CONNECTIONS BETWEEN PRECAST E L E M E N T S -

FAILURE WITHIN CONNECTION R E G I O N

Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/09/14. Copyright ASCE. For personal use only; all rights reserved.

By Catherine Wolfgram French, 1 Member, A S C E , Michael Hafner, 2

Associate Member, ASCE, and Viswanath Jayashankar, 3
Associate Member, ASCE

ABSTRACT: This paper presents the test results of three moment-resisting con-
nections between precast beam-column elements designed for regions of moderate
to high seismicity. The specimens are subjected to cyclic lateral load reversals that
are progressively amplified until the structures are taken to inter-story drifts (ratio
of lateral story displacement to story height) exceeding 4%. Two of the test spec-
imens represent exterior connections (BME, BMF), and the third represents an
interior connection (BMG). The connections include: BME—bolted, BMF—threaded
rebar with ordinary couplers, and BMG—threaded rebar with tapered-threaded
couplers. The plastic hinge is forced to develop within the connection region in
all three specimens. The specimens exibit ductility and energy dissipation char-
acteristics similar to those of ordinary reinforced concrete elements, as the con-
nection regions are designed to emulate ordinary reinforced concrete in a precast
system. The threaded rebar connections with tapered-threaded splices represent the
most favorable solution in terms of performance, fabrication, and economy.

INTRODUCTION

An investigation of the behavior of seven moment-resisting connections

between precast beam-column elements has been conducted at the University
of Minnesota. The connections chosen comprised posttensioned (BMA),
threaded rebar (BMB, BMF, BMG), composite (BMC), welded (BMD), and
bolted (BME) details. Three of the connections tested, BMA, BMC, and
BMD, were similar to details described in the Prestressed Concrete Institute
report "Connections for Precast Concrete Buildings Including Earthquake
Resistance" (Martin and Korkosz 1982). This report contains a series of
connection details that were rated by engineers and fabricators on the basis
of frequency of usage and quality of solution. The threaded rebar connec-
tions (BMB, BMF, and BMG) were selected by the writers to emulate the
performance of ordinary reinforced concrete in a precast system, and the
bolted connection detail was adapted from the steel industry.
Six of the specimens tested, B M A - B M F , represented exterior beam-col-
umn subassemblages; the seventh, BMG, represented an interior subassem-
blage. The models comprised a precast reinforced concrete column (14 x
14 in.) and a precast partially prestressed beam (10 x 19 in.). The primary
variable in the tests was the connection between the beam and column. To
facilitate comparison of test results, the specimens were subjected to the
same load history.
'Asst. Prof., Civ. and Min. Engrg. Dept., Univ. of Minnesota, Minneapolis, MN
55455.
'Struct. Engr., SOM, 33 W. Monroe, Chicago, IL 60603.
3
Struct. Engr., Railway Systems Inc., 464 S. Old Middletown Road, Media, PA
19063.
Note. Discussion open until May 1, 1990. To extend the closing date one month,
a written request must be filed with the ASCE Manager of Journals. The manuscript
for this paper was submitted for review and possible publication on September 16,
1988. This paper is part of the Journal of Structural Engineering, Vol. 115, No.
12, December, 1989. ©ASCE, ISSN 0733-9445/89/0012-3171/S1.00 + $.15 per
page. Paper No. 24192.

3171

J. Struct. Eng. 1989.115:3171-3192.

 The first four exterior connections, BMA-BMD, were designed to de-
velop a plastic hinge in the partially prestressed portion of the beam outside
of the connection region. The results of these tests are presented elsewhere
Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/09/14. Copyright ASCE. For personal use only; all rights reserved.

(French et al. 1989). Structures BME-BMG were designed to develop plas-

tic hinges in the connection region; the results of these tests (BME-BMG)
are presented in this paper.

DESCRIPTION OF CONNECTIONS

The cross-sectional details and moment capacities of the connections are

shown in Figs. 1-7. The provided (solid line) and induced (dashed line)
moment diagrams are included in the figures. The induced moment results
from the load applied at the end of the beam; it increases linearly from the
point of load application to the column face. The provided ultimate moment
capacity of the cross section was calculated using measured material prop-
erties (Table 1) with the Hognestad (1951) and Thompson and Park (1978)
stress-strain assumptions to model the concrete and reinforcement, respec-
tively. The ultimate capacity was defined by a concrete compressive strain
of 0.003 (Mattock et al. 1961). All of the longitudinal reinforcement in the
cross section was included in the computations. (The increased flexural ca-
pacities provided by the welded and bolted details at the face of the beams
in BMD and BME are not shown in the figures.) The intersection of the
induced and provided moment capacities represents the anticipated plastic
hinge locations in the structures.
The first four connections tested, BMA-BMD (Amu and French 1985;
Tarzikhan and French 1987), will be described here briefly. These connec-
tions were designed to develop a plastic hinge outside of the connection
region at a distance of approximately 35 in. from the column face. This

—10-—
t
3' Two 1-3/8'
Poat-Tenaioning Roda

9' Corrugated
Duct
i_ ®
3"
1

=3y
tfH_

3D' 35' 60' 90' 113' 120'

1
100 —- - — — " J ^ ^ "
200 [32.8 k - f t
Moment Induced
300

(k-ftl'

FIG. 1. Cross-Sectional Details and Moment Capacities: BMA (Posttensioned)

3172

J. Struct. Eng. 1989.115:3171-3192.

Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/09/14. Copyright ASCE. For personal use only; all rights reserved.

Four No. 9
Threaded Rebar

90' 113" 120"

(Yield Moment)
(170 k-ft) 200
300 Moment Induced

(k-ft) Moment Capacity

FIG. 2. Cross-Sectional Details and Moment Capacities: BMB (Threaded Rebar)

distance was chosen because it was the location at which the strands became
capable of developing ultimate capacity without slippage (35 in. is the trans-
fer length of 1/4-in. diameter prestressing strand).
Two 1-3/8 in. diameter (Grade 150) posttensioning rods were used for
the connection in BMA (Fig. 1). BMB contained four No. 9 (Grade 60)

three No. 9
Threeded Rebsr

Ouct (through column)

1-3/B/ Poat-
Tanaioning Rod
Caat-in-Place Concrete
Corrugeted Ouct

Cast-in-Plsce Concrete
^sss^y^syXss^sssssss'ss^s^^^SSI
k\\\\\\^\\\\\\\\\\\\\\\\^^^^
E_

100

200

300 Upward Loading (116.1 k - f t )

Downward Loading (198.6 k - f t )
(k-ft)
Moment Induced
Moment CBpacity

FIG. 3. Cross-Sectional Details and Moment Capacities: BMC (Composite)

3173

J. Struct. Eng. 1989.115:3171-3192.

Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/09/14. Copyright ASCE. For personal use only; all rights reserved.

