Professional Documents
Culture Documents
42 97 124
Earthquake Retrofit of Precast Panels
Reconnaissance Grouted Sleeve for Lateral-Force
Team Reports Connections Resistance
PCI Journal Reviewer Note of Appreciation
Every manuscript submitted to PCI Journal for publication as a peer-reviewed paper goes through a thorough tech-
nical appraisal. Specialists in the paper’s subject area conduct this review process.
Following are the members of this dedicated group of volunteer reviewers who in 2011 appraised prospective PCI
Journal papers. This year it was an especially demanding job as we worked to clear the backlog while also reducing
the length of time from submission to publication. PCI is grateful to each reviewer for the time and effort expended
in performing this valuable service to the engineering profession, the precast/prestressed concrete industry, and
ultimately the public.
Seismic
Preview of PCI’s New Zealand Earthquake Reconnaissance Team Report 42
42 Robert B. Fleischman, José I. Restrepo, Joseph R. Maffei, and Kim Seeber
52 Architectural Precast Concrete Panel Systems Used for Lateral-Force Resistance 124
J. Paul Hobelmann, Macarena Schachter, and Matthew C. Cooper
76
The PCI Journal (ISSN 0887-9672) is published quarterly by the Precast/
EDITORIAL DESIGN & PRODUCTION ADVERTISING SALES Prestressed Concrete Institute, 200 W. Adams St., Suite 2100, Chicago,
IL 60606. Copyright © 2012, Precast/Prestressed Concrete Institute. The
Rachel J. Detwiler, PhD, P.E. Editor-in-Chief Paul Grigonis Director, Publishing and Art Direction Kirstin Osgood Manager, Precast/Prestressed Concrete Institute is not responsible for statements
(rdetwiler@pci.org) (pgrigonis@pci.org) Sales and Member Development made by authors of papers or claims made by advertisers in the PCI Journal.
Michelle Burgess Managing Editor Ed Derwent Associate, Graphic Design (kosgood@pci.org) Original manuscripts and letters on published articles are accepted on
review by the PCI Technical Review Committee. No payment is offered. Direct
(mburgess@pci.org) (ederwent@pci.org) (312) 360-3206 all correspondence to PCI Journal at journal@pci.org or Precast/Prestressed
Carrie Wyrick Technical Editor Concrete Institute, c/o PCI Journal, 200 W. Adams St., Suite 2100, Chicago,
(carrie.wyrick@gmail.com) IL 60606. For information on advertising rates, send an email to adsales@
pci.org. Subscription rates are $70 per year and $175 for three years in the
Heather Green Copy Editor United States and $150 per year and $410 for three years for international.
Nyema Russell Administrative Assistant A single or back issue is $12. International subscriptions are delivered by an
(nrussell@pci.org) international carrier; allow one to three weeks.
This paper is milled from a 3rd certified source
2 W in t e r 2 0 1 2 | PCI Journal
Winter 2012 • Volume 57, Number 1
Editor’s Message 11
The Cover Story
Letters 13
Our Members 29
Project Spotlight 36
View the index for the 2011 PCI Journal articles on the PCI PCI Staff 151
website at www.pci.org/publications/journal/index.cfm.
Meet Doug Sutton 152
On the
WINTER 2012
VOLUME 57, NUMBER 1
Cover
Precast/Prestressed Concrete Institute
42 97 124
Earthquake Retrofit of Precast Panels
Reconnaissance Grouted Sleeve for Lateral-Force
Team Reports Connections Resistance
www.cegengineers.com
PIECES FIT
Architects . Engineers . Parking Consultants
The power of
collaboration
Over the past 26 years that I’ve been involved with PCI, I’ve been genuinely amazed by the
willingness of so many individuals to give of their time and talents to create, nurture, and
promote a system for building structures. I have been privileged to work alongside people
who are indeed titans of the industry and countless others also deserving recognition who
have demonstrated a passion for this work that continues to inspire.
The strength of this institute is most definitely in the experience, dedication, and perspira-
tion of so many volunteers. Its technical knowledge and focus are what make it so much
more than a mere trade organization. We have a strong institute and are financially sound
thanks to prudent management. We have an extremely strong and capable staff to help drive
the industry forward. We must not forget, however, that the heights we might reach and
things we might achieve will be the culmination of the work of the collective whole.
In his latest book, The Third Alternative, Stephen Covey writes about the power of seeing
beyond our selfish interests to the amplified results from synergistic collaboration. This, to
me, sums up PCI at its best. Our opportunities lie in continuing to make our voices heard,
listening to what others have to offer, and working together to keep precast/prestressed con-
crete in the vanguard of construction materials.
I attended a seminar recently about getting people fired up about formulating goals. Here
are some thoughts along those lines for your consideration as related to PCI involvement:
• Those of us on the downhill side of our careers, think about the importance of the
legacy we can leave behind.
• Those who have the majority of their careers ahead, make an impact to benefit the
global society by being on the cutting edge of technology, by pushing the envelope, and
by coming up with the means and methods to design and construct structures that offer
solutions to tough problems.
• Finally, business leaders, get involved and help shape the future of your industry.
We need to emphasize further establishing PCI as the preeminent authority on matters
related to precast/prestressed concrete, and we need to start thinking with a global perspec-
tive. Leveraging our technical brain trust for this and to affect the codes is paramount. We
must challenge ourselves to provide products and services so indisputably reliable that PCI
plant certification precludes consideration of any other alternatives, and we must further
expand our consideration of quality assurance by extending the effort to the field. We
should be able to ensure that the completed structure meets our design intent and standards
through competent inspections.
It is an honor and a privilege to serve in the position of chairman. I don’t take it lightly, and
I know I have big shoes to fill. I am thankful for the opportunity to try to make a positive
Greg Force | 2012 PCI Chairman difference. With the combined talents and efforts of this membership, I am confident of our
Tindall Corp. | Spartanburg, S.C. ability to do so. J
Something to
think about
Our industry has some important potential strategic advantages in the marketplace. But to
turn them into real advantages, we need to understand them fully and act on that under-
standing. I’d like to discuss two of them: high-performance systems and advanced materials.
• Sustainable: designed and built in a way that considers regional, global, and long-term
effects, intelligently using materials and energy to ensure the availability of sufficient
resources to future generations. This means, among other things, minimizing nonre-
newable materials, energy use, and the environmental footprint while maximizing the
Please send
correspondence to journal@
service life.
• Resilient: able to resist permanent damage from natural and accidental events (for
pci.org or Precast/Prestressed
example, severe weather, earthquakes, and physical impact) and still remain serviceable.
Concrete Institute, c/o PCI
This benefit is gaining importance, particularly in the wake of the highly destructive
Journal, 200 West Adams Street,
hurricanes and tornadoes of recent years.
Suite 2100, Chicago, IL 60606.
• Protective: more secure and comfortable for users and occupants because of structural
stability, fire resistance, blast resistance, noise isolation, or resistance to mold or rot.
Safety and security are difficult to measure but are increasingly desired as the world
seems more threatening.
• Constructive: designed and built to have positive functional, economic, and aesthetic
effects. This means, among other things, construction with minimal traffic disruption,
material handling, power and water use, dust, fumes, noise, and overall duration. It also
means a practical, beautiful structure that harmonizes with its environment.
Precast concrete structures are inherently strong in providing these attributes and PCI’s
marketing approach underscores this in many ways, but there’s much more we can do.
James G. Toscas, PCI President
A term that encompasses both sustainability and functional resilience is high performance.
We need to own it. “Designers Notebook: High Performance Precast Insulated Sandwich
Wall Panels” was published in the Summer 2011 issue of PCI’s Ascent magazine. However,
the broadest definition of high-performance is still vague at this point, and if we intend to
own it we need to define it.
To me, a high-performance system is one that has the versatility to meet a variety of needs.
High performance is the most general way to discuss many of the attributes we have been
emphasizing for years, including design flexibility, as touted by our “Discover the Freedom
of Precast” tagline.
“Freedom of Precast” was targeted specifically to architects, who value design flexibility but
often don’t associate it with precast concrete. Today, we need a core message that appeals to
all decision makers and influencers. High-performance systems could be it. If high-performance
systems can meet a wide variety of needs, then high performance can encompass a wide range of
capabilities. It would provide a common theme that suits a variety of times, regions, markets,
or targets and that logically underlies everything we say about precast concrete.
Why do we need to reach a broader range of groups? Today architects are not the only ones
deciding whether to use precast concrete. As sustainability becomes more important, we see
other stakeholders having increasing influence over materials and systems selection. Because
of resource limitations, we may decide, for now, to limit our marketing to one or two deci-
sion-maker or influencer groups, but the core message of our marketing should make sense
to anyone.
When the target audience changes, the focus may need to be changed—but the core message
should not. For instance, resilient and protective are two sides of the same coin: resilience pri-
marily benefits the owner, while protectiveness primarily benefits the user. Different focus,
but the core message—high performance—is the same.
This approach would fit well with PCI’s three-tiered marketing. The core message, which
would be promulgated nationally and even globally, would be the province of PCI itself.
National programs are best positioned to increase general awareness and spur interest.
The offshoots could be developed, customized to fit regional markets, and executed by our
regional affiliates with resources such as information, templates, and collateral materials
from PCI. PCI Producer Members can then be equipped and positioned to close the deal
based more on value and less on price.
PCI’s research and development efforts focus on practical results that can be immediately
applied, as they should. Who will conduct the research that has potential long-term payback
and would benefit precast concrete more than cast-in-place concrete?
We might have an answer to these questions right beside us. At the last meeting of the PCI
Foundation Board of Trustees in October 2011, Ken Lambla, dean of the School of Arts
and Architecture at the University of North Carolina at Charlotte, gave an exceptional and
thought-provoking presentation. After hearing status reports on the foundation’s highly suc-
cessful Architectural Design Studios, he tossed aside his prepared remarks and spoke off the
cuff. He outlined the huge potential of the Architectural Design Studio program and intro-
duced the idea of design research. Lambla was talking about the design of structures, but the
same concept could be applied to the design of concrete systems as well.
