Download as pdf or txt
Download as pdf or txt
You are on page 1of 11

Seventh National Seismic Conference on Bridges & Highways

Seismic Evaluation of Grouted Splice Sleeve


Connections for Precast Reinforced Concrete Bridge Piers
M.J. Ameli1, J.E. Parks1, D.N. Brown1, C.P. Pantelides1, J. Sletten2, and C.
Swanwick2

ABSTRACT

Connections between columns and footings and columns and bridge pier caps must be
able to withstand significant shear and rotational forces in large earthquakes. These connections
are important in the seismic design of bridges utilizing prefabricated bridge elements and
systems (PBES). PBES is proving to be an effective technique for accelerating bridge
construction. There is limited data for the expected performance of grouted splice sleeve
connections between precast bridge components in bridges exposed to seismic forces. Results of
quasi-static cyclic tests are presented for two splice sleeve connections. In both cases the splice
sleeves were located in the columns. In the first case, a splice sleeve connection with the bars
grouted at both ends (GGSS) between a reinforced concrete column and a footing was
constructed at half-scale of a typical bridge. The columns had an octagonal shape but the steel
reinforcing bars were arranged in a circular pattern. Results from tests of a splice sleeve
connection between a reinforced concrete column and a bridge pier cap, with the column bars
fastened to the sleeve and the dowels from the pier cap grouted in the column (FGSS) are also
presented. The performance of the two tests is compared in terms of strength, ductility and
energy dissipation. The performance is also compared to performance expectations for similar
monolithic connections.

1
Dept. of Civil and Environmental Engineering, University of Utah, Salt Lake City, UT 84112
2
Structures Division, Utah Department of Transportation, Salt Lake City, UT 84114

Paper No. A3-5 Page 1 Oakland, California | May 20-22, 2013


Seventh National Seismic Conference on Bridges & Highways

INTRODUCTION

Accelerated Bridge Construction (ABC) refers to a method of bridge construction that


incorporates innovative techniques, methodologies, and materials to efficiently reduce the
construction time and traffic disruption; it provides a higher level of work-zone safety for
workers and commuters, and implements environmental-friendly procedures.
Many Bridges have been built or rehabilitated following ABC standards. Local examples
in the state of Utah include the I-15 Provo Center Street Interchange, the Riverdale Road over I-
84, and the I-15 South Layton Interchange. Precast concrete deck panels, substructures, and
superstructures have been frequently utilized as effective ABC methods. The connection between
these precast elements is one of the most critical components of the structure, so researchers are
in the process of investigating the suitability of various connection configurations, especially in
moderate-to-high seismic regions. These connections, not only have to conform to ABC
standards in terms of overall construction delivery time, but also must resist high levels of
earthquake-induced deformations and stresses. Lateral load capacity, ductility levels, and
repairability are significant acceptance criteria for any connection considered in earthquake-
prone regions.
Research on bridge components connected by means of Grouted Splice Sleeves (GSS) is
limited. Aida et al. (2005) reported testing of three 75% scale specimens in Japan, two of which
used GSS column to footing connections. The specimens, representing railroad bridge column to
footing connections, were heavily reinforced and tested under cyclic loading to investigate their
inelastic performance. Specimens with GSS showed acceptable inelastic behavior in terms of
strength and ductility under cyclic loads, compared to the cast-in-place specimen.
Haber (2012) discussed the results of three specimens tested at the University of Nevada,
Reno. GSS were used in two specimens, one of which incorporated a precast pedestal to reduce
the moment demand over the coupler region, and decrease the severity of damage in the footing.
The third specimen was cast monolithically as the benchmark specimen. Due to the relatively
high rigidity of the column section over the splice sleeve region, plastic deformations did not
occur in the plastic hinge region; spliced rebar ruptured in the footing near the column-footing
interface. All specimens exhibited similar performance in terms of ultimate load capacity, energy
dissipation, and pushover response, but not ductility capacity.
An evaluation of several ABC connections in moderate-to high seismic regions was
summarized in NCHRP Report 698 (Lee et al., 2011). Connection types included bar couplers,
grouted ducts, pocket connections, socket connections, hybrid connections, integral connections,
and emerging technologies, such as shape memory alloys and elastomeric bearings. These
connection types were either being utilized in actual practice or developed in research. An
outcome of this study was the prioritization of research studies essential for each connection type
in order to fully understand their behavior under seismic actions; three research priorities were
suggested. First priority was given to the cyclic performance of the GSS with the spliced bars in
their plastic range of stress. The second priority was investigation of the strength details, such as
magnitude of stress that each bar can develop, together with the preferred location of the GSS
(e.g. in the column or footing), so that it does not affect overall response to earthquake loads,
while being easy to construct. The effect of surrounding concrete and level of confinement of the
GSS was prioritized as the last necessary study to be conducted on this type of ABC connection.
In an effort to carry out a comprehensive study on the performance of GSS bridge
connections in regions of high seismicity, a research program was developed at the University of