(Strand pattern
not shown
for clarity
see Fig. 2)

Moment Induced
Moment Capacity

FIG. 4. Cross-Sectional Details and Moment Capacities: BMD (Welded)

Four No. 3
Rebar ( Corners )

I -- /" <| Strands debonded

I ,' .' A this
in < region.

2 6 ' ^ 30' BO" 90' 113'1

100

200
_^L— 134.9 """""-
k-ft Moment InducEd
300
l Moment Capacity
(k-ft)

FIG. 5. Cross-Sectional Details and Moment Capacities: BME (Bolted)

3174

J. Struct. Eng. 1989.115:3171-3192.

 1 [——10 -**\
2.5*' '""*
Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/09/14. Copyright ASCE. For personal use only; all rights reserved.

Four No, 7
Threaded Rebar

3ET
Ordinary •
Threaded
Coupler 3EL

Moment Induced

Moment Capacity

FIG. 6. Cross-Sectional Details and Moment Capacities: BMF (Threaded Rebar)

threaded rebar (Fig. 2). Note that the term "threaded rebar" is used to denote
reinforcement for which the threads are hot-rolled on the bars during fab-
rication. BMB was designed to initiate a plastic hinge immediately adjacent
to the column face. After the reinforcement underwent strain hardening at
that location, the primary hinge was relocated outside of the connection re-
gion. BMC (Fig. 3) had an unsymmetrical beam cross section. The top half
of the beam comprised cast-in-place concrete with three No. 9 (Grade 60)
rebar. For the case of top of the beam in tension, a hinge was to develop
at 24 in. from the column face. The precast prestressed bottom half of the
beam, connected to the column with a 1-3/8 in. diameter (Grade 150) post-

<E70

120'113' 90' 90* 113' 120'

ttoaent Induced

Moment Capacity

FIG. 7. Cross-Sectional Details and Moment Capacities: BMG (Threaded Rebar)

3175

J. Struct. Eng. 1989.115:3171-3192.

 TABLE 1. Properties of Reinforcement
Bar Diameter Area Yield Ultimate
Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/09/14. Copyright ASCE. For personal use only; all rights reserved.

size (in.) (sq in.) Grade strength (ksi) strength (ksi) «*" £„"
d) (2) (3) (4) (5) (6) (7) (8)
No. 3 0.375 0.11 60 73.1 114.5 0.005 0.09
No. 5 0.625 0.31 40 56.4 81.9 0.005 0.09
No. 6 0.75 0.44 60 68.0 110.0 0.005 0.09
No. 7 0.875 0.60 60 65.8 103.7 0.005 0.28
No. 10 1.27 1.27 60 70.6 114.3 0.005 0.09
1/4 0.20 0.036 250 220.0 250.0 0.008 0.05
"es/, = strain at onset of strain hardening.
b
e„ = strain at ultimate.
Note: Modulus of elasticity = 29,000 ksi (27,500 ksi for 1/4 in. diameter strand); 1
in. = 25.4 mm; 1 ksi = 6.9 MPa.

tensioning rod as in BMA, was designed to develop a hinge at 35 in. from

the column face, as with the other models. Structure BMD (Fig. 4) com-
prised a welded assembly of plates and weldable Grade 40 bars.
Structures BME-BMG, which developed primary plastic hinges in the
connection regions (Hafner and French 1986; Jayashankar and French 1987),
are described next.

Bolted Connection (BME)

The prestressing strands in BME were each sleeved in plastic tubes (un-
bonded) over the first 26 in. of the beam to promote the formation of the
plastic hinge in an ordinary reinforced concrete section of the beam. Primary
longitudinal reinforcement in the failure region of the beam was provided
by four No. 6 rebar mechanically anchored at the interface to angle sections
cast into the top and bottom corners of the beam. Four vertical ducts were
cast between the angle sections to accommodate placement of 1-1/4 in. A490
bolts for connection of the beam and column elements. Four No. 3 hooks
were placed around the two ducts closest to the beam-column interface to
prevent bursting failure. Two fabricated T-sections with triangular gusset
plates were welded to plates that were anchored in the column with welded
headed studs. Four holes had been drilled in the stem of each tee for place-
ment of the connection bolts. The beam-column connection was completed
when the T-sections were bolted to the beam using a friction connection.
The cross-sectional details are shown in Fig. 5. It is evident from the
provided and induced moment diagrams that the plastic hinge was designed
to occur at approximately 26 in. from the column face.

Threaded Rebar Connection (BMF)

This connection was similar to the type of connection used for BMB with
two exceptions: the plastic hinge location and the anchorage of the connec-
tion reinforcement. BMB (Fig. 2) contained No. 9 threaded rebar anchored
at the exterior face of the column. The section was designed to distribute
the inelastic activity in the beam by yielding the bars at the interface and
then transferring the plastic hinge from the interface into the beam as the
bars began to strain harden. BMF used No. 7 threaded rebar to maintain a
3176

J. Struct. Eng. 1989.115:3171-3192.

 single plastic hinge at the interface, and the bars were anchored with 90°
hooks inside the column. This detail was used to prevent reinforcement slip-
page, as observed in BMB. Anchoring the bars at the exterior face of the
Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/09/14. Copyright ASCE. For personal use only; all rights reserved.

column in BMB was ineffective in resisting slip of the connection reinforce-

ment through the joint after yielding at the interface (AmU and French 1985).
Corrugated ducts extended from the face of the beam to blockouts located
35 in. from the column face to accommodate placement of the No. 7 threaded
rebar in the beam. Ordinary threaded rebar couplers were provided in the
column of BMF to connect the No. 7 threaded rebar in the beam with the
90° hooks cast into the column. Some inherent "slop" exists between the
couplers and threaded rebar (threads hot-rolled on bar), which is eliminated
during fabrication by grinding flush the ends of the two bars to be joined,
and subsequently tightening the bars against each other in the coupler. After
the No. 7 rebar were securely tightened into the couplers, anchor nuts were
attached to the threaded rebar in the blockouts to secure the rods against
embedded plates. The ducts and blockouts were subsequently grouted. The
anticipated failure zone at the interface is evident in Fig. 6.