The purpose of the PCI Foundation is to develop people and technology to improve society
using precast concrete structures. Its key concept is to make precast concrete a better way to
build. Better than what? Of course, we’d like to think it is better than steel or better than
cast-in-place concrete, but I think it’s sufficient to say simply “even better than it is today.”
Design research is what we need to identify and develop new ways to build with precast
concrete, to make it even better than it is today. Whether we want to know how we can use
a new material or how we can create one, the foundation may be the right research vehicle,
but finding new solutions won’t do any good if we’re not prepared to implement them. Our
industry needs to be ready to use new types of materials and build new types of structures.
Are we?
Such a future is possible for us, but we need to change the way people think. In the broadest
sense, this is the purpose of marketing.
The
cover story
Starting with this issue, we’re expanding on the theme by means of a cover story. The cover
story will connect the dots: research, design, construction, and performance. In some cases,
the cover story will highlight one of the peer-reviewed papers by showing how research has
been applied in the field. In other cases, the cover story will emphasize the design, construc-
tion, or performance of precast concrete and point back to the developments that helped
make it possible or show what we can learn from practical experience so that we do things
better in the future.
We can’t promise to include all of the stages in the process from beginning to end in every
issue, but we’ll include as many aspects as we can. We hope that the cover story will make
the academic papers more approachable and emphasize the importance of practical applica-
tions. We also hope to encourage—and maybe even inspire—academics to consider the end
use of their research as they plan their experimental work.
Please send
correspondence to journal@
This is our seismic issue, and the contents reflect the international nature of the topic. Our
cover story summarizes the findings of PCI’s New Zealand reconnaissance team on the per-
formance of precast concrete structures in the February 22, 2011, Christchurch earthquake.
pci.org or Precast/Prestressed
We also have a preliminary report on the performance of precast concrete structures in the
Concrete Institute, c/o PCI
March 11, 2011, Tohoku earthquake from PCI’s Japan reconnaissance team and a report on
Journal, 200 West Adams Street,
precast concrete structures in the earthquake that took place on February 27, 2010, off the
Suite 2100, Chicago, IL 60606.
Chilean coast from PCI’s Chile reconnaissance team. All three of these reports pertain to
structures built in accordance with modern building codes that were properly enforced. The
lessons learned by these and other teams of structural engineers will help to make our own
building codes that much better.
This issue also contains a note of appreciation to those who provided peer reviews of manu-
scripts being considered for publication in these pages. Whether the reviews resulted in
acceptance, rejection, or major revision of the manuscripts, this publication is much better
because of them. Thank you all!
Rachel J. Detwiler, PhD, P.E.
Current and former PCI staff members attend Dan Jenny’s 90th birthday celebration. From left are George Nasser, Brian
Goodmiller, Sidney Freedman, Dan Jenny, Paul Johal, and Jason Krohn. Courtesy of .Dan Jenny.
Level I/II
January 23–25, 2012 May 2012
World of Concrete, Las Vegas, Nev. Nashville, Tenn.
Level III
May 2012
Nashville, Tenn.
CFA/IES
January 23–25, 2012 May 2012
World of Concrete, Las Vegas, Nev. Nashville, Tenn.
30 W in t e r 2 0 1 2 | PCI Journal
James H. Gilbert
James H. Gilbert, PCI Fellow and president from 1965 to 1966, died September 4,
2011. He was 91.
After obtaining a BA from Dartmouth College, Gilbert joined C. W. Blakeslee
and Sons Inc., a subsidiary of the Westinghouse heavy construction division, in 1946.
Subsequently, he rose to become vice president, president, and chief operating officer
of the firm.
In 1958, Gilbert started the Prestressed division at Blakeslee. That same year, he
became a member of PCI. Gilbert was named chairman of the Blakeslee board in 1974.
He retired from Blakeslee in 1976 and then served as business advisor and special con-
sultant to the Heavy Construction division of Westinghouse Electric Corp. Gilbert
also played an important role in developing Precast Systems Inc. (PSI) and served as the first PSI president.
In addition to serving on the PCI Board of Directors, he was president of the Connecticut Society of Civil
Engineers. He was active in developing the PCI Plant Certification Program, and his firm served as a charter
certified plant. Gilbert was also chair of the 1974 FIP/PCI Congress Organizing Committee charged with over-
all administration of the New York, N.Y., congress under the direction of the PCI Board of Directors. In 1994,
Gilbert was named one of PCI’s first 25 fellows in recognition of his contributions to the precast/prestressed con-
crete industry and PCI.
32 W in t e r 2 0 1 2 | PCI Journal
First CarbonCast application used in Canada
International Precast Solutions LLC of River Rouge, Mich., a member of The Prestressed
Group and AltusGroup, produced and erected the first CarbonCast Enclosure System in
Canada at the new Cineplex Galaxy. CarbonCast High Performance Insulated Wall Panels
were used in the project.
International Precast Solutions is licensed to produce CarbonCast High Performance
Insulated Wall Panels, CarbonCast Insulated Architectural Cladding, and CarbonCast
Architectural Cladding after becoming the 14th producing member of AltusGroup in
August 2011. It is one of 30 manufacturing facilities nationwide that share more than
400 specification-oriented sales, marketing, and engineering professionals operating PCI-
certified plants.
—Sources: AltusGroup and International Precast Solutions
34 W in t e r 2 0 1 2 | PCI Journal
the search for excellence
200 West Adams Street 200 West Adams Street 200 West Adams Street
Suite 2100 Chicago, IL 60606 Suite 2100 Chicago, IL 60606 Suite 2100 Chicago, IL 60606
Phone: 312-786-0300 Phone: 312-786-0300 Phone: 312-786-0300
Fax: 312-621-1114 Fax: 312-621-1114 Fax: 312-621-1114
www.pci.org www.pci.org www.pci.org
Beach Prisms reduce ing waves before they reach the shoreline. Waves pass
through the slots in the triangular 4 ft × 4 ft × 10 ft (1.2
shore erosion near m × 1.2 m × 3 m) structures.
Chesapeake Bay home Several years of working through the permitting
process led to a final approval meeting in front of the
On September 1, 2006, Hurricane Ernesto roared up Virginia Marine Resources Commission (VMRC). John
Chesapeake Bay and nearly cut Bob and Chris White’s Klein, a consulting coastal engineer working with SMC,
bay-front property in Virginia in half. presented statistical data to the VMRC representing 20
“I can’t believe how powerful the water was,” Bob years of Beach Prism installation and testing. The board
White says. “Our home office was knocked off its foun- voted 11 to 0 in favor of the permit.
dation, the pool was full of mud and debris, and the Tractor trailers traveling from the SMC plant trans-
water had risen over our tennis court and was threaten- ferred the Beach Prisms in five loads to a barge. The full
ing the house.” Subsequent storms have continued to do installation consisted of 22 pieces of Beach Prism with 6
additional damage. in. (150 mm) spacing for a total of about 230 ft (70 m)
The Whites tried several methods to solve their of protection. Each piece was placed parallel to the beach
problems: bulkheads, stone revetments, and five wooden in about 1 ft (0.3 m) of water at low tide. Installation was
groins. More than four years ago, the Whites discovered completed the next day. The Whites hope that this will
Beach Prisms, made by Smith-Midland Corp. (SMC). provide some protection from storm damage and slow or
Beach Prisms are an open precast concrete product stop the erosion.
that works by reducing the amount of energy in incom-
A Virginia couple has Beach Prisms installed in their home on Chesapeake Bay to reduce beach erosion. Courtesy of Smith-Midland Corp.
This precast concrete T-wall in southern Iraq is shown following an impact from a
107 mm (4.21 in.) insurgent rocket about 3 ft (1 m) away. The crater is indicated with an
arrow. Courtesy of MSG James Doten, Minnesota Army National Guard.
This Hesco barrier in southern Iraq is shown following an impact from a 107 mm (4.21 in.)
insurgent rocket about 3 ft (1 m) away. Courtesy of MSG James Doten, Minnesota
Army National Guard.
Approved Provider
PCI is a registered continuing education provider with the American Institute of Architects, the National Council of
Examiners of Engineers and Surveyors, and the U.S. Green Building Council, and offers continuing education credit
accepted in all 50 states! PCI provides several options for you to engage in continuing education programs including
webinars, seminars, lunch-and-learns, and the new PCI eLearning Center. To learn more, visit www.pci.org/education.
Webinars
200 West Adams Street
Suite 2100 Chicago, IL 60606
Phone: 312-786-0300
Fax: 312-621-1114
www.pci.org
200 West Adams Street
Suite 2100 Chicago, IL 60606
Phone: 312-786-0300
Fax: 312-621-1114
www.pci.org
200 West Adams Street
Suite 2100 Chicago, IL 60606
Phone: 312-786-0300
Fax: 312-621-1114
PCI webinars are presented live each month by industry experts on a variety of topics, from design and construc-
www.pci.org
tion to sustainability and more. All webinars are FREE, one-hour long, and scheduled to provide a noon start time in
each time zone in the contiguous United States. Webinars provide an inexpensive way to keep up to date on new
and improved materials, products, concepts, and more while earning continuing education credits.
Seminars
PCI and its regional affiliates offer seminars all over the United States on a variety of topics. These seminars include
Quality Control Schools, Structural Design Seminars, Architectural Precast Seminars, lunch-and-learns, and more.
Visit www.pci.org/education for up-to-date seminar listings, additional information, and to register.
