering and Engineering
Mechanics, Univ. of Arizona, Tucson, AZ 85721-0072 (corresponding pretopped floor systems, as in the prototype structure, diaphragm
author). E-mail: rfleisch@email. carizona.edu action occurs entirely through the precast units rather than through
3
Associate Professor, Dept. of Civil Engineering, Lehigh Univ., a topping slab as is currently required [International Building Code
Bethlehem, PA 18015-4728. E-mail: cjn3@lehigh.edu (IBC) 2003] in high seismic design categories (SDCs). The
4
Graduate Student, Dept. of Civil Engineering, Lehigh Univ., reinforcement includes [refer to Fig. 1(a))] chord connectors con-
Bethlehem, PA 18015-4728. E-mail: rur206@lehigh.edu
tinuous across the precast unit to carry in-plane flexure and shear
Note. This manuscript was submitted on July 9, 2009; approved on
November 9, 2010; published online on December 3, 2010. Discussion per- connectors embedded in the units at discrete locations to transfer
iod open until March 1, 2012; separate discussions must be submitted for in-plane shear across the joints.
individual papers. This paper is part of the Journal of Structural Engineer- The ability of this reinforcement to achieve the expected joint
ing, Vol. 137, No. 10, October 1, 2011. ©ASCE, ISSN 0733-9445/2011/ opening demands associated with a flexure mechanism is a primary
10-1063–1074/$25.00. focus of the testing. This pertains not only to chord reinforcement;

JOURNAL OF STRUCTURAL ENGINEERING © ASCE / OCTOBER 2011 / 1063

J. Struct. Eng. 2011.137:1063-1074.


Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/05/14. Copyright ASCE. For personal use only; all rights reserved.

Fig. 1. Integrated physical test: (a) prototype structure; (b) analytical model (aerial photo courtesy of Wm. Blanchard Co.)

given the significant joint lengths, up to 18 m, shear reinforcement but for this study would limit demands placed on the specimen
can also be subjected to large opening deformations and requires during testing. Full details of the structure and diaphragm design
similar tensile deformation capacity. appear in Zhang (2010).

Prototype Structure Analytical Modeling

The parking structure is designed (IBC 2003) for a Charleston SC The structure is modeled in three-dimensions [refer to Fig. 1(b)]
site, seismic design category (SDC) D, soil class F (Ss ¼ 1:39, using the general-purpose FE program Ansys. Floor units are elas-
S1 ¼ 0:4). The seismic-resistant system is special reinforced con- tic plane stress elements; precast beams are elastic beam elements.
crete shear wall (R ¼ 6) designed according to ACI-318 [American The top floor is discretized with finer mesh. Diaphragm reinforce-
Concrete Institute (ACI) 2005]. The maximum (top-level) dia- ment is discretely modeled using element subassemblages involv-
phragm force produced by current code is 2,003 kN. A diaphragm ing nonlinear coupled (tension-shear) spring with a cyclic pinching/
force amplification factor, Ψd ¼ 1:5, is applied to this value to de- degradation capabilities in parallel with contact elements. The
sign each floor according to the new methodology (BSSC 2009). properties for these elements were obtained from earlier tests of
Table 1 summarizes the required and nominal strengths (for isolated precast diaphragm connectors (Naito et al. 2007). The
full-scale and half-scale) critical-flexure (CF) joint designs. With- shear wall is modeled using elastic shell elements with shear
out Ψd , current produces M u ¼ 3;660 kN·m. The section nominal deformation capabilities to capture in-plane and out-of-plane stiff-
flexural strength is determined using an analytically based design ness. Inelastic action at the shear wall base is modeled using two
procedure (BSSC 2009), and includes the tension contribution of pairs of nonlinear springs in parallel with contact elements to
shear reinforcement, in accordance with ACI-318 10.2.2/10.6.7 capture yielding of the wall attributable to biaxial moment. Lateral
(ACI 2005). restraint of gravity system columns is ignored. Full details of the
The shear connector design is controlled by maximum spacing discrete model and the connector elements appear in (Zhang 2010).
limits from construction practice (1.83 m), resulting in five connec-
Dynamic Analysis
tors. The design strengths of individual connectors (t n , vn ) are de-
termined from previous experiments as indicated. Although The interface PDHs are derived from the prototype structure
designed on the basis of flexure alone, the critical-flexure joint response to a sequence of increasing intensity ground motions. Five
is subjected to combinations of flexure, shear, and axial during motions are selected for the PDH test sequence (Zhang 2010).
the experiment. The first three motions represent seismic hazard for Charleston,
The CF joint is explicitly designed to create the test specimen. South Carolina (CH), corresponding to service (SVC), design basis
However, all other regions/levels of the diaphragm must also be (DBE), and maximum considered (MCE) earthquakes. The fourth
designed accurately so that the analytical model produces appro- is a bidirectional Charleston DBE motion intended to subject the
priate NLTDA results. Included are overstrength factors Ωv for di- CF joint to high combined axial, flexure, and shear attributable to
aphragm reinforcement in high shear regions to prevent nonductile the effects of transverse flexure in conjunction with in-plane twist-
limit states (Fleischman and Wan 2007). In preliminary analyses, ing and collector forces from the longitudinal component. The final
the diaphragm without Ωv suffered nonductile mechanisms away motion corresponds to a Berkeley (BK) MCE level. The prototype
from the high flexure region, not only structurally undesirable, structure design, which incorporated force amplification factors