Threaded Rebar Connection (BMG)

BMG was the only interior subassemblage tested. The connection repre-
sented a further modification of the connections used in exterior subassem-
blages BMB and BMF. No. 7 threaded rebar were used for the connection
reinforcement, as in BMF. The anchorage detail of BMF had been intended
to eliminate reinforcement slippage observed in the detail used for BMB. In
the tests of BMF, however, large slips were observed between the reinforce-
ment and the ordinary threaded rebar couplers. To eliminate this slip in the
test of BMG, tapered-threaded splices were used. Threads are mechanically
cut on the rebar to eliminate "slop" between the bar and the splice. The
threads are cut conically on the bar to prevent stress concentrations that would
be caused by an abrupt reduction in cross section. To reduce congestion, the
connection bars were run straight through the column using splices to couple
the reinforcement in the joint. The other ends of the rods were anchored
with nuts at 35 in. from the column face, similarly to BMF. The blockouts
and ducts were subsequently grouted.
It is not necessary to use threaded rebar (threads hot-rolled on bar) with
tapered-threaded splices. Threads can be cut onto ordinary rebar to mate with
the splices; this, in fact, was the intent of the manufacturer. The use of the
combination of threaded rebar and tapered-threaded splices is recommended
by the writers to facilitate mechanical anchorage of the bar in the beam away
from the column face. If mechanical anchorage is provided, the grout sur-
rounding the connection rod in the ducts need not be relied upon for an-
choring the rod. It would not be feasible to use tapered-threaded splices to
anchor the far end of the rod because it would not allow any length tolerance
in the field (length between tapered ends is fixed). The use of threaded rebar
provides unlimited length tolerance because a nut may be used to provide
mechanical anchorage anywhere along the length of the rod.
Because BMG represented an interior subassemblage, it was subjected to
a more severe load condition than the other models tested. Tension and
compression forces occur simultaneously in the reinforcement on each side
of an interior joint; consequently, if adequate anchorage is not available, the
resultant forces tend to cause bond deterioration and consequent slippage of
3177

J. Struct. Eng. 1989.115:3171-3192.

 continuous beam reinforcement through the joint. In the test of BMG, it was
assumed that the length of the bar through the joint region was not sufficient
to develop the required force entirely by bond. The minimum column width
Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/09/14. Copyright ASCE. For personal use only; all rights reserved.

suggested by ACI 352-85 to develop bond stresses was 17.5 in. for No. 7
bars; the provided column width was only 14 in. In addition, the splices that
comprised a total length of 4 of the 14 in. anchorage length had a smooth
surface. The possibility of wedge action created by the splices to resist the
bar force was ignored. To prevent slippage of the reinforcement through the
joint, plates were welded to the center of the tapered-threaded splices to
provide mechanical anchorage (Fig. 7—Detail A). The plates, referred to
as "bond plates" (Fenwick and Nguyen 1981), were well confined by the
transverse hoops of the column, so that even if the concrete cover spalled,
the plates could bear against the hoops. Besides preventing slippage, the
bond plates also contributed to the joint shear resistance by enhancing the
diagonal compressive strut.
Fig. 7 shows the cross-sectional details and the provided and induced bending
moment diagrams for the two beams of BMG: BMG-G1 and BMG-G2. Both
of the beams were originally designed to develop a plastic hinge at the beam-
column interface; however, the connection in BMG-G2 was later modified
to displace the hinge from the interface. This is discussed further in the next
section.

DESCRIPTION OF TEST SPECIMENS

The American Concrete Institute (ACI 318-83) "Building Code Require-

ments for Reinforced Concrete, Appendix A—Special Provisions for Seis-
mic Design" and ACI-ASCE Committee 352 (1985) "Recommendations for
Design of Beam-Column Joints in Monolithic Reinforced Concrete Struc-
tures" were used to design the structures.
The test setup was intended to simulate the joint region of a multistory
moment-resisting precast concrete frame. Pinned connections were used to
simulate inflection points assumed to occur near midheight of the columns
and midspan of the beams. Fig. 8 shows the general test setup of the exterior
subassemblages, BME and BMF, which were loaded with the hydraulic ac-
tuator attached vertically to the end of the beam (120 in. from the column
centerline). Note that the corbel, shown in the figure, was replaced by a
stiffened T-section for BME (Fig. 5). The interior specimen, BMG, was
loaded with the actuator attached horizontally to the bottom of the column.
The inflection points in the beams of BMG were simulated by means of pin-
ended rigid links (120 in. on either side of column centerline).
The precast columns were 14 in. square and 109 in. long, with a pin-to-
pin dimension of 120 in. (Fig. 8). The longitudinal reinforcement comprised
six No. 10 bars. Transverse reinforcement outside of a 20 in. distance from
both sides of the joint region consisted of No. 4 stirrups spaced at 5 in.
Within the depth of the joint and for 20 in. on each side of the joint face,
No. 4 stirrups were spaced at 2.5 in., with supplementary No. 3 and No. 4
ties placed parallel and perpendicular to the direction of bending, respec-
tively. No axial load was applied to the column. The compressive forces
provided by the addition of axial load would have enhanced the performance
of the joints.
The partially prestressed precast beams were 10 in. wide, 19 in. deep,
3178

J. Struct. Eng. 1989.115:3171-3192.

 SECTION A-A SECTION B-B
Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/09/14. Copyright ASCE. For personal use only; all rights reserved.