Paper No. A3-5 Page 2 Oakland, California | May 20-22, 2013


Seventh National Seismic Conference on Bridges & Highways

Utah, for conducting experiments on half-scale bridge column to footing and column to pier cap
splice sleeve connections. This paper discusses the design, construction, experimental
procedure, and preliminary results for two test specimens that incorporated two types of GSS
connections.

DESIGN AND CONSTRUCTION OF TEST SPECIMENS

Figure 1 shows the two types of GSS connections utilized for the test specimens. The
splice sleeve with the bars grouted at both ends (GGSS) was used in the connection of the
column to footing, and the one with one bar grouted and the other bar fastened (FGSS) was
incorporated in the column to pier cap connection. The specimens were designed and detailed to
simulate typical prototype bridges constructed in Utah, following the AASHTO LRFD Bridge
Design Specifications (2012), and the AASHTO Guide Specifications for LRFD Seismic Bridge
Design (2011), in accordance with capacity-based design procedures. The circular configuration
of column longitudinal bars and octagonal column cross section is the method of choice for
bridge designers in Utah, as shown in Figure 2.
The column dimensions were identical for both specimens, having a 21 in. square top
portion to provide a proper grip region, in order to engage with the lateral load transfer setup.
The cross section changed to an octagon to facilitate concrete casting. Six No. 8 bars, in addition
to a No. 4 spiral with a pitch of 2½ in., made up the column cage. The longitudinal and
transverse reinforcement ratios were 1.3% and 1.9% respectively, conforming to the Caltrans
Seismic Design Criteria (SDC). The footing was a 6ft x 3ft x 2ft precast element and consisted of
No. 8 longitudinal bars enclosed by No. 4 double hoops. The 9ft x 2ft x 2ft precast pier cap had
the same rebar size and arrangement as the footing. The footing and pier cap were designed to
remain linearly elastic and not undergo plastic deformations. The design inhibits shear failure
from occurring in the column, and the desirable column failure mode was set to either be flexural
or splice failure.

Figure 1. Two types of splice sleeves used in the test specimens.

Paper No. A3-5 Page 3 Oakland, California | May 20-22, 2013


Seventh National Seismic Conference on Bridges & Highways

The column cages, with the splice sleeves secured inside them, the footing and pier cap
were built in the laboratory. Dowel bars of proper length were extended out of the footing and
pier cap, according to the manufacturers’ guidelines. A step-by-step procedure was carried out
for the grouting operation, in which the grout was pumped into the sleeves, when installing the
precast components. A ¼ in.-thick bed grout was placed between the column and base, using the
same grout as in the sleeves. The specimens were instrumented with several strain gauges
especially in the plastic hinge region and in the joint area. Reinforcing steel, concrete, and grout
used in the test specimens were tested according to ASTM standards and the material properties
are summarized in Table I.

Figure 2. Details of the test specimens.