Table 1. Diaphragm Critical-Flexure Joint Design


M u (kN·m) V u (kN) N u (kN) Type Detail t n (kN)a vn (kN)b # mn (kN) M n (kN·m)c V n (kN) N n (kN)
Full 5,491 0 0 Chord #6 118 0 6 5,977 6,352 290 1,486
Shear JVI 14 58 5 375
Half 686 0 0 Chord #3 29.5 0 6 747 794 73 372
Shear JVI 3.5 14.5 5 47
a
Connector (chord and JVI, refer to Fig. 7) tension strength (Naito et al. 2007).
b
Connector (chord and JVI, refer to Fig. 7) shear strength (Naito et al. 2007).
c
Analytically based procedure (BSSC 2009).

1064 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / OCTOBER 2011

J. Struct. Eng. 2011.137:1063-1074.


from the emerging design methodology, corresponds roughly to the Analytical Results
current diaphragm design for the Berkeley SDC. Thus, this last test Key structure response is summarized in full-scale. The shear wall
is not only intended to examine “overload” conditions, but also to moment-rotation is shown in Fig. 3 for selected motions: CH DBE
make inferences on current design. shear wall rotational demands are within expected life-safety tar-
The structural model itself was scaled with outputs used directly gets per IBC (2003). Fig. 3 shows that the longitudinal wall par-
to create the PDHs, rather than analyzing a full-scale structure and ticipates in the bidirectional motion. Fig. 4 shows the critical (3rd)
scaling outputs. The ground motions are therefore modified by in- floor inertial force time histories across the earthquake motions.
creasing acceleration and compressing time by the scale factor. The design force, Ψd F px is indicated on the force plot.
Similitude was verified through comparison to scaled results from Force-deformation responses at the CF joint are shown in Fig. 5
a full-scale model. for the center shear connector and the chords. The dotted lines in-
Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/05/14. Copyright ASCE. For personal use only; all rights reserved.

The dynamic analysis is conducted using the Newmark dicate connector element monotonic backbone curves. The chord
integration method (two substeps per time step of 0:05=0:025 s connector failure deformation capacity from isolated connector
for Charleston/Berkeley compressed records). The convergence tests (Naito et al. 2007) is shown. Recall that the critical chord
technique is a modified Newton-Raphson iteration method with switches for each motion. Key observations include: the SVC re-
force tolerance error of 0.5% (ANSYS 2007). Equivalent damping sponse is elastic; the chord yields in the DBE; the shear response
was taken conservatively as 0% on the basis of recent findings on enters the inelastic regime in the MCE; and maximum chord de-
formation exceeds the deformation capacity during the BK MCE.
viscous damping present in nonlinear systems (Panagiotou et al.
2006). The final two runs required 2% equivalent damping to
converge. Test Specimen
Sequence of Loading The test specimen is composed of two half-scale pretopped precast
double-tee (DT) floor units, connected across their joint by HDE
Seismic response of parking structure diaphragms is complex
diaphragm reinforcement. Testing at half-scale allowed industry
(Fleischman et al. 1998): the leading flat is under bending and ten- partners to produce identical units for a half-scale shake table test
sion; the trailing flat under bending and compression (see Fig. 2). (Schoettler et al. 2009) and permitted the test footprint to fit on
The outer chord of the leading flat is critical (tension component of the laboratory strong-floor (at Lehigh University, Bethlehem,
bending adding to axial tension) for either vibration direction. Pennsylvania).
Thus, more test life was produced from a single specimen by suc- Key physical dimensions of the test specimen are scaled in half.
cessively alternating the experimental substructure from north to Fig. 6(a) shows the DT cross section in full-scale. The 102-mm
south, essentially flipping the critical chord from top to bottom
in the test sequence (see Table 2).
The longitudinal walls are offset about floor midspan, not un-
common in parking structures. This layout eliminates the confining
effect of the longitudinal wall on the CF joint, while also introduc-
ing longitudinal wall collector forces to this joint, thereby creating a
highly demanding condition.