FIG. 8. Test Setup for Exterior Subassemblages

and 120 in. long (Fig. 8). The prestressed reinforcement consisted of 18
1/4-in. diameter 250 ksi seven-wire strands. Each strand was prestressed to
70% of the nominal tensile strength at a constant eccentricity. The non-
prestressed longitudinal reinforcement comprised four No. 3 bars symmet-
rically placed in the corners of the beam cages. Transverse reinforcement in
the beams comprised No. 3 stirrups spaced at 4 in. on center to satisfy ACI
318-83 A.3.3. The stirrups were closed with 135° bends and 10 bar-diameter
extensions to conform with ACI 352-85 4.2.2.6. This provision was not
satisfied for specimens BMA-BMC; consequently, the concrete deteriorated
in the vicinity of the plastic hinges located in the blockout regions of these
models because of inadequate confinement.
Distributing the plastic hinge away from the mortar joint was investigated
with one of the beams (BMG-G2) of BMG. The beam-column interface re-
gion of BMG-G2 was strengthened, as an afterthought, by providing addi-
tional No. 3 bars (Fig. 7) projecting 11 in. out of the east column face. 1-
in. diameter holes were drilled into BMG-G2 to mate with the No. 3 bars.
The holes were subsequently grouted. All of the prestressing strands in BMG-
G2 were debonded to a depth of 5 in. from the interface. Subsequent analysis
of the beam indicated that the plastic hinge of BMG-G2 would develop within
5 in. of the interface.

Material Properties
The reinforcement properties are listed in Table 1. The nominal concrete
compressive strength was 6,000 psi. The actual compressive strengths ranged
from 7,400 to 7,600 psi for the beams and from 6,900 (BMG) to 8,900
(BME and BMF) for the columns at the time of test.

Instrumentation
Strain gauges were used to measure strains in the prestressing strands and
reinforcing bars from the time the strands were stressed through completion
of the tests. A load cell mounted in the hydraulic actuator measured the
3179

J. Struct. Eng. 1989.115:3171-3192.

 applied loads. The reactions at the ends of the beams in the interior subas-
semblage, BMG, were measured with Clevis pin load cells mounted in the
rigid links. Linear voltage differential transducers (LVDT's) were used to
Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/09/14. Copyright ASCE. For personal use only; all rights reserved.

measure vertical beam and horizontal column displacements with respect to

an external reference frame. Cumulative rotations of the beam relative to the
column face were measured with LVDT's attached to the top and bottom of
the beam at 7, 14, 42, and 56 in. from the column face. These rotations
provided information regarding the distribution of the inelastic activity in the
beams. LVDT's were also used to measure the joint shear deformation and
slip of the reinforcement relative to the couplers in the joint.

Load History
The loading sequence comprised cycles to peak displacements defined in
terms of the nominal yield displacement (1 in.), which was calculated for a
reinforced concrete subassemblage of the same dimensions and capacity as
specimen BMB using the effective moment of inertia for the beam and the
gross moment of inertia for the column. Load cycles were repeated at each
peak displacement to observe the specimen strength and stiffness degrada-
tion. To facilitate comparison of test results, the same beam-end deflection
history was used in all of the tests. Peak values of the cycles corresponded
to displacement ductility factors (DDF = ratio of the beam end displacement
to the nominal yield displacement) of two cycles each at 0.75 and 2, and
three cycles each at 3 and 4, and, finally, to the limiting stroke of the ac-
tuator. In the case of structure BMG, the actuator was repositioned to take
the structure to the full stroke in one direction (DDF = 10). The entire
loading history corresponded to a nominal cumulative displacement ductility
of 53, where the cumulative displacement ductility is defined as the sum of
the DDF's of all cycles.

TEST RESULTS

Figs. 9(a~c) are photographs of the specimens loaded to a DDF of 4 (in-

terstory drift of 3.3%). Note in Fig. 9(c), BMG-G1 is located on the right-
hand side of the photograph; BMG-G2 is on the left. The following sections
provide a brief description of the behavior of the specimens observed during
the tests.

Model BME (Bolted)

During the first load cycle to a DDF of three, the structure attained its
maximum strength of 16.1 kips in the upward loading direction. The cal-
culated capacity was 18.6 kips. During loading to the limiting stroke of the
actuator (6.24 in.), an abrupt loss in strength was observed [Fig. 10(a)],
associated with major crushing of the top concrete. The load history rep-
resented a cumulative displacement ductility of 59.
The major hinging zone occurred in the beam between 8 and 20 in. from
the column face. Excessive diagonal shear cracking was observed in the
connection region [Fig. 9(a)]. This region of the beam, stiffened by trian-
gular gusset plates attached to the bolted T-sections, was quite rigid. Con-
sequently, force transfer through this region was similar to that of the joint
region—with the exception that this section was not designed as a joint. The
3180

J. Struct. Eng. 1989.115:3171-3192.

Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/09/14. Copyright ASCE. For personal use only; all rights reserved.

I hi

FIG. 9. Specimens (DDF = 4): (a) BME (Bolted); (b) BMF (Threaded Rebar); (c)
BMG (Threaded Rebar)

3181

J. Struct. Eng. 1989.115:3171-3192.

 20.0 - Load (kips)

M
Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/09/14. Copyright ASCE. For personal use only; all rights reserved.

15.0 -

10.0 -

5.0j

10.0 -7.5 -5.0 -/•S/l % W/?S 5.0 7.5 10.0

Displacement (in.)

^^15.0-

-20.0 -

20.0 - Load (kips)

W
15.0 -

10.0 -

5.0 -

0.0 -7.5 -5.0 / ) 2.5 5.0 7.5 10.0

Displacement (in.)

-15.0 -

J
-20.0

Load (kips)

FIG. 10. Load versus Deflection Response: (a) BME (Bolted); (b) BMF (Threaded
Rebar); (c) BMG-G1 (Threaded Rebar)

3182

J. Struct. Eng. 1989.115:3171-3192.

 Load (kips)
Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/09/14. Copyright ASCE. For personal use only; all rights reserved.

40.0 -| Load (kips)

M
30.0 -

20.0 -

10.0 -

0.0 -7.5 -5.0 ^ ^ l ' . 5 ^ sir 7.5 10.0

Displacement (in.)