TABLE I. MATERIAL PROPERTIES (ksi)


Column Rebar
Concrete Grout
Specimen Longitudinal (No. 8) Transverse (No. 4)
Yield Ultimate Yield Ultimate 28-day Test day 28-day Test day
GGSS-1 68 93 63 103 5.3 5.9 14.4 14.4
FGSS-1 75 103 63 103 5.3 6.2 12.5 13.3

Paper No. A3-5 Page 4 Oakland, California | May 20-22, 2013


Seventh National Seismic Conference on Bridges & Highways

TEST PROCEDURE

Linear variable differential transformers were mounted on the column end—five on each
side—to measure relative vertical displacements between sections, and provide data for
generated strain and moment-curvature analyses. String potentiometers were also used to capture
the horizontal displacement of the column top.
Reversed cyclic quasi-static loading was applied to both test specimens using a 120-kip
servo-controlled actuator, with an overall stroke of 18 in., as shown in Figure 3. The tests were
carried out using displacement control, using a history comprised of peak displacements as
multipliers of the predicted yield displacement; the first step corresponded to half the yield
displacement. Two cycles per displacement level were employed as presented in Figure 4. The
axial load application system consisted of a cylindrical 500-kip hydraulic actuator, a 3ft-long
stiffened W14x90 steel section, a 3in. x 15in. x 36in. steel plate, a spherical plate, and 14ft-6in.
long 150-ksi threaded rods. A constant axial load of 110 kip was applied corresponding to 6% of
the axial capacity of the column.

Figure 3. Test Setup.

Paper No. A3-5 Page 5 Oakland, California | May 20-22, 2013


Seventh National Seismic Conference on Bridges & Highways

Figure 4. Applied Displacement History.

TEST RESULTS

The hysteresis response of the specimens, shown in Figure 5, illustrates the significant
stages of performance. These stages mark the formation and extension of cracks and concrete
spalling initiation, yield penetration, and failure of the specimens (due to either rebar fracture in
GGSS-1, or rebar pull-out from the splice sleeves in FGSS-1). The hysteresis loops were wide
and stable for GGSS-1. Pinching occurred in the hysteresis response of FGSS-1 and resulted in
narrower loops, especially after 4% drift; this is attributed to pullout of the bars from the FGSS-1
sleeves for this specimen. Both specimens had nearly the same load capacity of 49 kips, although
FGSS-1 showed a 15% lower value in the pull direction, resulting in an asymmetrical inelastic
response. Strength degradation for FGSS-1 began at the 3% drift level, corresponding to the pull
direction, while it started at the 5% drift level in the push direction. Specimen GGSS-1 exhibited
a slight strength reduction at a drift level of 6%, in both directions; this specimen had a strength
reduction greater than 20% in both directions during the first cycle of the last drift level (i.e.
approximately 9% drift), associated with fracture of the two extreme bars due to low cycle
fatigue. FGSS-1 withstood all of the drift levels up to 6%, when a drop of greater than 20%
occurred in the capacity, due to two of the bars pulling out of their sleeves, during the second
cycle of the 6% drift level in both directions.
Figures 6-8 show damage patterns at different drift levels for both test specimens. The
specimens were painted white; grid lines divided the concrete surface into 4in. squares. A black
marker was used to mark cracks for GGSS-1 and a red marker for FGSS-1 at each drift level.
Damage was concentrated in two key sections—namely, the bed grout and the section just above
the sleeves; a green dashed line identifies the location of the top of the sleeves. Another major
crack occurred at the end of the spiral-overlapping zone, which was formed because of using two
spirals with different outside diameters, the exterior spiral that surrounded the sleeves and the
interior that enclosed the bars.

Paper No. A3-5 Page 6 Oakland, California | May 20-22, 2013


Seventh National Seismic Conference on Bridges & Highways

Figure 5. Hysteretic Behavior of the Two Specimens with Damage States.