Fig. 3. Shear wall base rotation

9,000
PDH1 PDH2 PDH3 PDH4 PDH5

Fig. 2. Diaphragm deformation shape: (a) north leading; (b) south 6,000
leading
Ψd F px
Force (kN)

3,000

Table 2. Test Sequence 0


Test Earthquake Intensity Direction Test panel
0 40 80 120 160 200
PDH 1 Charleston (CH) SVC Transverse South -3,000
PDH 2 Charleston (CH) DBE Transverse North
PDH 3 Charleston (CH) MCE Transverse South -6,000
Time (s)
PDH 4 Charleston (CH) DBE Bidirection (Bi) North
PDH 5 Berkeley (BK) MCE Transverse South Fig. 4. Diaphragm force at critical floor (3rd)

JOURNAL OF STRUCTURAL ENGINEERING © ASCE / OCTOBER 2011 / 1065

J. Struct. Eng. 2011.137:1063-1074.


(a) PDH1 60 PDH60
2 PDH60
3 PDH60
4 PDH60
5
30 30 30 30 30
Shear sliding Sliding (mm)
0 0 0 0 0
-4 -2 -30 0 2 4 -4 -2 -30 0 2 4 -4 -2 -30 0 2 4 -4 -2 -30 0 2 4 -4 -2 -30 0 2 4
-60 -60 -60 -60 -60 V (kN)
(b) 1200 1200 1200 1200 T1200
(kN)
600 Top Chord 600 600 600 600
Opening Opening (mm)
0 0 0 0 0
-10-600 0 10 20 30 -10-600 0 10 20 30 -10-600 0 10 20 30 -10-600 0 10 20 30 -10-600 0 10 20 30
-1200 -1200 -1200 -1200 -1200
Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/05/14. Copyright ASCE. For personal use only; all rights reserved.

(c) 1200 1200 1200 1200 T1200


(kN)
600 Bottom Chord 600 600 600 600
Opening
0 0 0 0 0
-10-600 0 10 20 30 -10-600 0 10 20 30 -10-600 0 10 20 30 -10-600 0 10 20 30 -10-600 0 10 20 30
Failure
-1200 -1200 -1200 -1200 -1200 opening

Fig. 5. Reinforcement: (a) connector shear; (b) top chord tension; (c) bottom chord

(4-in.) flange is typical of pretopped floor systems currently in use well within the stem leading to low prestress level in the flange, the
in lower seismic hazards. DT specimens were not prestressed.
Fig. 6(b) shows the half-scale DT specimen. Important actions The HDE reinforcement details are shown in full-scale and half-
occur in the plane of the DT flange, thus the specimen stem geom- scale in Fig. 7: (a) a partially unbonded dry chord connector (Naito
etry was chosen for ease of forming and load handling, rather than et al. 2007); and (b) the JVI Vector, a common DT flange-to-flange
complete verisimilitude. Likewise, as DT prestressing is typically connector (Shaikh and Feile 2004). These details exhibited HDE
behavior in isolated connector tests (Naito et al. 2007).
Connector strengths are controlled by yielding of the anchoring
102 mm elements by providing overstrength to the plates, welds, and slug.
2.44 m
197 mm The half-scale chord detail is a six #3-bar configuration. Debonding
Flange is provided through padded foam extending 3.2 mm from the back
Standard of the faceplate. Mechanical debonding increases inelastic defor-
Prestressing mation prior to bar fracture; the compressible filler material also
762 mm chamfer
strands significantly reduces high stresses associated with dowel action.
Stem The scaled JVI Vector connector is a specially made piece,
121 mm developed and fabricated specifically for the testing program by
(a) JVI Inc. (Lincolnwood, Illinois). Tests of individual half-scale
connectors (Natio and Ren 2008) indicated good similitude to
51 mm full-scale tests.
1.22 m Flange
330 mm Standard Production of Specimens
2 #4
chamfer The specimen was constructed in accordance with current practice
204 mm
at an industry partner manufacturing facility (High Concrete Inc.,
(b) Denver, Pennsylvania). A standard single-tee form was modified to
create the half-scale cross section. The DTs were poured using self-
Fig. 6. DT section: (a) full-scale; (b) half-scale consolidating concrete with nominal strength of 48 MPa. Actual

JVI leg
FULL (2) #5 Gr.60 HALF (6) #3 Gr.60
HALF

E7018
14mm FULL
Faceplate PL
E7018
216x51x9.6(mm) 6.4mm 102mm 6.4 3.2 51mm
Unbonded 423mm 211mm
102mm length 51mm
6.4mm filler 3.2mm
324x25x4.8(mm)

Slug 216x38x9.5(mm) 8 216mm E7018


324mm 4.8 324x13x9.5(mm)
Slug PL 85x25.4x9.5(mm) Slug 45x9.5x4.8(mm)
(a) (b)