'j/Mo.o -

-30.0 -

-40.0 -

FIG. 10. Load versus Deflection Response: (d) BMG-G2 (Threaded Rebar); (e)
BMG

observed behavior was similar to that exhibited by BMD, which contained

a welded detail stiffened at the top of the beam by a gusset plate and stiff-
ened at the bottom of the beam by the corbel.
Very little diagonal shear cracking was observed in the joint itself. Al-
though the diagonal compression strut is usually located within the limits of
the beam depth, the strut in the joint of this specimen extended well outside
of this region because of the spacing of the welded headed studs anchoring
the beam to the column. The increased length and inclination of the strut
provided a greater area of concrete to resist the shear, and there was an
increased number of stirrups crossing the potential diagonal splitting surface.
The maximum deformations due to shear in the joint were small. They were
on the order of 0.5 to 0.75 milliradians, which represented less than 1% of
the total inter-story drift (ratio of lateral story displacement to story height).

Model BMF (Threaded Rebar)

The maximum strength of the specimen (12.2 kips) was reached in the
upward loading direction during the first cycle to a DDF of three [Fig. 10(b)].
3183

J. Struct. Eng. 1989.115:3171-3192.

 The calculated capacity was 13.6 kips. The specimen maintained 91% of its
maximum measured capacity at the end of the test (DDF = 5.6). The load
history represented a cumulative displacement ductility of 63.4 (additional
Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/09/14. Copyright ASCE. For personal use only; all rights reserved.

half-cycle in negative direction).

The inelastic activity of BMF was concentrated at the beam-column in-
terface [Fig. 9(b)]. Although a major failure did not occur during the test,
significant slippage associated with the threaded rebar connector was ob-
served. The slippage comprised three components: the No. 7 longitudinal
bars in the beam slipping with respect to the couplers, the couplers slipping
with respect to the No. 7 hooks that anchored them in the column, and bond
deterioration due to poor grout quality in some of the beam ducts. The slip-
page was evidenced by severe pinching of the load-deflection curve shown
in Fig. 10(b). The total magnitude of slip that occurred was on the order of
0.20 in. for the two bottom No. 7 bars and approximately 0.15 in. for the
two top bars. Had this slippage not occurred, the energy dissipation capacity
of this structure would have likely been much higher.
The slippage associated with the coupler resulted from the plastic hinge
forming at the face of the coupler and the cyclic displacements that took the
rebar into the inelastic range. When the connection was fabricated, the re-
inforcement joined in the coupler was torqued such that the bars were bear-
ing against each other, thereby eliminating any slop between the reinforce-
ment and the coupler. This works well for gravity-loading applications, but
not for large lateral load reversals. The yielding at the face of the coupler
caused the reinforcement to elongate, and as the loads were cycled the re-
inforcement was worked back and forth in the coupler such that the bars
were no longer in bearing, thus enabling slippage of the reinforcement rel-
ative to the coupler.
The joint did not undergo significant deformation due to shear. The max-
imum magnitude of the joint shear deformation was on the order of only 2
milliradians, which represented approximately 3% of the inter-story drift.

Model BMG (Threaded Rebar—Interior Subassemblage)

Figs. 10(c) and 10(d) show plots of load versus deflection for BMG-G1
and BMG-G2, respectively, of structure BMG. The loads were measured
with the Clevis pin load cells mounted in the rigid links at the ends of the
beams; the displacements were measured at the point of load application at
the base of the column. From geometry, the beam end deflections were the
same as those applied at the base of the column that corresponded to the
displacement histories applied to the other models.
BMG-G1 and BMG-G2 carried different portions of the total lateral load
because supplementary reinforcement had been added at the interface of BMG-
G2 to distribute the plastic hinge [Fig. 9(c), BMG-G2 on the left side of
figure]. BMG-G1 had a single concentrated crack at the beam column-in-
terface. Throughout the test, BMG-G2 carried more load than BMG-G1.
When BMG was taken to the limiting stroke of the actuator (5.6 in.), the
beams still exhibited an increase in strength, indicating that the structure had
not yet reached the ultimate load. The maximum loads carried by each of
the beams at this displacement level were 13.6 and 15.1 kips, respectively,
compared to the calculated strengths of 13.7 and 16 kips for BMG-G1 and
BMG-G2, respectively. The jack was then repositioned and the structure was
taken to the full stroke (10 in.) in one direction. The power to the Clevis
3184

J. Struct. Eng. 1989.115:3171-3192.

 pin load cells was inadvertently cut when the jack was repositioned. Con-
sequently, the loads at the ends of the beams were not measured during this
portion of the test. Fig. 10(e) shows the load-versus-deflection plot for BMG,
Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/09/14. Copyright ASCE. For personal use only; all rights reserved.

in which the load and deflection represent those measured laterally at the
base of the column. An approximation of the maximum loads of BMG-G1
and BMG-G2 was obtained by adding one-half the difference between the
jack load measured at 5.6 in. and 8 in. (displacement at which the peak load
occurred) to the beam.loads measured at 5.6 in. This provides only a rough
estimation of the maximum loads, because the load was not distributed equally
between the two beams. The approximation results in maximum loads of
14.5 and 16.0 kips for BMG-G1 and BMG-G2, respectively. These loads
are likely to represent an upper bound for BMG-G1 and a lower bound for
BMG-G2.
Because of the slippage problems observed with the couplers in the test
of BMF, and because BMG was an interior joint, particular attention was
paid to the possibility of slippage assocated with the tapered-threaded splices
in BMG. To monitor any slippage, stiff wires were anchored to the bars and
routed through greased ducts in the specimen to LVDT's located at the ex-
terior face of the column. One of the LVDT's measured the displacement
of the slip wire attached to the tapered-threaded splice, while the other LVDT
measured the displacement of the slip wire attached to the No. 7 threaded
rebar located 1 in. from the splice. There was no evidence of relative slip
observed between the coupler and the No. 7 threaded rebar. The readings
were approximately equal to the elongation of the bar (assuming a constant
strain over a 1 in. length from the coupler), and the distance moved by the
slip wire due to the rotation of the beam. The absence of slip was also
evident from the load-deflection curves, which did not indicate pinching.
This indicates good performance of the tapered-threaded splices subjected to
load reversals.
The joint shear deformation measured for BMG was higher than those of
the other models, which is to be expected, because the joint shear is larger
in a laterally loaded interior joint than in an exterior joint. In the test of
BMG, the joint shear contribution to the inter-story drift exceeded 10%.