The first crack formed at the bed grout section, accompanied by another crack just above
the sleeves, during the first cycle of the 1% drift level, for both specimens. All of the major
cracks developed by the end of the 3% drift level. Figure 6 shows the extent of damage after this
step for both specimens. Spalling initiated during the first cycle of the 3% drift level, progressed
at the corners of the octagonal column section. The damaged area had a height of 4in. on both
sides of the column and a crack width of 0.009 in. was measured for GGSS-1, and 0.002 in. for
FGSS-1.
Cracks widened and spalling progressed at higher drift levels. Another major stage in the
performance of the specimens was observed at the 6% drift level, when yield penetration resulted
in concrete crushing at top of the footing, adjacent to the column for GGSS-1. The penetration
depth measured 1½ in. and 1in. on the west and east side of the column, respectively. Figure 7
presents the state of damage of both specimens. On the contrary, yield penetration did not
develop for FGSS-1, as the bars gradually pulled out of the sleeves. The cone shape of the
expelled grout was obvious, as shown in Figure 7(c); the column longitudinal bars barely yielded
at the pier cap to column interface. Testing of FGSS-1 was terminated after completion of the
6% drift level, when the capacity dropped by more than 20%.
The height of the spalling region was found to be 8in. and 12in. on the west and east side
of the column for both specimens, and the spiral became partially exposed. The width of
representative cracks was 0.03in. and 0.01in. for GGSS-1 and FGSS-1, respectively. The bed
grout deteriorated at the column peripheral of both specimens; the permanent crack at the bed
grout for FGSS-1 was larger and measured 0.1in.
During the last drift level of 9% for GGSS-1, twisting of the column became evident; all
six longitudinal column bars underwent some twisting. Concrete spalling grew larger in terms of
area and depth, and the spiral and sleeves were exposed, as shown in Figure 8. The two extreme
bars fractured in the first cycle of the push and pull directions of the 9% drift level, implying low

Paper No. A3-5 Page 7 Oakland, California | May 20-22, 2013


Seventh National Seismic Conference on Bridges & Highways

(a) GGSS-1 (b) FGSS-1


Figure 6. Damage Patterns at 3% Drift Level.

(a) GGSS-1 Cracks and Spalling (b) GGSS-1 Yield Penetration

(c) FGSS-1 Bars Pulled Out of the Sleeves (d) FGSS-1 Cracks and Spalling
Figure 7. Damage Patterns at 6% Drift Level.

cycle fatigue failure. Rebar fracture occurred 1in. to 1½ in. below the surface of the footing,
where there was no confining transverse reinforcement. All six bars buckled considerably at this
unconfined section located at the top of the joint area. Post-test investigations ascertained that the
confined concrete core remained undamaged, and the sleeves themselves did not slip for either
specimen. The footing and pier cap remained perfectly elastic as capacity-protected members
with minor hairline cracks in the joint region.

Paper No. A3-5 Page 8 Oakland, California | May 20-22, 2013


Seventh National Seismic Conference on Bridges & Highways

(a) GGSS-1 After Failure (b) GGSS-1 Rebar Fracture


Figure.8 Damage Patterns at 9% Drift Level.

Figure 9 shows average force-displacement curves that were constructed considering the
peaks of the first complete cycle at each drift level. The rising portions of the curves are almost
identical, while the inelastic portions differ in terms of load and displacement capacity.
Specimen FGSS-1 displaced more than GGSS-1 beginning at the 3% drift level; this is attributed
to the gradual pull-out which caused early failure of FGSS-1. The displacement ductility was
obtained for both specimens using idealized elasto-plastic behavior. The displacement ductility is
equal to 5.8 and 4.5 for GGSS-1 and FGSS-1, respectively. These values will be compared to
upcoming tests of control specimens for column to footing and column to pier cap connections.
In the absence of such results, and according to Priestley et al. (1996), the displacement ductility
for the cast-in-place specimen was predicted as 10.0, regardless of the flexibility levels offered
by either the footing or the beam.
One of the main features of a bridge ductile element in high seismic regions is its ability
to dissipate energy through inelastic deformations. The presence of mild steel in the plastic hinge
region capable of undergoing inelastic behavior is very important for achieving the required
amount of energy dissipation. The location of maximum deformation demand on the rebar was
shifted to the section below the column-to-base interface, due to the presence of the splice
sleeves. GGSS-1 dissipated more energy than FGSS-1, since it was a more ductile specimen,
with a ductile failure mode. This is shown in Figure 10, in terms of dissipated energy per cycle
and cumulative dissipated energy. Both specimens dissipated nearly the same amounts of energy
up to a drift level of 4%. FGSS-1 exhibited lower energy dissipation than GGSS-1 in the
remaining cycles before failure. Overall, GGSS-1 showed an acceptable energy dissipation
capacity.