200 West Adams Street I Suite 2100 I Chicago, IL 60606-5230
Phone: 312-786-0300 I Fax: 312-621-1114 I www.pci.org
U C AT IO
ED
N
NUING
CON
Phone: 312-786-0300 I Fax: 312-621-1114 I www.pci.org
Seismic
Preview of PCI’s New Zealand Earthquake 42
Reconnaissance Team Report
Robert B. Fleischman, José I. Restrepo, Joseph R. Maffei,
and Kim Seeber
Preview of PCI’s
New Zealand earthquake
reconnaissance team report
Precast concrete structural Emulative design, which requires critical sections to re-
design in New Zealand spond under load as they would for cast-in-place concrete,
together with the capacity design philosophy, was crucial
recast concrete has been used in New Zealand for gaining the acceptance of structural engineers in New
since the early 1960s and became widely accepted Zealand. In capacity design, the primary structural system
in the 1980s as an effective construction method for resisting seismic forces is first designed to have a
for seismic resistance. Since then, precast concrete suitable mechanism of accommodating nonlinear lateral
has been extensively used there. Park described the evolution deformation, with the engineer explicitly defining the loca-
of precast concrete design in that country.1 tions of plastic hinges. Plastic hinge regions are designed
and detailed for strength and ductility during a severe
During the building boom of the mid- to late 1980s, the speed earthquake. Then the remainder of the structural system is
of construction and high quality of precast concrete gave it provided with sufficient strength to avoid other possible
significant advantages over cast-in-place concrete. The New failure modes and maintain ductility after flexural over-
Zealand concrete design standard2 of the time included com- strength develops at the plastic hinges.
prehensive provisions for the seismic design of cast-in-place
concrete structures but did not cover all aspects of precast Resisting lateral forces
concrete structures. A study group of the New Zealand from earthquakes
Concrete Society was formed in 1988 to summarize data on
precast concrete design and construction, recommend best Moment-resisting frames have been the most common
practices, and identify areas for further research. A 1995 revi- seismic-force-resisting system for tall buildings in New
sion of the New Zealand concrete design standard3 included Zealand. A beam sidesway mechanism, in which plas-
new provisions for precast concrete structures based on the tic hinges form in the beams due to strong columns and
research of the early 1990s.1 weak beams, is the preferred nonlinear mechanism. To
ensure that flexural hinging does not occur at locations not
Earthquake
of February 22, 2011
New Zealand lies on the Pacific Rim and has a history Figure 1. PricewaterhouseCoopers building under construction in Christchurch,
New Zealand.
of strong earthquakes. However, Christchurch is about
240 km (150 mi) from the nearest known major fault, and
thus was not expected to experience a large earthquake. It Case study:
is in a moderate seismic hazard zone comparable to that of PricewaterhouseCoopers
Portland, Ore. The February earthquake, along with a less- office building
damaging earthquake the previous September, indicates the
existence of fault lines much closer to Christchurch. The 21-story PricewaterhouseCoopers office building was
constructed in the 1980s (Fig. 1). It was designed with
In the February 22 earthquake, an interval of extremely an exterior moment-resisting frame and constructed us-
high-intensity shaking lasted approximately 8 sec. The ing a detail in which the precast concrete joint (Fig. 2) is
average response spectrum of the ground motions recorded placed over the column below. Vertical ducts in the precast
in the city center was approximately equivalent to the concrete joint allow the vertical reinforcement from the
Christchurch maximum considered earthquake and was column to be threaded through and into the column above.
stronger for substantial ranges of typical building natural The ducts are grouted once the joint is in place.1
frequencies. For individual records, the ground motion
contained frequency ranges where it was substantially Figure 3 shows the perimeter moment-resisting frame
stronger, similar to the maximum considered earthquake in (background) and the interior “gravity” frame (foreground)
Berkeley, Calif.6 The February event was extremely strong following the earthquake of February 22. As can be seen,
near the epicenter, with vertical accelerations of up to 2g. the columns are more closely spaced in the perimeter
Soil liquefaction was significant and widespread due to the frame.
presence of a high water table in sandy soil. Deep founda-
tions are fairly rare in Christchurch, and several buildings A visual survey of this building by the authors showed a
and bridges showed evidence of support movement or dif- pattern of damage consistent with capacity design (Fig. 4).
ferential settlement. The lowest two floors, situated above a parking garage and
undergoing less deformation, showed no plastic hinging.
The lower- to midlevel floors showed the greatest beam de-
Final report
Figure 3. Corner joint in first-level parking podium area after February 22, 2011, earthquake in New Zealand. Photo courtesy of Franz Kahn.
of which suffered serious damage, will be published in a 5. The Canterbury Earthquakes Royal Commission.
future issue of the PCI Journal. canterbury.royalcommission.govt.nz.
Preview of PCI’s
Japan Earthquake
Reconnaissance Team Report
he authors formed the PCI reconnaissance team The tsunami inundated many locations along the north-
that visited selected locations (Fig. 1) in Japan east coast of Honshu. Depending on offshore topography
from June 27–30, 2011, following the Tohoku and distance from the epicenter, the depth of water above
earthquake of March 11, 2011. grade on shore varied from 1 m (3.3 ft) to 8 m (26 ft) in
the coastal plain region of Natori to more than 18 m (59
Tohoku earthquake ft) in the Ria coastal town of Onagawa.3 A tsunami wave
typically crests before reaching the shore, creating a bore,
At 2:46 p.m. on March 11, 2011, a magnitude 9.0 earthquake which can move rapidly and produce impulsive forces on
occurred near the east coast of Honshu, the main island of Ja- structures in its path. The subsequent inundation produces
pan.1 This earthquake, known as the 2011 Tohoku earthquake, high-velocity water flow around and through structures.
caused strong ground shaking throughout much of eastern The effects of these forces on structures were widespread
Honshu and triggered tsunami waves that caused widespread in the tsunami-affected coastal regions.
damage at many locations along the northeast coast of Japan.
Earthquake-resistant
The epicenter of the earthquake was located at 38.297° building design in Japan
N, 142.372° E,1 about 70 km (44 mi) off the coast. Strong
ground shaking occurred in Sendai, the major city (130 km The lack of widespread significant structural damage to
[80 mi]) nearest the epicenter. Peak ground accelerations buildings from ground motions associated with the 2011
exceeding 2g were recorded at locations near the east coast Tohoku earthquake has been noted by other observers.2
of Honshu.1 Some ground shaking–induced damage was This result has been attributed to important revisions over
observed even in the Tokyo metropolitan region, more than the past 40 years to the Japanese national building code
350 km (210 mi) southwest of the epicenter; 1 however, wide- and related national standards. The Japanese national
spread structural damage was not observed across eastern code, the Building Standard Law of Japan (BSLJ), which
Honshu.2 includes the Building Standard Law, the Enforcement
The 1968 Tokachi-oki earthquake caused significant dam- The authors noted a lack of widespread earthquake damage
age to buildings, and a revision to the BSLJ5 reduced the to buildings in the affected areas, including Sendai, where
spacing of steel ties in reinforced concrete columns to 100 the recorded peak ground acceleration on rock exceeded
mm (4 in.). In 1971, a major revision of the AIJ standard 0.34g. The authors observed some damage to conventional
for reinforced concrete7 incorporated ultimate-strength reinforced concrete buildings at the Aobayama Campus of
design for shear of beams and columns, including more Tohoku University, which is located in a hill zone above
stringent shear reinforcement requirements. These changes the city of Sendai.
are comparable to important code changes in the United
States following the 1971 San Fernando earthquake in The authors observed the exteriors of several precast
California. Post-1971 reinforced concrete structures per- concrete structures, all of which performed well during
formed much better in the 1995 Kobe earthquake than their the earthquake with no indication of significant damage.
pre-1971 counterparts, primarily because of the improved Several of these structures are discussed in the following
shear design of columns.8 paragraphs.
The 1978 Miyagiken-oki earthquake caused significant The Shiogama Municipal No. 2 Junior High School
damage to buildings and led to a 1981 revision of the (Fig. 2) was retrofitted in 2010 using an external precast
BSLJ,6 which introduced two-phase earthquake-resistant concrete frame braced with tension ties. The existing
design. The first-phase design (essentially the allowable cast-in-place concrete building was deficient in torsional
stress design from the previous BSLJ) is intended to pro- resistance to earthquake loading. The north facade had stiff
tect a building against loss of function in ground motions shear walls, while the south facade had moment frames,
expected to occur several times during its lifetime, with creating a torsional stiffness irregularity. The tension-tie-
peak ground accelerations in the range of 0.08g to 0.10g. braced precast concrete frame increased the lateral stiff-
The second-phase design is intended to ensure safety under ness along the south facade. The pieces included precast
a ground motion expected to occur once in the lifetime of concrete columns anchored to the foundation using type II
a building, with peak ground accelerations in the range of mechanical splices, architectural medallions on top of the
0.3g to 0.4g. Post-1981 structures designed by the two- columns, and drop beams between the medallions post-
phase procedure performed well in the 1995 Kobe earth- tensioned together on-site. To enhance the rigidity of the
quake. precast concrete frame, X-bracing with steel rods covered
with plastic sleeves was used in the openings. Cast-in-
Current earthquake-resistant design of reinforced concrete place concrete diaphragms (Fig. 3) between the existing
buildings in Japan is consistent with the 1981 BSLJ6 and structure and precast concrete frame were used to transfer
the 1971 AIJ standard for reinforced concrete.7 lateral shear forces to the external precast concrete frame.
No damage to this structure was observed.
Figure 5. Minor spalling at the base of the columns of Namagachi Lions Tower.
Jason Lien is vice president of Code, earthquake, Japan, preview, seismic, tsunami.
design for the corporate office of
EnCon. He is on the Technical Reader comments
Activities Coouncil and the
Continuing Education Committee. Please address any reader comments to journal@pci
He is chair of the Building .org or Precast/Prestressed Concrete Institute, c/o PCI Jour-
Information Modeling Committee. nal, 200 W. Adams St., Suite 2100, Chicago, IL 60606. J
He is a registered Professional Civil Engineer in
Colorado, California, New Mexico, Wyoming, and the
Territory of Guam.