COMPARISON OF TEST SPECIMENS

In this section, the performance of the test specimens will be compared

based on their strength, ductility, and energy dissipation capacities. The
specimens were all subjected to the same beam end displacements to facil-
itate comparison of test results. However, the results must be normalized
for the following reasons: each of the specimens had different design ca-
pacities; the primary hinging region occurred at several different beam lo-
cations (for the same beam end displacement different rotational demands
were imposed on the beams); BMG was an interior connection, whereas all
other specimens were exterior connections (this contributed to the difference
in rotational demands imposed on the beams).

Strength
The strengths of the specimens were evaluated by comparing the calcu-
3185

J. Struct. Eng. 1989.115:3171-3192.

 IB.6
6.

1
1=1
2.; 5.
Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/09/14. Copyright ASCE. For personal use only; all rights reserved.

1I
3.E

BME BMF BMG-G1 BMG-G2

(Bolted) (Threaded Rebar) (Threaded Rebar) (Threaded Rebar)

• Measured Maximum Load (kips)

Calculated Maximum Load (kips)

FIG. 11. Comparison of Measured and Calculated Flexural Capacities (kips)

lated flexural capacities (see "Description of Connections" section) with the

actual strengths measured during the tests. As shown in Fig. 11, structures
BME and BMF attained approximately 87-90% of their calculated strengths,
and BMG-G1 and BMG-G2 attained or slightly exceeded their calculated
values.
The reduced measured strength of BME may be attributed to the shear
deterioration observed within the connection region. The plastic hinge was
to develop at approximately 26 in. from the interface; however, the extensive
diagonal shear cracking observed within the connection region promoted the
formation of the plastic hinge between 8 and 20 in. from the interface, thereby
causing a reduction in the load required to develop the ultimate capacity at
that cross section.
The only difference between BMF and BMG-G1 was the type of coupler
used in the column to splice the beam reinforcement to the reinforcement in
the joint. As a result, these models were expected to develop the same flex-
ural capacities (the slight variation in calculated strengths reflects the dif-
ference in concrete compressive strengths of these two models). BMF was
unable to attain the calculated strength because of the slippage associated
with the coupler. In this case, the slippage of the reinforcement relative to
the coupler accounted for at least 30% of the drift. Consequently, the struc-
ture was not subjected to large enough displacements in the test to achieve
its ultimate capacity.
BMG-G1 and BMG-G2 both reached or exceeded their respective ultimate
capacities. Though an internal joint is subjected to a more severe loading
condition due to the presence of compression on one face and tension on
the other, the tapered-threaded splices with bond plates worked well to elim-
inate slippage between the coupler and the reinforcement when subjected to
cyclic loading and yielding at the face of the splice.

Ductility
Ductility was evaluated in two ways: by the magnitude of the displacement
level (i.e., drift level) reached by the structures without a significant loss of
strength, and with respect to the number of increased displacement cycles
that occurred without the structures undergoing a significant loss of strength
after the ultimate strength was reached. The ultimate strengths of BME and
BMF were reached during the first cycle to a displacement of 3 in., whereas
3186

J. Struct. Eng. 1989.115:3171-3192.

 for BMG there was a steady increase in resistance offered by the beams until
a displacement of 8 in. was reached. The larger drift associated with BMG
resulted because it was an interior joint; consequently, it developed increased
Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/09/14. Copyright ASCE. For personal use only; all rights reserved.

column rotation and joint shear deformation in comparison to the exterior

specimens. At the end of the tests, BME and BMF were taken to DDF's of
6.2 and 5.6 (inter-story drifts of 5.1 and 4.7%), respectively, at which level
they still maintained 72 and 91% of their respective maximum capacities
measured at a DDF of three. The reduction in strength of BMG was only
approximately 6%, as the structure was taken to the maximum lateral dis-
placement of 10 in. (inter-story drift of 8.3%).
In general, all of the specimens tested in this investigation exhibited good
ductility characteristics. Structures BME-BMG exhibited greater ductility
compared to that observed in the tests of BMA-BMD, which developed plas-
tic hinges in the partially prestressed regions of the beam. (In the tests of
BMA-BMD, the structures maintained approximately 78% of their maximum
capacities when taken to DDF's of 4.5 [interstory drifts of 3.8%].) The en-
hanced ductility of BME-BMG may be attributed to the formation of the
plastic hinges in the connection regions that contained ordinary reinforce-
ment (a more ductile stress-strain curve than that of prestressing strand).

Energy Dissipation Capacity

Because BME-BMG developed plastic hinges in ordinary reinforced re-
gions of the beam, these structures also exhibited better energy dissipation
characteristics in comparison to those exhibited by BMA-BMD (Thompson
and Park 1980). The ductile behavior of the ordinary reinforcement leads to
larger load-deflection hysteresis loops, and consequently greater energy dis-
sipation.
Because the strengths of the three specimens BME-BMG were different,
a direct comparison of the energy dissipation values does not give a good
indication of the performance of the connections. In a previous paper com-
paring the response of structures BMA-BMD (French et al. 1989), the total
energy dissipated through a beam end displacement of 4 in. (DDF = 4) was
normalized with respect to the energy absorbed. The energies absorbed and
dissipated were taken as the areas under and enclosed within the load-de-
flection curves, respectively. The ratios obtained for all seven specimens,
BMA-BMG, are shown in Fig. 12(a) for comparison.
It should be noted that at a beam end displacement of 4 in. (nominal DDF
= 4) BMA-BMF had already reached their maximum capacities; however,
BMG-Gl and BMG-G2 did not reach their maximum capacities until a beam
end displacement of 8 in. Consequently, much more inelastic activity of
structures BMA-BMF is reflected in the bar chart of Fig. 12(a), and the
values given for BMG-Gl and BMG-G2 are misleading. The beam rotation
component of the drift was smaller in BMG than in the other specimens
because BMG was an interior connection, and consequently the column ro-
tation and joint shear deformation comprised larger proportions of the drift
in BMG. It is interesting to note, however, that the ratios for BMG-Gl and
BMG-G2 were still nearly as high as that given for BMF.
Despite the problems associated with shear degradation in the connection
region of BME, this specimen exhibited the best energy dissipation char-
acteristics. The plastic hinge was well distributed in the region of the beam
that contained the debonded tendons. The ratios of energy dissipated to en-

3187

J. Struct. Eng. 1989.115:3171-3192.

 o

o
Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/09/14. Copyright ASCE. For personal use only; all rights reserved.