Fig. 7. Full- and half-scale connectors: (a) unbonded chord; (b) JVI Vector

1066 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / OCTOBER 2011

J. Struct. Eng. 2011.137:1063-1074.


concrete compressive strength, measured from cylinder tests Experimental Setup
conducted according to ASTM C39 (ASTM 2011b), averaged
54  0:7 MPa. The specimen is placed in a multi-degree-of-freedom (MDOF)
The panels include: (1) mild steel reinforcement, #3 and #4 re- loading fixture developed specially for the test program (see Fig. 9).
bar, of ASTM A706 (ASTM 2009) grade 60 steel; and (2) welded The fixture provides simultaneous control of three DOFs in the
wire reinforcement (WWR) conforming to ASTM A185 (ASTM plane of the specimen through three 1,250-kN actuators (ACT)
2007). Mill certified yield and tensile strength F Y , F U are 490, attached to a movable support beam: ACT1 and ACT2 control dis-
732 MPa (#3 bars), and 475, 759 MPa (#4 bars). Bar ultimate elon- placement perpendicular to the joint at specimen top and bottom
gation was measured in tests as approximately 0.20. The WWR (producing opening/closing of the joint); ACT3 controls displace-
possessed a measured tensile strength of 725 MPa and ultimate ment parallel to the joint (producing sliding displacement). Inde-
strain capacity of 0.03. Embedded connector plates and slugs were pendent control of these actuators allows for application of the
Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/05/14. Copyright ASCE. For personal use only; all rights reserved.

fabricated from A36 steel (ASTM 2011a). All welds were con- interface PDHs determined in the analytical simulations, thereby
ducted using the shielded metal arc welding (SMAW) process using subjecting the specimen to nonproportional combinations of in-
E7018 electrodes in accordance with American Welding Society plane shear, axial, and bending actions.
(AWS) (2004) standards. The test specimen is connected to a fixed support beam on one
face and a movable support beam on the other (W30 × 326 steel sec-
Specimen Detail
tions). The movable support rests on a pair of polytetrafluorethylene-
The specimen reinforcement layout is shown in Fig. 8. In addition coated steel plate sliders, providing in-plane motion with minimum
to primary reinforcement (see Table 1), each panel contains friction. Similar sliders under the panel stems eliminate specimen
WWR for temperature/shrinkage reinforcement (ACI 2005). Con- sag attributable to self-weight in the absence of prestressing. The
ventional reinforcing bars (two #4 bars at 44.5 mm) are placed at connection between panel and support consists of: (1) grouted
the bottom of the stem in lieu of prestressing [see Fig. 8(b)]. Anchor through-rods at 610 mm (seven per beam) attached to framing angles
holes, reinforced with L-shaped #4 bars, are provided at the fixture bolted to the fixture beam; (2) welding of the chord connector
supports. back-end face plate to the fixture beam (shown in Fig. 10).

4.88 m (16')
610 mm 5 connectors @ 914 mm= 3.66 m 610 mm
229 mm JOINT FACE 67 mm
UP

UP

UP

UP

UP
FOR

FOR

FOR

FOR

FOR
Chord connector
610 mm JVI Vector
1.22 m
(4' ) Stirrup hooks
610 mm 28.6 mm dia. holes

140 mm SUPPORT FACE WWR 305x915-MD83xMD29(mm)


5 spaces @ 54=270 mm 7 Anchor holes @ 610 mm =3.66 m
(a)

4.88 m (16' )
51 mm 2 # 4 bars Stem Flange 330 mm 381mm

(b)

Fig. 8. Specimen plan view and side elevation (one panel of two)

7.62 m
Fixed Fixed Movable Fixed
support support support Framing angles support
4.88 m
ACT1 A
ACT2
Teflon sliders
(b) A-A
1.22 m

5.17 m

Fixed support
Framing
BOTTOM

Specimen panel A angles


TOP

Specimen panel B
ACT3 Fixed Support
Movable support
1.37 m 4.88 m 1.37 m 5.17 m
A
(a) (c)

Fig. 9. MDOF loading fixture: (a) plan view; (b) section; (c) photo

JOURNAL OF STRUCTURAL ENGINEERING © ASCE / OCTOBER 2011 / 1067

J. Struct. Eng. 2011.137:1063-1074.


Joint opening across each connector (D1–D7); sliding along the
joint at three equal gages of 1.52 m (D8–D10); total panel defor-
mation between supports at chord (C5, C1) and panel centerline
(C8); movement of the panel relative to the supports at each chord
(C2, C4, C6, and C7). Actuator displacements (Δ1 , Δ2 , and Δ3 )
and restoring forces (F 1 , F 2 , and F 3 ) are measured by internal
LVDTs and external load cells, respectively.

Loading Protocol
Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/05/14. Copyright ASCE. For personal use only; all rights reserved.