Observations
from the February 27, 2010,
earthquake in Chile
The 2010 earthquake that is the subject of this paper struck Region Offshore Maule, Chile
in an area previously identified as a seismic gap extending 60 mi NW of Chillán, Chile
from Constitución in the north to Concepción in the south
with a projected worst-case potential to produce an earth- 65 mi NNE of Concepción, Chile
quake with M between 8.0 and 8.5.2 The rupture extended Distances
beyond the northern and southern boundaries of the gap, 70 mi WSW of Talca, Chile
overlapping extensive zones already ruptured in 1985 and
1960.3 210 mi SW of Santiago, Chile
Figure 1. The source of the Chile earthquake is at the convergence of the Nazca and the South American plates. Source: Roberto Leon presentation at www.eqclear
inghouse.org/20100227-chile/wp-content/uploads/2010/04/Leon-Chile-Earthquake.pdf. Note: 1 mi = 1.61 km.
Figure 3. Accelerograms and corresponding acceleration and displacement response spectra (β = 5%) from downtown Concepción, Chile. Source: Boroschek et al.
2010. Notes: g = acceleration due to earth’s gravity; Sa = spectral acceleration; Sd = spectral displacement; β = damping coefficient. 1 cm = 0.4 in.
Purlins
Edge channel Gable-shaped girder
Column
Figure 10. Framing system that uses long-span gable beams on columns and is clad with double-tee walls in an industrial building with offices.
over the purlins, but there did not appear to be connections the other end, just inside the two-level bay with offices at
between planks to form a diaphragm. There was no diagonal the front of the building. There were interior walls across
bracing in the plane of the roof deck. The wall cladding was the building at the first bay that included two floors of of-
connected to the precast concrete roof using long threaded fice space, but this did not provide effective bracing to the
rods that connected to the channels on the sides and the framing at the distant end. The column-to-girder connec-
girders at the ends. These connections appeared to provide tions apparently provided sufficient continuity for frame
out-of-plane resistance but not a load path to transfer lateral behavior in the direction of the frame. Figure 11 shows a
forces into the plane of the wall cladding. sketch of the building plan.
The shear walls for these buildings were isolated cast-in- The team learned from the owner’s representative that
place concrete walls in two bays on each side. These bays one of the buildings had soil saturation problems during
were the last short bay at one end and the second bay from construction that required soil improvement to a depth of
Shear wall
Shear wall
Purlins
Girder
Figure 11. Roof framing plan used in an industrial building with offices. Note: All measurements are in meters. 1 m = 3.28 ft.
Figure 13. Rear elevation of the same office and warehouse building showing Figure 15. Bent and broken connection angle that failed to hold walls to edge
fallen exterior walls and exposed transverse walls and end columns. beam at roof.
4.6 ft (1.4 m). The other did not. The structure without soil the front, with column spacing at about 20 ft (6 m) and
improvement dropped the precast concrete purlins and roof stairways every 80 ft (24 m). Transverse walls separating
in the two bays with cast-in-place concrete shear walls. the units were spaced at 40 ft (12 m), and the length from
The other building did not suffer damage. It appears that front to back was from 66 ft to 82 ft (20 m and 25 m). The
the lateral bowing in the roof girders caused a failure at the rear was enclosed by vertical precast concrete walls with
pins and loss of bearing for the purlins. loading docks and doors. Figures 12 and 13 show photos
of the collapse.
Although many precast concrete buildings constructed
using shear walls have performed well in past earthquakes, Although there appeared to be an ample number of walls
the LFRS requires a complete load path that ties all compo- to provide lateral bracing for the structure, failures likely
nents together. In this case, it appears that the roof framing occurred because of inadequate connections and a lack of
with spaced purlins and without connections between roof seismic detailing for strength and ductility. Most of the
planks lacked a diaphragm. Failures occurred at the pinned walls that clad the exterior of the stairs on the front eleva-
purlin bearings where movement between the roof and the tion fell away from the structure because of out-of-plane
supporting girders was not sufficiently restrained. forces that caused the connections to fail. Failed gravity
columns showed bar bucking and a lack of confinement
Another combined office and warehouse consisting of a (Fig. 14).
precast concrete building with transverse and longitudinal
walls suffered major local failures and partial collapse. On the rear elevation, the tops of the walls were connected
The structure had two-level beam and column framing on to the spandrel girders at the roof through a thin angle that
spanned between embedments in the walls and girders. bays. The framing included precast concrete girders in both
Figure 15 shows a remnant of bent and broken angle still the longitudinal and transverse directions, bearing on top of
welded to embedments in the back of the fallen spandrel precast concrete columns with wet-cast joints. Figure 17
panel. The welds to the plates embedded in the spandrel shows one interior joint. The roof sloped from the sides
girder were torn loose. Figure 16 shows that the continu- to the center girder line. The roof deck was supported be-
ous angle allowed the welds to be made even if the plates tween the girders with simply supported precast concrete
in the walls did not align with the plates in the spandrel joists. The roof plane, in lieu of a diaphragm, had diagonal
girder. The out-of-plane forces, however, caused bend- bracing below the roof joists that was connected between
ing in the angle and prying on the welds. There were also girders to plates and gussets. The diagonal bracing was
locations that appeared to have field-installed anchor rods also composed of precast concrete joists, similar to those
in the walls that projected into the cast-in-place concrete in Fig. 18. The only significant damage to this structure
topping over the roof. Walls had shallow breakout cones from the earthquake was that all of these diagonal braces
that appeared to correspond to short lengths of bent dowels disconnected at the plates at the girders and fell.
projecting from the edge of the topping. In addition, the
walls were relatively thin, about 6 in. (150 mm), and the The girders were formed as I-shaped sections with rect-
wall reinforcement comprised at least three sizes of mild angular end blocks, similar to bridge girders used in the
steel reinforcement no larger than no. 3 (10M) bars.
Moment-resisting frames
Figure 19. Total–precast concrete school framing system used in more than 40 There were several examples of industrial buildings with
schools that experienced no damage from the earthquake. precast concrete framing supported by cantilevered col-
umns as the LFRS.
their bearings on the roof girders, so the secondary framing were fixed to the precast concrete columns with exterior
is made continuous. This detail adds some redundancy to plates bolted around the columns. The earthquake caused
the roof system and contributes to the overall structural the canopy to sag and the columns to crack near the beam
integrity. The spread-beam system, however, does not form connections. The cantilevered behavior of the columns
a continuous diaphragm capable of redistribution of the resulted in flexural cracks near the bottoms of the columns.
forces. The building suffered no structural damage to the The team also found that there were some local spalls at
primary LFRS. the bearing of a roof beam where the width of the beam
spanned across the joint between the ends of the roof gird-
This building was clad to about two-thirds of the exterior ers. The observed damage in these areas was not severe
wall height with horizontal precast concrete walls. These and was being repaired.
walls were not intended to act as shear walls, and some of
the upper panels fell from the structure as the cladding con- The design of cantilevered column systems was shown to
nections failed. This aspect is discussed in the section on be effective.
precast concrete cladding.
Advanced seismic-force-resisting
Another example of precast concrete framing with cantile- systems
vered columns was found in Coronel, south of Concepción,
at the Parque Industrial Escuadrón, adjacent to the failed There were several examples of framing systems using
gable frames described earlier. This recent construction advanced concepts that proved effective during the earth-
also used long-span gable-shaped roof girders and spread- quake.
beam framing to form a warehouse for the fish meal opera-
tion. The gable-shaped girders are pierced with round holes Base-isolated offices As a demonstration of lami-
in the webs to reduce their weight. Again, the trapezoidal nated base-isolation rubber bearings, the manufacturer con-
spread roof beams were made continuous across these gird- structed a total–precast concrete office building supported
ers with cast-in-place concrete joints. One section of this on slide bearings at the corners and on base isolators at the
building was about 52 ft (16 m) tall, with beams framing interior bays. Figure 22 shows a view of the isolation bear-
with pinned end connections at two levels above and below ings on one side.
the girders. This building survived the earthquake without
damage. The building is two bays by five bays, with a square
module of 26 ft (8 m). The structure is only two stories
A second example in Coronel was found at a manufactur- tall, but the company has supplied isolation bearings to
ing facility. The framing was similar to that of the fish meal fifteen other buildings with similar design. The structure
facility. Figure 21 shows an interior view of this building experienced no damage during the earthquake, though the
showing the gable-shaped girders and spread trapezoidal slide bearings showed movement of about 5 in. (130 mm)
roof beams. Field-cast joints created continuity in the roof in both longitudinal and transverse directions. The owner
beams across the girders. reported that books standing on end in the structure did not
fall over.
This structure was damaged in three areas. In one loca-
tion there was a long exterior cantilevered canopy over a It was reported that these bearings were used in some
loading area that was framed with steel beams. The beams residential buildings, some bridges, and at buildings at the
The clear height under the roof girders is 40 ft (12 m), and
they are 5 ft (1.5 m) deep. Roof beams span between the
girders to support the roof decking. Because of the long
spans of the girders, there is additional support provided
by cable stays. To hold these stays above the roof, columns
were added above the top intersection of the tilted braced
Figure 23. Interior view of a convention hall with tilted braced frames.
frame columns; these added columns lean outward. The
joints tying the lower and upper columns together were
made with field-cast concrete. Figure 24 shows an exterior
view of the cable stays, braced frames, and columns.