(C
SI :k\\\\\\\\\\\\\\\^ c
a>
;K\\\^\\WN II ©
E

gkWWWWWW^
°k\\\\\\\\\\\\\\\^^
I
gkWWWWWW :k\\\\\\\\\\\\\\\^

gk\\"\W^\^^ Q
°k\\\\\\\\\\\\\\\^

3188

J. Struct. Eng. 1989.115:3171-3192.

 Load
Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/09/14. Copyright ASCE. For personal use only; all rights reserved.

+p s - • f'r/^y

>
+Aa Deflection

fjii- "%

1P 1
2.6 2.6

8.6 8.4
7.6

1I 1 1
6.9

BMA BMB BMC BMD BME BMF BMG-G1 BMG-G2

tPost-tenaioned) (Threaded Rebar) (Composite) (Welded) (Salted) (Threaded Rebar) (Threaded Rebar) (Threaded Rebar)

FIG. 13. Energy Dissipation Index, D,

ergy absorbed for BME-BMG all exceeded those given for structures BMA-
BMD (0.42-0.46), which developed plastic hinges in the partially pre-
stressed portions of the beams at 35 in. from the column face.
To eliminate the differences in inelastic demands associated with interior
versus exterior subassemblages, the energy dissipated through rotation of the
beams was investigated at particular values of cumulative rotation. Fig. 12(b)
shows the rotational energy dissipated by the beams normalized with respect
to the energy absorbed for the similarly designed beams of BMF and BMG.
The ratios are given for cumulative rotations of 10, 20, and 30 milliradians
measured within the first 14 in. of the beam relative to the column face
(which encompasses the plastic hinge region located near the interface). At
30 milliradians, the ratio of total energy dissipated to absorbed was 0.74 for
BMF, 0.84 for BMG-G1, and 0.78 for BMG-G2. The ratio of energy dis-
sipated to absorbed was greatest for BMG-G1, even though BMG-G2 was
slightly stronger than BMG-G1 and also had distributed cracks. This may
be attributed to the rotation in BMG-G1 occurring in the nonprestressed por-
tion of the beam. In BMG-G2, a large part of the rotation was observed in
the partially prestressed portion of the beam between 7 and 14 in. from the
interface; the strands were only debonded to 5 in. from the interface. Be-
cause prestressing strands tend to relieve their stress when the specimen is
brought back to its original position, they exhibit lower energy dissipation
characteristics. At rotations beyond 30 milliradians, BMG-G1 and BMG-G2

3189

J. Struct. Eng. 1989.115:3171-3192.

 continued to carry additional load, while BMF showed degradation of strength
due to slip of the threaded rebar relative to the coupler.
Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/09/14. Copyright ASCE. For personal use only; all rights reserved.

Another method to evaluate the energy dissipation characteristics is to nor-

malize the total energy dissipated in all of the cycles that can resist a load
greater than or equal to 75% of the yield load with respect to the energy
absorbed at yield (Darwin and Namai 1986). The resulting ratio is called the
"dissipation index" (D,). The dissipation indices, D,'s, are shown in Fig. 13
for B M A - B M G . The good energy dissipation characteristics associated with
plastic hinging in an ordinary reinforced concrete portion of the beam are
evident from the high dissipation index values associated with BME and
BMG-G1. The lower values obtained for the other specimens may be at-
tributed to the slippage associated with the ordinary threaded rebar coupler
(BMF) and the hinging within the partially prestressed portion of the beams.

CONCLUSIONS AND RECOMMENDATIONS

1. All of the specimens exhibited adequate strength. BMG resisted loads

greater than the design capacity. The capacities of structures BME and BMF
were approximately 87-90% of their respective design capacities. The reduced
capacities of BME and BMF resulted from shear degradation in the connection
region of the beam (BME) and slippage of the reinforcement in the coupler (BMF).
2. All of the specimens displayed good ductility characteristics. All were
subjected to inter-story drifts (ratio of the column end displacement to story height)
exceeding 4.0% without undergoing much strength degradation.
3. Energy dissipation capacity was greatest when hinging occurred in non-
prestressed regions. As a result, BME-BMG exhibited better energy dissipation
characteristics than structures BMA-BMD, for which hinging was forced to oc-
cur in the partially prestressed regions of the beams. The best characteristics
were observed for BME and BMG-G1.
4. The slightly lower energy dissipation capacity exhibited by BMF was at-
tributed to slippage of the threaded rebar relative to the coupler. Reinforcement
slippage allows the structure to undergo increased deflection without an increase
in resistance (load), which results in pinching of the load-deflection curve and
corresponding decreases in measured energy dissipation.
5. There was no benefit observed in spreading the hinge region in BMG-
G2. Although it reduced the rotational demand imposed at the interface and re-
duced the likelihood of relative slip between the splice and the connection bar,
it did not increase the energy dissipation capacity of the beam. Comparing the
results of BMG-G1 and BMG-G2, the energy dissipation capacity of BMG-G1
(hinge concentrated at the interface) was comparable to or slightly greater than
that of BMG-G2 because the hinging migrated into the partially prestressed por-
tion of BMG-G2. The energy dissipation capacity would be likely to increase
for a distributed hinge if the inelasticity is maintained in the ordinary reinforced-
concrete region of the beam. This was evidenced in the test of BME, for which
the strands were sleeved within the connection region.
6. The slippage associated with the coupler of BMF might be eliminated by
tightening a nut onto the threaded rebar at the face of the coupler. The appli-
cation of the nut could, in effect, prestress the rod in the coupler to prevent
"slop." If the plastic hinge is still maintained at the coupler face, however, the
rod may still elongate at the interface and reintroduce "slop" into the connection

3190

J. Struct. Eng. 1989.115:3171-3192.

 through cyclic loading. Consequently, this modification should be tested before
conclusions on its effectiveness are drawn. The results of the present study in-
Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/09/14. Copyright ASCE. For personal use only; all rights reserved.