The interface PDHs are applied to the specimen using the same
time discretization as the NLTDA. Because of the test specimen
high elastic stiffness, displacement commands associated with
Fig. 10. Overhead photograph of PDH test low force levels are of the same order of magnitude as play in
the actuator clevis. For this reason, actuator displacement com-
Specimen Installation mands (Δ1 , Δ2 , and Δ3 ) were controlled through a multiple loop
architecture using external LVDTs C5, C1, and D9. Each outer loop
The panels were installed as follows: (1) panel A is placed in the displacement step (one PDH step) is divided into smaller substeps
fixture and welded to the fixed support; (2) through-bolts are in- (actuator command increments of 0.152 mm; changed to 0.051 mm
serted into the anchor holes, fully tightened, and grouted; (3) panel B after PDH2) at approximately the actuator resolution (0.102 mm).
is attached to the movable support in an identical fashion; (4) the In the inner loop, each substep is applied through actuator displace-
actuators are attached to the movable support and brought online ment control until displacement targets are achieved at C5, C1,
(hydraulic pressure is then maintained until completion of testing); and D9. The substeps are repeated, with actuators not achieving
and (5) the panels are welded to each other. their target within a tolerance (originally 0.051 mm; changed to
The welding procedure follows typical field construction 0.076 mm after PDH1) slightly extended or retracted, until the
practice (see Fig. 10): (1) a steel rectangular slug is tack welded outer loop full step is achieved on all feedback channels.
between connectors, moving from top to bottom; and (2) the se-
quence is repeated with the appropriate size fillet weld as specified PDH Transformations
in the detail drawings (field weld shown in Fig. 7). Force and displacement components are shown in Fig. 12 for:
(a) the FE model; (b) the test specimen; and (c) at the joint.
Instrumentation The FE model interface DOFs define the PDHs (equivalent to
Seventeen relative displacements are measured on the specimen us- LVDT test control DOFs) using Eq. (1); likewise, the PDHs are
ing linear variable displacement transducers (LVDTs) (see Fig. 11): transformed into actuator command DOFs [Eq. (2)] and calculated
global displacements [Eq. (3)]:
LVDT C5 ¼ PDHt ¼ U x2  U x1 ð1a Þ
Stem 229 mm (typ.) Fixed support
Anchor 406 mm (typ.)
C4 C7 C1 ¼ PDHb ¼ U x4  U x3 ð1b Þ
C5 C8 C1
305 mm
51mm D2
D8 D9 D10 76 mm (typ.) D9 ¼ PDHv ¼ U y6  U y5 ð1c Þ
U PFOR

U
P

P
FOR

FOR

FOR

FOR

152 mm D1 D3
F1, ∆1 D4 D5 D6 D7 610 mm F2, ∆2
610mm 5 LVDT @ 914 mm=3.66 m
C2
82 mm (typ.)
C6 Δ1 ¼ C5 þ θ · ðc þ d 1 Þ ð2a Þ
F3, ∆3
BOT.
TOP

Movable support
Δ2 ¼ C1  θ · ðc þ d 1 Þ ð2b Þ

Fig. 11. Instrumentation layout Δ3 ¼ D9  θ · ðb þ eÞ ð2c Þ

y
Uy ∆1 , F1 F1
c ∆v ∆a
Ux b e
d1 Ttop
1 Ttop 2 LVDT C5 yF1
T j1
LVDT D9

Vmid V T j2 yt
y1
d a T j3 y2
5 6 N y3
M T j4
T j5 y4 yb
LVDT C1 y5
3 Tbot 4 Tbot yF2
d1
θ
c
∆2 , F2 F2
∆3 , F3 F3
(a) (b) (c) (d)

Fig. 12. DOF and force transformations: (a) FE model; (b) test specimen; (c) deformed shape; (d) free body diagram

1068 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / OCTOBER 2011

J. Struct. Eng. 2011.137:1063-1074.


θ ¼ ðC5  C1Þ=d ð3a Þ derived PDHs. Three differences are noted: (1) because of the
time-intensive nature of the testing control convergence protocol,
Δa ¼ ðC5 þ C1Þ=2 ð3b Þ the test input data (black lines) were truncated to include only the
major response regions of each earthquake motion; (2) control
Δv ¼ D9 ð3c Þ of sliding displacements in early tests was poor, resulting in
signal error, and resolved as discussed in “Loading Protocol” [see
Similarly, the actuator forces are transformed to panel forces Fig. 13(c)]; (3) during the Charleston MCE (roughly midpoint of
[Eq. (4)] and internal forces [Eq. (5)] on the basis of Fig. 12(d): testing), a large shear excursion occurred [see Fig. 13(c), PDH3].
M ¼ ðF 1  F 2 Þðc þ d 1 þ d=2Þ  F 3 ðb þ eÞ ð4a Þ Though unplanned, it permits the examination of the effect of shear
degradation on flexural response. The failure event, weld fracture
on the chord, is noted as the abrupt termination of the test curve
Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/05/14. Copyright ASCE. For personal use only; all rights reserved.