Catholic University and the University of Chile. They were The only damage to this building from the earthquake
used in a dock at the port of Coronel, which was reported was from failure of the cladding panel connections. It was
to be the only dock not damaged by the earthquake. reported that the top panels pulled out of plane and fell off
the structure. Some of the panels shown in the photograph
Unbonded prestressed concrete frames and were replaced temporarily while waiting for new panels
walls A precast concrete manufacturer constructed a five- with additional connections to be fabricated.
story structure at its convention/exhibition site that uses
unbonded post-tensioned walls and frames following the Precast concrete stadiums
research of the PCI PRESSS (Precast Seismic Structural
Systems) program. The structure is braced in the short The PCI team investigated stadiums framed with precast
direction by post-tensioned shear walls placed at the ends concrete columns, beams, rakers, and risers in Chillán and
of the building. The post-tensioning strands are located in nearby Chillán Viejo.
near the center of the walls. In the other direction, there are
three bays framed with unbonded post-tensioned moment- The stadium in Chillán was framed with precast concrete
resisting frames. Although the erection of the structure was for the seating areas on four sides enclosing the playing
complete, the building was unfinished. The first floor was in field. The seating was shaded with a fabric roof within
operation as a kitchen for the convention center. The upper steel frames supported by cantilevered steel columns
floors were not yet completed. The structure experienced no attached to the tops of precast concrete columns on the
damage from the earthquake. perimeter; these were braced by the raker beams and
transverse beam framing. The project was planned on a
tight schedule, and when the precast concrete manufacturer
In Chile, precast concrete cladding panels are used on At the convention center complex described previously, the
industrial buildings and on some low-rise office buildings. exhibition and convention halls were clad with long hori-
Precast concrete cladding was not observed on high-rise zontal precast concrete panels. The panels were stacked
structures. Most of the panels observed performed well. two to four units high, with the primary weight transferred
There were some cases where connections between the through the panel below and then to the foundation. One of
cladding and the supporting structure failed. the buildings had sloped precast concrete columns, creat-
ing a braced frame, so that part of the panel weight was
Most of the precast concrete cladding panels observed were carried by its connections to the columns (Fig. 23 and 24).
nonstructural. These panels were subject only to inertial The end annex to that building (Fig. 27) and the adjacent
seismic forces and wind loads. For effective support, the exhibit hall had vertically stacked panels. In neither case
connections of these panels should accommodate move- did the precast concrete walls reach the level of the roof.
ment of the supporting structure. Without this flexibility, Metal cladding was used to complete the enclosure.
cladding panels can attract unintended forces. Examples of
successful and unsuccessful performance were observed. The wall panels were attached to the concrete columns
with long slotted embedments (Fig. 28). The embedments
There were several warehouse-type buildings using vertical are commercial inserts commonly used for precast con-
double-tee wall panel cladding that performed well. These crete connections. The slotted inserts are oriented vertically
Precast concrete
bridge construction
Figure 31. Bridge section at the Bio Bio River crossing with partial diaphragms, vertical restraint bars, and end stoppers.
bearing was not sufficient for the longitudinal displacement The abutment bearing was cast as a level ledge, and then
between the piers. This was one of several failed multispan the bearing blocks with varying thicknesses were placed
bridges that dropped the ends of spans. to form the final bearing elevations. This convenient
method used to construct the correct bearing elevations
Figure 30 also shows an elevated bearing surface that is lacks the lateral restraint of bearing details using pockets.
a reminder that actual construction geometry is often not The earthquake forces caused lateral displacement and the
as simple as the illustration of the geometry shown in the failure of the stopper at one end of the pier. The failure did
earlier figures. not indicate a lack of reinforcing, though the horizontal
confinement did not appear to contain all of the vertical
Figure 31 shows another bridge section that includes reinforcement. The design, however, imposed all the re-
the center crown and cross slope for drainage that must quirement of lateral restraint on the end stopper. Figure 33
often be included. It is common for the beam bearings shows the failed stopper at the abutment. This view also
to be stepped to provide the crowned shape or to provide shows a wide crack in the bottom of the edge girder from
for cross slope or superelevation in the roadway above. the impact with the stopper.
These features of practical geometry can result in the loss
of lateral restraint to the girders; this was observed in the A bridge in Santiago designed with vertical restraint bars
Llacolen Bridge. The second bridge was a multispan bridge attached to precast concrete girders and bridge pier showed
crossing an inlet of the Bio Bio River. no evidence of displacement. Damage at this bridge was
seen in subsidence of the fill that formed the bridge ap-
This bridge had some of the characteristics of more recent proach and embankment at the grade separation.
bridge construction details but did include partial-depth
diaphragms between the girders at the bearings. Figure 31 Building code
also includes the diaphragms, the vertical restraint bars,
and the end stoppers. Figure 32 shows a view of the bear- Buildings and other structures in Chile must be designed
ing at two interior girders. and constructed in compliance with the Chilean Building
Code NCh433-2009.6 This code is applied in addition to
the specific design code for each of the materials and aims
to achieve structures that meet the following objectives:
R Cmax
2 0.90SA0 /g
3 0.60SA0 /g
4 0.55SA0 /g
5.5 0.40SA0 /g
6 0.35SA0 /g
7 0.35SA0 /g
NCh433-20096 has four building categories: A, B, C, The seismic coefficient C is obtained from Eq. (2).
and D. These are comparable to ASCE 7/05 occupancy
2.75A0 a T lk
categories IV (essential facilities, hazardous facilities), III n
• Apply capacity design concepts to determine axial 1. The whole flange width of a flanged section (T, L, C,
force and shear, considering the effect of connection or other cross-sectional shapes) must be considered in
with the slabs. calculating combined flexural and axial load strength.
• Add slenderness restrictions to avoid transverse bend- 2. The contribution of the total amount of longitudinal
ing of boundary elements and the panel. For this, study reinforcement must be considered in determining
the New Zealand19 and Canadian20 code recommenda- combined flexural and axial load strength.
tions.
3. Longitudinal reinforcing bar diameter must be less
• Revise the displacement spectrum. than or equal to one-ninth of the least dimension of
the boundary element (wall thickness) (Fig. 34).
• Avoid adding to the resistance side; add to the demand
side. Study the effect of resistance (strength) on dis- 4. Transverse reinforcing bar diameter must be greater
placement demand. than or equal to one-third of the diameter of the re-
strained longitudinal bar (Fig. 34).
• Optional displacement-based design has been intro-
duced in annex B. 5. Transverse reinforcement must be anchored to ex-
treme longitudinal bars in a wall.
Patricio Bonelli, a professor at Universidad Técnica Fed-
erico Santa María, also shared the following recommenda- 6. Standard hooks must be used with transverse rein-
tions and reflections: forcement as defined in section 7.1 of ACI 318-08:
135° or 180° bend plus 6db extension, but not less than
• Carefully study the demands of displacement and 3 in. (75 mm) at the free end of the bar. In ACI 318-
rotation. In Concepción, 15 cycles of 0.2g acceleration 08,16 a standard hook is defined as a 180° bend plus
were observed at a period of 1.5 sec. 4db extension or a 90° bend plus 12db extension. Thus,
these two requirements are contradictory. The second
• Microzonation of cities is desirable. requirement (135° or 180° bend plus 6db extension),
being more restrictive than the first, is assumed to 11. Transverse reinforcement through the length of lap
govern. splices of longitudinal bars in walls must satisfy the
condition in Fig. 35.
7. In special structural walls, the net tensile strain in the
extreme tension steel εt must be equal to or greater 12. In section 21.9.6.2 of ACI 318-0816 on boundary ele-
than 0.004 when the concrete in compression reaches ments of special structural walls, the lower-bound limi-
its assumed strain limit of 0.003. (Section 10.3.5 of tation of 0.007 on δu /hw in Eq. (21-8) shall not apply.
ACI 318-08 requires εt at nominal strength to be not
less than 0.004 for nonprestressed flexural members 13. Section 21.9.6.4(a) of ACI 318-0816 may be replaced
and nonprestressed members with factored axial com- by the following:
pressive load less than 0.10 fclAg .)
The boundary element shall extend horizontally from
8. The transverse dimension (thickness) of special struc- the extreme compression fiber a distance not less than
tural walls must be greater than or equal to one-six- cc determined from Eq. (3).
teenth of the lateral unsupported member length under
cc c 1
compression, ℓu /16. = - (3)
,w ,w dul
600
9. In special structural walls, when shear has not been hlw
calculated using capacity design rules, the maximum
shear obtained from design load combinations that When this option is applied, the term δu/hw in ACI 318-
include load effects of earthquake E shall be calculated 0816 section 21.9.6.2 must be replaced by d'u /h 'w . The
with E assumed to be 1.4 times that prescribed by the lateral design displacement δu comes from NCh433-
legally adopted general building code for earthquake- 199611 (modified in 20096) section 5.9.5.
resistant design. ACI 318-08 section 21.3.3(b) lists the
comparable factor as 2, which is applicable to beams 14. Replace ACI 318-0816 section 21.9.6.2(b) with “The
and columns of intermediate moment frames. boundary element reinforcement shall extend vertical-
ly from the critical section a distance not less than ℓu.”
10. Transverse reinforcement through the length of lap
splices of longitudinal bars in walls must satisfy items 4, 15. Transverse reinforcement in boundary elements in
6, and 11 of this list when either (a) or (b) is true. walls, when required, must satisfy ACI 318-08 section
21.9.6.4 and (a) and (b).
a) T
he longitudinal reinforcement ratio, defined as
ΣAb/hs, is greater than 2.8/fy.
b) T
he cover of a longitudinal bar with nominal diam-
eter db is less than 2db.
b) W
here vertical reinforcement can yield, the maxi- The 2010 emergency changes to Chile's building code have
mum longitudinal spacing of transverse reinforce- far-reaching implications for the special structural wall de-
ment in the boundary shall not exceed the smaller sign provisions in ACI 318-0816 section 21.9. Changes will
of six times the longitudinal bar diameter and 8 in. be considered for possible inclusion in the next version of
(200 mm). ACI 318 and, if adopted, may have significant effects on
design practice in the United States.
Conclusion
Acknowledgments
The majority of structures performed acceptably or better,
considering the severity of the 2010 Chile earthquake. The A version of this report is under review for inclusion in a
number of deaths and the amount of property loss, while special edition of Earthquake Spectra to be published by
quite significant, were not disproportionate to the severity the Earthquake Engineering Research Institute (EERI) in
of the earthquake. Much of this is attributable to Chile’s June 2012. The guest editors are Jack P. Moehle and Da-
history of adoption and implementation of adequate build- vid Frost. This report appears here by permission of EERI.
ing codes.