SUMMARY AND CONCLUDING REMARKS

Design, construction, and preliminary results for two half-scale test specimens were
discussed. Two different grouted splice sleeves were utilized to connect the precast components.
In both cases the splice sleeves were located in the columns. Column longitudinal bars and
footing dowels were confined by the grout inside the splice sleeve in GGSS-1. In FGSS-1 the
column longitudinal bars were fastened to the top of the sleeves at the threaded end, and the
beam dowels were grouted inside the sleeves. Both specimens demonstrated similar performance
with respect to load capacity; however, they performed differently in terms of ductility and

Paper No. A3-5 Page 9 Oakland, California | May 20-22, 2013


Seventh National Seismic Conference on Bridges & Highways

Figure 9. Average Load-Displacement Response.

Figure 10. Hysteretic Energy.

dissipated energy. The overall deformation behavior of both specimens under cyclic loads
represented more of a rigid rotation than flexural bending, resulting in the formation of just a few
severe cracks in the column. Major crack formation in addition to concrete spalling occurred at a
drift level of 3%. At a 6% drift level, FGSS-1 failed due to bars pulling out of the sleeves, while
yield penetration occurred in GGSS-1. Specimen GGSS-1 failed at a 9% drift level, as a result of

Paper No. A3-5 Page 10 Oakland, California | May 20-22, 2013


Seventh National Seismic Conference on Bridges & Highways

fracture of two longitudinal column bars below the surface of the footing. Displacement ductility
was 5.8 and 4.5 for GGSS-1 and FGSS-1, respectively; these values are applicable in regions of
moderate-to-high seismicity, according to Caltrans SDC that requires the minimum local
displacement ductility of 3.0.
There was a significant difference in energy dissipation capacity of the two specimens.
GGSS-1 dissipated at least twice the hysteretic energy of FGSS-1 because of the wide and stable
hysteresis loops together with a greater displacement capacity.
In upcoming phases, alternative approaches with a different sleeve configuration will be
investigated for both categories of column-to-footing and column to pier-cap connections. The
cyclic performance of monolithic control specimens will also be investigated. Consequently, a
better understanding of the results will be possible through comparisons with the control
specimens.

ACKNOWLEDGMENTS
The authors would like to acknowledge the financial support of the Utah Department of
Transportation, the New York State Department of Transportation and the Texas Department of
Transportation. The authors wish to thank Professor Lawrence D. Reaveley for his invaluable
input throughout the project. Special thanks are extended to Mark Bryant for his untiring efforts
and support.

REFERENCES
Aida, H., et al. (2005). “Cyclic Loading Experiment of Precast Columns of Railway Rigid-Frame Viaduct Installed
with NMB Splice Sleeves,” Proceedings of Japan Concrete Institute, 27 (2), 20 pp.
American Association of State Highway and Transportation Officials (AASHTO). (2012). AASHTO LRFD Bridge
Design Specifications, Customary U.S. Units, 6th ed., AASHTO, Washington, D.C.
American Association of State Highway and Transportation Officials (AASHTO). (2011). AASHTO Guide
Specifications for LRFD Seismic Bridge Design, 2nd ed., AASHTO, Washington, D.C.
California Department of Transportation. (2010). “Seismic Design Criteria,” Division of Engineering Services,
Sacramento, CA.
Haber, Z. B. (2012). "Seismic Performance of Emulative Precast Bridge Column Elements with Grouted Coupler
Connections," Proceedings of International Bridge Conference, Pittsburgh, Pennsylvania.
Lee, M., Wernly, M. Garrett, B., Stanton, J. F., Eberhard, M. O. and Weinert, M. (2011). “Application of
Accelerated Bridge Construction Connections in Moderate-to-High Seismic Regions,” NCHRP Report 698,
Washington, D.C.
Priestley, M. J. N., Seible, E, and Calvi, M. (1996). Seismic design and retrofit of bridges. John Wiley & Sons, Inc.,
New York, N.Y.

Paper No. A3-5 Page 11 Oakland, California | May 20-22, 2013

You might also like