N ¼ F1 þ F2 ð4b Þ
in PDH5.
V ¼ F3 ð4c Þ

Experimental Results
X
5 X
2
T top yt þ T ji yi þ T bot yb ¼ F i yFi  F 3 ðb þ eÞ ð5a Þ Results are presented as tested (half-scale). Thus, full-scale defor-
i¼1 i¼1
mation capacity, force, and moment strength are larger by the value,
square, and cubic of the scale factor, respectively. For reference,
X
5 X
2
T top þ T ji þ T bot ¼ Fi ð5b Þ analytical predictions using the FE model of Fig. 1(b) are shown
i¼1 i¼1 with test results. Calibration of the FE model, on the basis of this
testing, is reported elsewhere (Zhang 2010).
V mid ¼ F 3 ð5c Þ
Force Response Histories
where
Fig. 14 shows the measured force response histories [Eq. (4)].
T ji ¼ αt;c ðyi =yt;b ÞT top;bot ð6Þ Nominal design strengths are indicated. Though individual internal
force components rarely reach design strength, their combined ef-
and α = chord-to-shear connector stiffness ratio in tension fects consistently yield the diaphragm reinforcement, as shown in
(αt ¼ 19:1) or compression (αc ¼ 5:2), as determined from testing subsequent figures. Note the effect of control error on shear force
(Natio and Ren 2008). accuracy in the early tests and the lower shear forces after the unan-
ticipated high shear load in the subsequent tests.
PDH Test Input
Connector Response
The PDH test is conducted in quasi-static fashion (displacement
command speed of 0:0254 mm=s). The actual test control histories The chord connector hysteretic response [Eq. (5) versus Eq. (1)] is
[Eq. (1)] are shown in Fig. 13 superimposed on the analytically shown in Fig. 15. For the SVC earthquake, the forces are suffi-
ciently low so that play in the actuator clevis leads to a “squaring”
of the hysteresis. The progression of damage in the chords is as
follows [See Fig. 15(a)]: (1) hairline cracks appear in the chord
region during essentially elastic response at SVC level; (2) the
top chord forms moderate cracks in the DBE, where chord yield
strength is reached but no large inelastic deformation demand is
incurred; (3) major cracking is evident in the bottom chord in
the MCE where the chord incurs significant inelastic opening de-
formation; (4) major cracking/crushing is exhibited in the top chord
for the Bi DBE, in which inelastic opening occurs in combination
with significant compression cycles; and finally, (5) the slug of the
bottom chord connector fractures at a maximum deformation
capacity of 10 mm.
The hysteretic response and damage development for JVI Vector
j5 [see Fig. 12(d)] is shown in Fig. 16. Cracking begins during the
DBE; the surrounding concrete starts spalling during the MCE after
4-mm shear displacement. The connector fractures under tension in
the presence of significant shear (50% of V n ) during the BK MCE
at approximately a 9-mm opening.

Joint Moment-Rotation
The joint moment-rotation response is shown in Fig. 17. Fig. 17(e)
indicates a maximum rotation capacity for the joint of approxi-
mately 0.0025 rad.

Joint Opening Profiles


Fig. 18(a) shows the distribution of tension deformation along the
Fig. 13. PDHs: (a) opening at top; (b) opening at bottom; (c) sliding
total length of the chord during the CH MCE. The response shown

JOURNAL OF STRUCTURAL ENGINEERING © ASCE / OCTOBER 2011 / 1069

J. Struct. Eng. 2011.137:1063-1074.


Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/05/14. Copyright ASCE. For personal use only; all rights reserved.

Fig. 14. Joint force: (a) axial; (b) shear; (c) moment

Fig. 15. Chord responses: (a) top photo; (b) top response; (c) bottom response; (d) bottom photo

1070 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / OCTOBER 2011

J. Struct. Eng. 2011.137:1063-1074.


Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/05/14. Copyright ASCE. For personal use only; all rights reserved.

Fig. 16. JVI response: (a) opening; (b) sliding; (c) photo

on either side of the first major opening excursion (indicated) is


typical for response prior to and following that point during the
entire sequence. First, the slip on the back end of the specimen
is negligible. Prior to the excursion, the experimental substructure
opening is accommodated almost entirely within the panel (strain
penetration attributable to panel cracking) as inferred by the differ-
ence between C1 and the sum of D2, C6, and C7. Following the
excursion, opening does concentrate at the joint, but the panel
contribution is still significant. Finally, the joint does not close
at the chord following the major opening excursion. Fig. 18(b)
shows the cumulative chord opening through the test sequence.
Fig. 19 shows the profile along the specimen (at maximum
opening) for (a) total opening across the panels; and (b) joint open-
ing. Although plane sections are enforced for the total opening, the
joint itself forms a bowed shape. This distribution occurs because
the deformation demand at the chord is distributed (via cracking)
along the entire panel width, whereas the JVI Vector connectors’
inherent tension flexibility allows relaxation of the concrete panels
[see Fig. 19(c)].