The PCI team consisted of Ned Cleland of Blue Ridge
The PCI team concentrated on precast concrete structures. Design in Winchester, Va.; Susan Dowty of S. K. Ghosh
With the exception of the out-of-date gable frame system Associates Inc. in Aliso Viejo, Calif.; S. K. Ghosh of S. K.
observed in one location south of Concepción, the pre- Ghosh Associates Inc. in Palatine, Ill. (team leader); Ray
cast concrete building systems generally performed well. McCann a structural engineer in Napa, Calif.; and Dante
In some cases, the LFRS performed, but the absence or Sanguineti of Pomeroy Corp, in Perris, Calif. Augusto
weakness of diaphragm framing resulted in local failures. Holmberg of Instituto del Cemento y del Hormigón de
Where lateral forces were resisted by cantilevered columns Chile was the PCI team’s local contact and host. Patricio
and distribution of loads through diaphragm action was not Bonelli, a professor at Universidad Técnica Federico Santa
essential, the structural framing of the buildings did not María, provided much valuable information concerning
experience significant damage. Chile's building code and changes to that document. He
also led three of the team members on a tour of Viña del
Some of the buildings inspected showed the success of Mar. His help is gratefully acknowledged. The contribution
advanced precast concrete seismic systems, which reflects of the section on cladding by Ray McCann and review of
research conducted in the United States. Some used tech- the original version of this report by all team members are
nology associated with other materials on the U.S. market. much appreciated. The authors are grateful to Prabuddha
The example of the reinforced concrete braced frames Dasgupta of S. K. Ghosh Associates Inc., whose help in
showed the success of a system that is not included in the the preparation of the manuscript was invaluable.
defined systems in ASCE 7-10.8 The PCI team found a
mature and sophisticated precast concrete industry that has
successfully considered and solved problems of earthquake
7. EERI. EERI Chile Earthquake Clearinghouse. www 20. Canadian Standards Association (CSA). 2004 (reaf-
.eqclearinghouse.org/20100227-chile/. firmed 2010). Design of Concrete Structures. CAN/
CSA-A23.3-04 (R2010). Toronto, ON, Canada: CSA.
8. American Socity of Civil Engineers (ASCE) 7. 2010.
Minimum Design Loads for Buildings and Other 21. 2011. Decreto numero 117—Aproeba Reglamento que
Structures. ASCE 7-10. Reston, VA: ASCE. fija el Diseño Sísmico de Edificios [Decree number
117—approves regulation which establishes seismic
9. ASCE 41. 2006. Seismic Rehabilitation of Existing design of buildings]. In Diario Oficial de la República
Buildings. ASCE 41-06. Reston, VA: ASCE. de Chile [Official diary of the Republic of Chile].
Santiago, Chile.
10. Los Angeles Tall Buildings Structural Design Council.
www.tallbuildings.org. 22. INN. 2008. Hormigón Armado—Requisitos de Diseño
y Cálculo [Reinforced concrete—requirements for de-
11. INN. 1996. Diseño Sísmico de Edificios [Earthquake- sign and calculation]. Official Chilean Standard NCh
resistant design of buildings]. Official Chilean stan- 430. 2008. Santiago, Chile: INN.
dard NCh 433. 1996. Santiago, Chile: INN.
23. 2011. Decreto numero 118—Aproeba Reglamento
12. International Code Council. 1997. 1997 Uniform que fija los requisites de Diseño y Cálculo para el
Building Code. International Conference of Building Hormigón Armado [Decree number 118—approves
Officials: Whittier, CA. regulation which establishes the requirements for de-
sign and calculation of reinforced concrete]. In Diario
13. American Concrete Institute (ACI) Committee 318. Oficial de la República de Chile [Official diary of the
2005. Building Code Requirements for Structural Con- Republic of Chile]. Santiago, Chile.
crete (ACI 318-05) and Commentary (ACI 318R-05).
Farmington Hills, MI: ACI.
Ab = area of an individual longitudinal bar R = reduction factor ranging between 2 for structural
systems of limited ductility and 7 for ductile struc-
A 'b = area of an individual transverse bar tural systems in NCh433
c = largest neutral axis depth calculated for the factored Sa = spectral acceleration
axial force and nominal moment strength consistent
with δu Sd = spectral displacement
Cmax = maximum value of seismic coefficient T* = period of mode with the highest translational
equivalent mass in the direction of analysis
db = nominal diameter of longitudinal bar
β = damping coefficient
dbt = nominal diameter of transverse bar
δu = lateral design displacement
E = effects of earthquake
d'u = design drift measured between the top and the con-
fcl = compressive strength of concrete sidered level
fy = yield strength of longitudinal reinforcement εt = net tensile strain in the extreme tension steel
ℓw = length of wall
N
ew types of seismic-resistant building and bridge
structural systems (for example, shear wall, mo-
ment-resisting frame, bridge pier) using unbonded
post-tensioning tendons have emerged since the 1990s.1–5
The post-tensioning tendon anchorages in these structures
are among the most important components because the
entire post-tensioning force is transferred at these loca-
tions. While unbonded post-tensioning has been available
in the United States since the 1950s (for example, concrete
slab systems) and standards exist for its application,6–8 the
■ This paper presents an experimental investigation of unbonded use of this construction method in seismic zones is more
post-tensioning strand-anchorage systems subjected to a recent. Unbonded tendons result in a uniform distribu-
variety of loading conditions. tion of the strand elongations over the unbonded length,
thus reducing the maximum strand strains as the structure
■ The research focuses on the strand stresses and strains at wire deflects laterally during an earthquake. However, depend-
fracture inside the anchors, considering the following loading ing on the geometry and type of the structure, the intensity
parameters: loading rate, eccentricity between the strand ends, of the earthquake, and the initial strain (that is, prestrain)
postyield cyclic loading, and initial stress. level, maximum strains significantly greater than those
under gravity loading can develop in the strands during
■ In addition to the strand-anchorage experiments, updated seismic loading.
stress-strain relationships are proposed for the PCI Design
Handbook: Precast and Prestressed Concrete for 0.5-in.- For example, consider the hybrid precast concrete shear
diameter (13 mm) and 0.6-in.-diameter (15 mm) strands based wall described in Smith et al.4 Despite the placement of the
on experiments that achieve free-length fracture of the test post-tensioning tendons near the wall centerline to mini-
specimens. mize the strand elongations, the expected maximum strand
H H
BW
Boss
BID BID
BOD BOD
Two piece Three piece
H
TOD dmw
L dp
BID
dow
Cast anchor Barrel anchor
Nine-hole multistrand anchor Seven-wire strand
Figure 2. Materials tested in this research. Note: Diagrams are not drawn to scale. BID = anchor bottom inside diameter; BOD = anchor bottom outside diameter
(equal to TOD in case of barrel anchors); BW = wedge bottom outside width; dmw = diameter of middle strand wire; dow = diameter of outer strand wire; dp = outside
crown-to-crown diameter of strand; H = anchor or wedge height; IW = wedge inside width; L = cast anchor length; TID = anchor top inside diameter; TOD = anchor
top outside diameter (equal to BOD in case of barrel anchors); TW = wedge top outside width; W = cast anchor width.
tested with a monotonic machine position displacement rate • anchor wedges (two-piece or three-piece wedges with
between 0.197 in./min (5.00 mm/min) and 0.887 in./min or without binding rings)
(22.5 mm/min), which corresponds to a strand strain rate of
0.0047/min to 0.021/min for the minimum strand free length • ASTM A41617 seven-wire, uncoated, low-relaxation
of 3.5 ft required by ICC-ES. An extensometer with a mini- post-tensioning strand (0.5 in. [13 mm] or 0.6 in.
mum gauge length of 36 in. (910 mm) must be used within [15 mm] nominal diameters) with a nominal ultimate
the free-length of the strand to measure the strains. strength fpu of 270 ksi (1860 MPa)
The ICC-ES requirements also serve as the basis for the An important factor in the selection of each anchor con-
loading parameter tests in this paper, though most of these figuration for testing was the use of wedges and anchors
other procedures exceed the requirements of ICC-ES in that were compatible with each other and were provided by
vigor. The ICC-ES cyclic loading requirements, for exam- the same manufacturer.
ple, are provided for fatigue load tests. As described later
in the paper, the ICC-ES 50-cycle fatigue load upper stress Table 1 lists the geometric properties of the anchor speci-
limit of 0.85fpu (where fpu is the nominal strand strength), mens used in the loading parameter tests, named accord-
which roughly corresponds to a strand strain of 0.008, is ing to the anchor type and the nominal diameter of the
not high enough to observe the postyield cyclic behavior strand used with the anchor. Four monostrand cast anchors,
that may occur in strands under extreme seismic loading. two monostrand barrel anchors, one machined nine-hole
multistrand anchor (multi-0.6A), and one cast seven-hole
Test specimen properties multistrand anchor (multi-0.6B) representing three differ-
ent manufacturers were included in the investigation. As
The materials tested in this research (Fig. 2) consist of the described in the section titled “Multistrand Anchorages,”
following: the multistrand anchors were tested using only one hole
per test, as if they were monostrand anchors.
• monostrand cast anchors
Table 2 lists the tested wedges, which are named ac-
• single-use monostrand barrel anchors cording to the number of wedge pieces (2P for two-piece
wedges or 3P for three-piece wedges) and the nominal
• multistrand anchors (shown here as a nine-hole cylin- strand diameter (0.5 in. [13 mm] or 0.6 in. [15 mm]).
drical anchor but also tested with a seven-hole anchor)
Anchor H, in. L, in. W, in. TOD, in. TID, in. BID, in. BOD, in.
Note: BID = bottom inside diameter; BOD = bottom outside diameter; H = height; L = length; n.a. = not applicable; TID = top inside diameter; TOD =
top outside diameter; W = width. 1 in. = 25.4 mm.