dicate that the use of the ordinary threaded rebar couplers should be restricted
in high-intensity seismic regions, where it is desirable to dissipate a large amount
of energy as the structure is cycled into the inelastic range. If this detail is to
be used in such a location, the plastic hinge region should be located at a distance
from the coupler.
7. The use of the tapered-threaded splices in BMG eliminated slippage of
the reinforcement relative to the coupler, despite yielding of the bars adjacent
to it.
8. The bond plates worked well in the joint of BMG to provide anchorage
to the splices and reinforcement through the joint. The use of such bond plates
alleviates limitations imposed on column widths and beam bar sizes required to
insure adequate reinforcement anchorage in the joint. This implies that column
and beam sizes may be reduced, which would in turn, reduce the overall weight
of the building and increase the structural flexibility. Consequently, this may
lead to a reduction in the design lateral load for earthquake resistance; however,
a more flexible structure is associated with larger lateral drifts.
9. The bond plates also served to enhance the diagonal compressive strut in
the joint.
10. Additional transverse shear reinforcement is necessary in the connection
regions of beams with rigid details like specimens BMD and BME. For specimen
BME, increased shear reinforcement would be required in the connection region
of the beam within 12 in. of the column face. If the connection region had
contained adequate shear reinforcement, overall behavior of this connection would
have likely been improved.

In summary, the three connections described in this paper, B M E - B M F ,

behaved well especially in comparison to the results obtained for specimens
B M A - B M D , in which hinging was forced to occur in the partially pre-
stressed regions of the beams. Of all seven connections, the bolted connec-
tion (BME), was one of the most difficult to fabricate. Some of the short-
comings of the fabrication of BME could be avoided by welding the connection
T-sections to the columns at the precasting plant, rather than overhead weld-
ing them in the field. On site, these T-sections would facilitate placement
of the beam elements. The primary shortcoming of BME is that it requires
close tolerances in the alignment of the ducts and holes in the plate and T-
sections to accommodate placement of the friction-anchored bolts.
Overall, the threaded rebar connection with tapered-threaded splices ap-
pears to hold the most promise for use in regions of moderate to high seis-
micity. The threaded rebar connection with tapered-threaded splices is easy
to fabricate and emulates the behavior of ordinary reinforced concrete within
the precast system.

ACKNOWLEDGMENTS

This research investigation was sponsored by the National Science Foun-

dation under Grant No. CES-8451536. The writers also wish to acknowledge
Dywidag Systems International Inc., Erico Products Inc., Minnesota Rebar,
and Spancrete Midwest Co. The work was carried out at the Civil and Min-

3191

J. Struct. Eng. 1989.115:3171-3192.

 eral Engineering Structures Laboratory at the University of Minnesota. The
views expressed herein are those of the writers and do not necessarily reflect
those of the sponsors. The assistance of P. Bergson and D. Chandra is greatly
Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/09/14. Copyright ASCE. For personal use only; all rights reserved.

appreciated.

APPENDIX I. UNIT CONVERSION

1 in. = 25.4 mm;

1 kip = 4,450 N;
1 ksi = 6.9 MPa;
1 kip-ft = 1,360 N-m.

APPENDIX II. REFERENCES

Amu, O. O., and French, C- W. (1985). "Moment resistant connections in precast

structures subjected to cyclic lateral loads." Struct. Res. Report No. 87-07, Dept.
of Civ. and Mineral Engrg., Univ. of Minnesota, Minneapolis, Minn.
Building code requirements for reinforced concrete. (1983). ACI 318-83, Amer. Concr.
Inst., ACI Comm. 318, Detroit, Mich.
Darwin, D., and Nmai, C. (1986). "Energy dissipation in RC beams under cyclic
load." / . Struct. Engrg., ASCE, 112(8), 1829-1846.
Fenwick, R. C , and Nguyen, H. T. (1981). "Reinforced concrete beam-column
joints for seismic loading." School of Engrg. Report No. 220, Dept. of Civ. Engrg.,
Univ. of Auckland, Auckland, New Zealand.
French, C. W., Amu, O., and Tarzikhan, C. (1989). "Connections between precast
elements—Failure outside connection region." / . Struct. Engrg., ASCE, 115(2),
316-340.
Hafner, M., and French, C. W. (1986). "Moment resistant connections between pre-
cast elements subjected to cyclic lateral loads." Struct. Res. Report No. 87-09,
Dept. of Civ. and Mineral Engrg., Univ. of Minnesota, Minneapolis, Minn.
Hognestad, E. (1951). "A study of combined bending and axial load in reinforced
concrete members." Bull. Series No. 399, Univ. of Illinois Engrg. Exp. Station,
Urbana, 111.
Jayashankar, V., and French, C. W. (1987). "An interior moment resistant connec-
tion between precast elements subjected to cyclic lateral loads." Struct. Res. Re-
port No. 87-10, Dept. of Civ. and Mineral Engrg., Univ. of Minnesota, Minne-
apolis, Minn.
Martin, L. D., and Korkosz, W. J. (1982). "Connections for precast concrete build-
ings including earthquake resistance." Tech. Report No. 2, Prestressed Concr. Inst.,
Chicago, 111.
Mattock, A. H., Kriz, L. B., and Hognestad, E. (1961). "Rectangular concrete stress
distribution in ultimate strength design," J. Amer. Concr. Inst., ASCE, 57(8),
875-926.
"Recommendations for design of beam-column joints in monolithic reinforced con-
crete structures." (1985). J. Amer. Concr. Inst., 82(3), 266-283.
Tarzikhan, C , and French, C. W. (1987). "Welded and composite connections be-
tween precast elements subjected to cyclic lateral loads." Struct. Res. Report No.
87-08, Dept. of Civ. and Mineral Engrg., Univ. of Minnesota, Minneapolis, Minn.
Thompson, K. J., and Park, R. (1978). "Stress-strain models for Grade 275 rein-
forcing steel with cyclic loading," Bull. New Zealand Nat. Soc. Earthquake Engrg.,
11(2), 101-109.
Thompson, K. J., and Park, R. (1980). "Ductility of prestressed and partially pre-
stressed concrete beam sections." Prestressed Concr. Inst. J., 25(2), 46-70.

3192

J. Struct. Eng. 1989.115:3171-3192.