Stiffness Degradation
Stiffness degradation through the test sequence is shown in Fig. 20
Fig. 17. Moment versus joint rotation: (a) SVC; (b) DBE; (c) MCE;
including (a) chord tension (K t ); (b) chord compression (K c ); (c) ro-
(d) Bi DBE; (e) BK MCE
tation (K θ ); and (d) shear (K v ). As seen, the tension, compression,

JOURNAL OF STRUCTURAL ENGINEERING © ASCE / OCTOBER 2011 / 1071

J. Struct. Eng. 2011.137:1063-1074.


Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/05/14. Copyright ASCE. For personal use only; all rights reserved.

Fig. 18. Chord opening: (a) distribution for CH MCE; (b) cumulative

Top yc/(d/2) yc/(d/2)


1 CH SVC 1
Flexible
yc 0.5 CH DBE 0.5 chord
CH MCE bars Flexible
d 0 CH Bi DBE 0 JVI
BK MCE
-0.5 -0.5 Stiff
chord
-1 -1 Slug
Bottom Opening (mm) Opening (mm)
-5 0 5 10 15 -5 0 5 10
(a) (b) (c)

Fig. 19. Opening profile at maximum: (a) total; (b) joint; (c) deformed shape

(a) 150 (b) 500


Kt Kc
400
Kc(kN/mm)
Kt(kN/mm)

100 Test-top 300


Test-bottom 200
50
Initial 100
0 0
(c) 150,000 1st major (d) 90
opening Shear Shear
Kθ(kN·m/rad)

Kv(kN/mm)

overload overload
100,000 70
Kθ Test Kv
50,000 50 1st major
opening
0 30
PDH1 PDH2 PDH3 PDH4 PDH5 PDH1 PDH2 PDH3 PDH4 PDH5

Fig. 20. Stiffness degradation: (a) chord tension; (b) chord compression; (c) rotation; (d) shear

(a) 1 Top (b)


1
0.5 0.5
250
0 0
-0.5 -0.5
150
-1 -1
Bottom 600 1,600 2,600 step
50 1,000 1,500 2,000 2,500
N (kN)

(c) 1 (d) 1
0 50 100 150
-50 0.5 0.5 Compression
0 0 No axial
-150 Tension&yield
Before shear overload -0.5 -0.5
After shear overload -1 -1
-250 500 1,500 2,500 3,500 600 1,600 2,600
Kv (kN/mm)
Fig. 22. Neutral axis: (a) CH CVS; (b) CH DBE; (c) CH MCE;
Fig. 21. Shear stiffness under axial forces (d) Avg. of CH DBE

1072 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / OCTOBER 2011

J. Struct. Eng. 2011.137:1063-1074.


and rotation stiffness degrades as the specimen accumulates d, d 1 = distance from the center of top chord to bottom
damage. A close-up of K θ near the shear overload excursion indi- chord and to edge of plane;
cates that the major chord tension stiffness reduction is attributable e = half of height of movable support steel beam;
to the first major opening excursion [see Fig. 20(c)], whereas the F px = diaphragm design force;
shear overload event only affects the shear stiffness [see Fig. 20(d)]. F Y , F U = steel yield and ultimate strength;
Fig. 21 shows the effect of axial forces on specimen shear stiff- F 1 , F 2 , F 3 = actuator forces;
ness, indicating a significant effect with significantly higher values K t , K c , K θ , = chord tension, chord compression, joint rotation,
under axial compression and slightly lower values under axial ten- K v and joint shear stiffness;
sion. Fig. 22 shows a scatter plot of the neutral axis location during M, V, N = joint moment, shear, and axial force;
the testing as calculated from LVDTs D1-D5 and plotted for peak M n , V n , N n = diaphragm moment, shear, and axial force
strength;
Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/05/14. Copyright ASCE. For personal use only; all rights reserved.