The properties provided in the tables are average measure- and outside crown-to-crown diameter of strand dp—are all
ments taken from a minimum of three samples. Variations within ASTM A416 tolerance limits. The cross-sectional
exist in the average dimensions of both the anchors and area ap of each strand was determined from measured
the wedges depending on the manufacturer and the model. dimensions and weights of the strand samples10,16 (that is,
Binding rings hold the wedge pieces together, particularly measured strand areas rather than nominal areas were used
during seating. The wedges without rings (always two- in the research).
piece wedges in this sample set) do not have the notches
shown in Fig. 2. Test setup
Table 3 shows the properties of the tested strands, named As described in Walsh and Kurama,10 each strand speci-
according to the nominal diameter and differentiated men was positioned through the crossheads of a 600 kip
for each different source (spool) from which the mate- (2700 kN) hydraulic universal testing machine with a
rial was obtained. Strands 0.5A and 0.6A, from the same post-tensioning anchor placed on the outer surface of each
manufacturer, were used exclusively in the testing of the crosshead. All of the monostrand barrel anchors and the
strand-anchorage assemblies, while all of the 10 different nine-hole multistrand anchor (multi-0.6A) bore directly
strands listed, representing five different manufacturers, on 1.5-in.-thick (38 mm) steel plates, each with a central
were used in the free-length fracture tests described at the hole oversized 1/16 in. (1.6 mm) with respect to the nominal
end of this paper. The strand properties in the table are strand diameter. The bearing plates were screwed to the
average measurements from a minimum of three samples. machine crossheads and the central holes were aligned
The measured wire and strand diameters—diameter of vertically using a laser alignment tool such that each strand
middle strand wire dmw, diameter of outer strand wire dow, sample was placed in the same position with minimal
Strand spool dmw, in. dow, in. dp, in. ap, in.2 fpm,free-length, ksi εpf,free-length
Note: a p = strand cross-sectional area; dmw = middle-wire diameter; dow = outer-wire diameter; dp = strand diameter; fpm,free-length = maximum free-length
strength; εpf,free-length = free-length fracture strain. 1 in. = 25.4 mm; 1 ksi = 6.895 MPa.
accidental eccentricity between the ends of the strand. the extensometer gauge length, little strand deformation
Further validation on the concentric axial loading of the occurred at the fracture location due to the lack of neck-
test specimens (for example, by testing strand samples with ing (that is, reduction in wire diameter) in the fractured
extended free-lengths) was also conducted.10,16 wires (Fig. 1). Thus, the extensometer strains from the
post-tensioning strand-anchorage tests provide an accu-
At the beginning of each test, the crossheads of the testing rate measurement of the strand strains up to fracture. All
machine were positioned such that a strand free length of test equipment components, including the load cell and
about 72 in. (1800 mm) was achieved between the top and position rate of the testing machine, as well as the gauge
bottom post-tensioning anchors after seating the anchor lengths and strain measurements of the extensometers,
wedges. The anchor wedges were placed and tapped in light- were calibrated at least yearly on-site. Comparisons and
ly by hand using the back of a spare anchor barrel before a validation of the measurements from the two extensom-
preload of approximately 750 lb (3300 N) was applied to eters were also made.10,16
the strand. Then the wedges were checked for proper initial
seating, the extensometers were initialized, and the strand For the monostrand cast anchors with the wedge hole
was loaded to fracture by moving the top machine crosshead protruding from the bottom surface of the anchor (referred
while the bottom crosshead remained stationary. to as the anchor boss [Fig. 2]), each anchor was screwed to
a fitted steel support plate10 to use the full bearing area of
During each test, the strand stress and strain data, as well the anchor, simulating an anchor embedded inside con-
as the time elapsed, applied load, crosshead displacement, crete. The necessity for these support plates was proved by
and strand elongation (based on the strain data) were preliminary tests that compared strand-anchorage perfor-
recorded continuously at a frequency of 100 Hz. The strand mance with and without support plates. The support plates
stress was calculated by dividing the measured load by the were machined such that the bottom of the anchor boss
measured strand area ap (Table 3). Each test was concluded was flush with the underside of the plate.
following the point at which the first strand wire fracture
occurred, resulting in a sudden drop in stress. After the test, Multistrand anchorages
the number of simultaneous wire fractures and the loca-
tions of the fractures were recorded. The multistrand anchorages tested in this research had
multiple wedge holes on the perimeter and one central
Two extensometers (with 2.0 in. [50 mm] and 36 in. hole (Fig. 2 shows a nine-hole anchor). As specified by the
[910 mm] gauge lengths) were placed within the free Post-Tensioning Institute’s (PTI’s) Acceptance Standards
length of the strand (as required by ICC-ES) between the for Post-Tensioning Systems,18 these anchors were tested
anchors. While both extensometers were used to measure one strand at a time because of the uncertainties associ-
the strand strains up to fracture, the strain measurements ated with loading multiple strands simultaneously (for
presented in this paper are from the 36 in. extensometer example, difficulties in loading the strands evenly). Only
unless stated otherwise. While the strand wire fractures oc- the outside wedge holes were used because of the different
curred inside the post-tensioning anchors, and thus outside anchor geometry around the middle hole. After each test,
200,000
Stress, psi
0.70fpu
150,000
100,000
50,000
0
0 1 2 3 4
Time, min
Figure 3. The effects of various loading parameters on the strand fracture strains were investigated through different testing procedures. Note: fpu = nominal ultimate
strand strength. 1 in. = 25.4 mm; 1 psi = 6.895 kPa.
the anchor wedges and a short gripped length of the test procedure was applied to three to twelve samples of each
strand were left inside the wedge hole. For the nine-hole of the anchor-wedge configurations from Tables 1 and 2.
anchors (multi-0.6A), which were tested resting directly on
the 1.5-in.-thick (38 mm) bearing plates connected to the • ICC: This control procedure10,16 conformed to the
top and bottom machine crossheads, the strand was ground static testing requirements of ICC-ES. After a preload
flush with the anchor bottom to allow subsequent testing of of approximately 750 lb (3300 N) was applied, the
the remaining holes. strand was loaded to fracture by monotonically mov-
ing the top machine crosshead upward to result in a
The seven-hole anchors (multi-0.6B) were each composed constant nominal strain rate of about 0.02/min.
of two parts, with a wedge plate placed on top of a 4-in.-
deep (100 mm) hollow cylindrical anchor pedestal. The • SLO: In the SLO procedure, a slower monotonic
wedge plate–pedestal assembly was placed directly on the strand strain loading rate of about 0.01/min was used.
bearing plate at each end of the strand, and the relative This slow loading rate was studied because the control
position of the machine heads was decreased to maintain ICC rate of 0.02/min was at the high end of the ac-
a 72 in. (1830 mm) strand free-length between the wedge ceptable ICC-ES range, allowing for the investigation
plates. Also, because the wedge plates were tested resting of any differences in the strand behavior due to the
on the hollow pedestals, after each test, the short gripped different strain rate.
lengths of the fractured strand only needed to be cut short
(that is, not ground flush with the bottom of the wedge • ECC: In the ECC procedure, a strand end eccentricity
plate) to fit inside the pedestals. of 3.3% was applied at the control monotonic strain
rate of 0.02/min. The importance of testing specimens
During the investigation of the multistrand anchors, both eccentrically stems from the differential transverse
the wedge hole locations used in the different tests and the displacements the strand ends can undergo in an
test results were closely monitored to determine whether unbonded post-tensioned structure under lateral load-
differences in strand fracture strains might have occurred ing. Furthermore, load eccentricity may occur due to
due to the presence of previously loaded wedge holes the presence of strand flares inside the trumpet at the
(containing short lengths of fractured strand) next to the end of a multistrand tendon. A strand end eccentricity
hole being tested. No significant change in strand-anchor as large as 3.3% can develop under such conditions,
performance was found because of the previously loaded which was simulated by offsetting the hole in the top
holes in the wedge plates. bearing plate by 2.4 in. (61 mm) (Fig. 3).
Note: CV = coefficient of variation; ICC = standard monotonic test procedure based on International Code Council Evaluation Service11 requirements;
n.a. = not applicable.
Note: CV = coefficient of variation; ICC = standard monotonic test procedure based on International Code Council Evaluation Service11 requirements;
n.a. = not applicable; SLO = monotonic test procedure with slower loading rate than ICC procedure.
Note: CV = coefficient of variation; ECC = monotonic test procedure with eccentric loading; n.a. = not applicable.
Note: CV = coefficient of variation; CYC = cyclic test procedure with same loading rate as ICC procedure; ICC = standard monotonic test procedure
based on International Code Council Evaluation Service11 requirements; n.a. = not applicable.
Note: CV = coefficient of variation; CYC-FST = cyclic test procedure with faster loading rate than CYC procedure; n.a. = not applicable.
Note: CV = coefficient of variation; n.a. = not applicable; PRE = monotonic test procedure including a prestress.
Note: CV = coefficient of variation; EQK = cyclic test procedure with a prestress, eccentricity, and fast loading rate; n.a. = not applicable.
280,000 280,000
Fracture stress, psi
270,000 270,000
260,000 260,000
250,000 250,000
240,000 240,000
230,000 230,000
0.00 0.01 0.02 0.03 0.04 0.05 0.00 0.01 0.02 0.03 0.04 0.05
Fracture strain Fracture strain
ICC (0.5 in.) SLO (0.5 in.) ICC (0.5 in.) ECC (0.5 in.)
ICC (0.6 in.) SLO (0.6 in.) ICC (0.6 in.) ECC (0.6 in.)
280,000 280,000
Fracture stress, psi
Fracture stress, psi
270,000 270,000
260,000 260,000
250,000 250,000
240,000 240,000
230,000 230,000
0.00 0.01 0.02 0.03 0.04 0.05 0.00 0.01 0.02 0.03 0.04 0.05
Fracture strain Fracture strain
ICC (0.5 in.) CYC (0.5 in.) ICC (0.5 in.) CYC-FST (0.5 in.)
ICC (0.6 in.) CYC (0.6 in.) ICC (0.6 in.) CYC-FST (0.6 in.)