moment.
mn , vn , t n , = connector moment, tensile, and shear strength;
M u , V u , N u = diaphragm design moment, shear, and axial force;
Conclusions PDHt , PDHb , PDHv = Interface predetermined displacement
histories at top, bottom, and sliding;
The critical-flexure joint in a pretopped concrete diaphragm was R = response modification coefficient;
tested at half-scale under displacement histories from NLDTA. S1 Ss = mapped spectral acceleration for 1-s period, short
The testing provided the following information: period;
1. The initial stiffness and strength of the joint in flexure T j = axial force of JVI;
and shear; T top , T bot = axial force of chord at top and bottom of panel;
2. The manner in which these values degrade during seismic U x , U y = x and y direction displacement;
response; V mid , V j = shear force of joint and JVI;
3. The cyclic rotational capacity of the joint; yi , yc = distance to neutral axis, to centriod;
4. The failure modes, including mapping of damage to seismic α = chord to shear connector stiffness ratio;
hazard level; Δ1 , Δ2 , Δ3 = actuator displacements;
5. The behavior under combined M-V-N loading; and Δa = average joint opening;
6. The deformation profile, both across the chord region and Δv = joint sliding;
along the joint. θ = rotation of joint;
The following design conclusions are made: Ψd = diaphragm force amplification factor; and
1. Using the new design methodology, the critical flexural joint Ωv = diaphragm shear overstrength factor.
survived its designated MCE earthquake. Under current de-
sign, nominally equivalent to the BK site, the joint will likely
References
fail during its designated MCE earthquake.
2. The unbonded dry chord connector exhibited good inelastic American Concrete Institute (ACI). (2005). “Building code requirements
deformation capacity corresponding to a joint rotation of for structural concrete.” ACI 318-05 and Commentary, Farmington
0.0025 rad (20-mm opening in half-scale). Hills, MI.
3. The JVI Vector connector exhibited good inelastic tension ASTM. (2007). “Standard specification for steel welded wire reinforce-
compliance, achieving a 9-mm opening (half-scale) before ment, plain, for concrete.” A185-07, West Conshohocken, PA.
pulling out from the surrounding concrete. ASTM. (2009). “Standard specification for low-alloy steel deformed and
4. The joint rotational stiffness degrades to approximately half its plain bars for concrete reinforcement.” A706-09, West Conshohocken,
PA.
original value under expected rotation targets.
ASTM. (2011a). “Standard specification for general requirements for rolled
5. The joint shear strength and stiffness significantly degraded structural steel bars, plates, shapes, and sheet piling.” A6-11, West
after a large sliding deformation (ductility ∼4) but maintained Conshohocken, PA.
a reduced load carrying capacity. ASTM. (2011b). “Standard test method for compressive strength of cylin-
drical concrete specimens.” C39-11, West Conshohocken, PA.
American Welding Society (AWS). (2004). AWS D1.1-2004 Structural
Acknowledgments Welding Code—Steel, 19th Ed., Miami.
Ansys 11 [Computer software]. Ansys, Canonsburg, PA.
This research was supported by the Precast/Prestressed Concrete Ansys. (2007). Elements (00853) and theory reference (00855), SAS IP,
Institute (PCI), the Charles Pankow Foundation, and the National Inc., Canonsburg, PA.
Building Seismic Safety Council (BSSC). (2009). “Seismic design meth-
Science Foundation (NSF) under Grant CMS-0324522. The
odology for precast concrete floor diaphragms.” Part III, 2009 NEHRP
authors are grateful for this support. Any opinions, findings, and Recommended Seismic Provisions, Federal Emergency Management
conclusions or recommendations expressed in this material are Agency, Washington, DC.
those of the authors and do not necessarily reflect the views of Fleischman, R. B., Sause, R., Pessiki, S., and Rhodes, A. B. (1998).
the National Science Foundation. The contributions of NEES@ “Seismic behavior of precast parking structure diaphragms.” PCI J.,
Lehigh staff, in particular Tommy Marullo, are noted. 43(1), 38–53.
Fleischman, R. B., and Wan, G. (2007). “Appropriate overstrength of shear
reinforcement in precast concrete diaphragms.” J. Struct. Eng., 133(11),
Notation 1616–1626.
International Building Code (IBC). (2003). International building code,
2003 Ed., International Code Council, Inc., Falls Church, VA.
The following symbols are used in this paper: Naito, C., and Ren, R. (2008). “Development of a seismic design
b = width of panel; methodology for precast diaphragms—Phase 1c half-scale.” ATLSS
c, a = distance from center of axial actuator to chord, Rep. #08-09, Advanced Technology for Large Structural Systems
between center of axial actuators; (ALTSS) Center, Lehigh Univ., Lehigh, PA.

JOURNAL OF STRUCTURAL ENGINEERING © ASCE / OCTOBER 2011 / 1073

J. Struct. Eng. 2011.137:1063-1074.


Naito, C. J., Ren, R., Jones, C., and Cullen, T. (2007). “Development Schoettler, M. J., Belleri, A., Zhang, D., Restrepo, J., and Fleischman, R. B.
of a seismic design methodology for precast diaphragms—Phase 1b (2009). “Preliminary results of the shake-table testing for development
summary report.” ATLSS Rep. 07-04, ATLSS Center, Lehigh Univ., of a diaphragm seismic design methology.” PCI J., 54(1), 100–124.
Lehigh, PA. Shaikh, A. F., and Feile, E. P. (2004). “Load testing of a precast concrete
Panagiotou, M., Restrepo, J., Conte, J. P., and Englekirk, R. E. (2006). double-tee flange connector.” PCI J., 49(3), 84–95.
“Shake table response of a full scale reinforced concrete wall building Zhang, D. (2010). “Examination of precast concrete diaphragm seismic re-
slice.” Structural Engineering Association of California Convention, sponse by three-dimensional nonlinear dynamic transient analyses.”
Sacramento, CA, 285–300. Ph.D. thesis. Univ. of Arizona, Tucson, AZ.
Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/05/14. Copyright ASCE. For personal use only; all rights reserved.

1074 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / OCTOBER 2011

J. Struct. Eng. 2011.137:1063-1074.

You might also like