Figure 4. Effects of loading parameters on strand fracture stress versus fracture strain. Note: CYC = cyclic test procedure with same loading rate as ICC procedure;
CYC-FST = cyclic test procedure with faster loading rate than CYC procedure; ECC = monotonic test procedure with eccentric loading; ICC = standard monotonic test
procedure based on International Code Council Evaluation Service11 requirements; SLO = monotonic test procedure with slower loading rate than ICC procedure. 1 in.
= 25.4 mm; 1 psi = 6.895 kPa.
260,000 260,000
250,000 250,000
240,000 240,000
230,000 230,000
0.00 0.01 0.02 0.03 0.04 0.05 0.00 0.01 0.02 0.03 0.04 0.05
Fracture strain Fracture strain
CYC (0.5 in.) CYC-FST (0.5 in.) ICC (0.5 in.) PRE (0.5 in.)
CYC (0.6 in.) CYC-FST (0.6 in.) ICC (0.6 in.) PRE (0.6 in.)
280,000 280,000
Fracture stress, psi
260,000 260,000
250,000 250,000
240,000 240,000
230,000 230,000
0.00 0.01 0.02 0.03 0.04 0.05 0.00 0.01 0.02 0.03 0.04 0.05
Fracture strain Fracture strain
ICC (0.5 in.) EQK (0.5 in.) CYC-FST (0.5 in.) EQK (0.5 in.)
ICC (0.6 in.) EQK (0.6 in.) CYC-FST (0.6 in.) EQK (0.6 in.)
Figure 5. Further comparisons among the test results. Note: CYC = cyclic test procedure with same loading rate as ICC procedure; CYC-FST = cyclic test pro-
cedure with faster loading rate than CYC procedure; EQK = cyclic test procedure with a prestress, eccentricity, and fast loading rate; ICC = standard monotonic
test procedure based on International Code Council Evaluation Service 11 requirements; PRE = monotonic test procedure including a prestress. 1 in. = 25.4 mm;
1 psi = 6.895 kPa.
are small on average, indicating that the dominant param- dard deviation and coefficient of variation due to the varied
eter is the postyield cyclic loading. The loading parameter loading parameters are generally greater than the decreases
from the PRE procedure has a slightly negative effect, as in these measures, and the overall data ranges from the
controlled by the greatly reduced fracture strains for only varied loading parameter procedures in Fig. 4 and 5 tend to
one of the six anchor-wedge configurations, whereas the be larger than those from the base procedure.
SLO procedure has a slightly positive effect and the ECC
procedure has almost no effect on average. However, these Stressing-end effect
are average trends from the entire ensemble and the effects
of the loading parameters greatly depend on the anchor- Figure 6 combines the data from the EQK and PRE
wedge configuration. procedures, exhibiting the strand fracture end from each
test. Twenty-five tests from this combined data pool had
The smallest fracture strains from the varied loading wire fractures occur inside the bottom anchor, 24 tests had
parameter test data are always smaller than the smallest fractures occur inside the top anchor (the prestressing end),
fracture strains from the base test data (that is, the left-most and one test had wire fractures occur simultaneously inside
yellow markers are always to the left of the left-most red both anchors. Based on these data, it is concluded that the
markers) (Fig. 4 and 5). This means that the varied loading wire fractures occurred fairly randomly inside the top or
parameter always results in the single worst strand-anchor bottom anchor, and the use of a prestressing jack in the
performance in each comparison chart. The effects of the EQK and PRE procedures did not have an observable ef-
loading parameters on the standard deviation and coef- fect on the fracture end. Tests with fractures inside the top
ficient of variation of the fracture strains vary significantly. anchor at the prestressing end, however, did appear to have
Nevertheless, Tables 4–10 show that the increases in stan- a slightly larger range of fracture strains.
Stress, psi
Fracture strain
0.03
150,000
0.02 100,000
0.40fpu
0.01 50,000
0.00 0
-2 -1 (25 points)0 Top anchor1(24 points) 2 0 2 4 6 8 10 12
Bottom anchor
Time, min.
PRE EQK
Figure 7. Fifty-cycle loading pattern with strain rate of 0.056/min com-
Figure 6. The test data from the EQK and PRE procedures are combined, exhib- pleted without strand fracture. Note: fpu = nominal ultimate strand strength.
iting the fracture end from each test. Note: EQK = cyclic test procedure with a 1 psi = 6.895 kPa.
prestress, eccentricity, and fast loading rate; PRE = monotonic test procedure
including a prestress.
260,000 260,000
250,000 250,000
240,000 240,000
230,000 230,000
0.00 0.01 0.02 0.03 0.04 0.05 0.00 0.01 0.02 0.03 0.04 0.05
Fracture strain Fracture strain
ICC CYC ICC CYC
Figure 8. Among the procedures investigated in this paper, postyield cyclic loading appears to be the common loading parameter that tends to cause significant
reductions in the strand fracture strains. Note: CYC = cyclic test procedure with same loading rate as ICC procedure; ICC = standard monotonic test procedure based
on International Code Council Evaluation Service11 requirements. 1 psi = 6.895 kPa.
Strand
Differential seating
Anchor
Wedge Wedge
0.05 0.05
0.04 0.04
Fracture strain
Fracture strain
0.03 0.03
0.02 0.02
0.01 0.01
0.00 0.00
0.00 0.05 0.10 0.15 0.20 0.00 0.05 0.10 0.15 0.20
Wedge differential seating, in. Wedge differential seating, in.
Cast-0.5B anchors with 2P-0.5C wedges Cast-0.5C anchors with 2P-0.5C wedges
Figure 9. Poor wedge seating may be associated with low fracture strains, but low fracture strains are not entirely dependent on poor wedge seating.
Note: 1 in. = 25.4 mm.
Fracture strain
age in Table 11 provides the probability of observing the 0.03
measured difference in the test averages between the two
procedures if the actual averages from the two procedures 0.02
(which could only be determined by running an unfeasi-
bly large number of tests) were, in fact, equal. A p-value 0.01
of 0.27% for the first anchor-wedge configuration, for
0.00
example, indicates that it is nearly impossible for the actual 0.00 0.05 0.10 0.15 0.20
averages from the two procedures to be equal. Conclu-
Wedge differential seating, in.
sively, the CYC procedure causes a significant reduction in
the average fracture strain of this particular anchor-wedge
Figure 10. Wedge differential seating measured for all of the tests conducted in
configuration. The p-value of the second anchor-wedge
this research project. Note: 1 in. = 25.4 mm.
configuration is not nearly as conclusively small, but it still
indicates a favorable degree of likelihood that the CYC
procedure causes a reduction in the average fracture strain the consistency and ease with which the 0.10 in. limit is
of that particular configuration. applied rendered it useful in invalidating certain test data in
this research. Of course, the cutoff differential seating could
Figure 8 plots the fracture stress and strain for each of the be adjusted anywhere from about 0.08 in. (2.0 mm) to about
anchor-wedge configurations in Table 11, comparing the 0.12 in. (3.0 mm) in accordance with these data.
data from the two loading procedures. Not only are the
averages of the ICC data higher than the averages of the Results from the strand free-
CYC data for both configurations but consistent with the length fracture experiments
trend observed in earlier plots, an ICC data point represents
the largest fracture strain in each configuration and a CYC As a corollary to the post-tensioning anchorage experiments,
data point represents the smallest fracture strain in each samples from all of the 10 strand spools (Table 3) were
configuration as well. tested to result in free-length seven-wire fracture of each
specimen in accordance with ASTM A370.20 These tests
Wedge differential seating were conducted using the same setup and equipment as the
strand-anchorage experiments described previously, but with
An analysis of the relationship between wedge differential special “sand-grip” anchors10,16 to achieve free-length failure.
seating and fracture strain was conducted. The data for Table 3 lists the average maximum free-length strength
this investigation was accumulated from the normal testing fpm,free-length and fracture strain εpf,free-length for each strand.
pool, and no test was run specifically to control the amount
of wedge differential seating because the investigators Figure 11 shows the average stress-strain behaviors from
wanted to observe the tendency of wedge seating under all tests of the five different spools in each strand diameter,
normal conditions. All of the tests presented thus far in the demonstrating considerable differences between the 0.5 in.
paper had small differential seating between the anchor (13 mm) and 0.6 in. (15 mm) strand sizes. As described
wedges. Any test with more than 0.10 in. (2.5 mm) of dif- in the following section, this information becomes impor-
ferential wedge seating at wire fracture was discarded from tant when determining a relationship that can be used to
the main test data pool but was recorded for the purposes correlate strand strains with stresses for design purposes.
of the particular investigation described in this section. The free-length fracture strain is reached at an apparent
stress slightly smaller than the maximum strength because
Figure 9 shows a photograph of well-seated wedges and of the necking (that is, reduction of wire diameter prior
also illustrates poor wedge differential seating. The test to fracture) that occurs in the strand wires in the case of
data in this figure, taken from two different anchor-wedge free-length fracture (Fig. 11). This behavior is different
configurations, also exhibit that poor wedge seating may from the brittle wire fractures observed inside the post-
be associated with low fracture strains but that low fracture tensioning anchors as described in the previous section,
strains are not entirely dependent on poor wedge seating. which occurred with no visible necking (Fig. 1), resulting
in the maximum strength and fracture strain being reached
Figure 10 shows the wedge differential seating measured for at virtually the same point on the stress-strain curve.
all of the tests conducted in this research project (including the
tests described in Walsh and Kurama)10 compared with their Strand linear-elastic modulus
fracture strains. This data pool includes wedges of several
different heights. Therefore, 0.10 in. (2.5 mm) represents a The PCI Design Handbook: Precast and Prestressed
different percentage of each wedge’s height. Nonetheless, Concrete21 assumes a linear-elastic modulus Ep equal to
250,000
Stress, psi
200,000
150,000
100,000
50,000
0
0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08
Strain
0.5 in. average 0.6 in. average
Figure 11. Strand free-length fracture experiments were conducted to determine the actual strength and strain capacity of the strands.
Note: 1 in. = 25.4 mm; 1 psi = 6.895 kPa.
28,500 ksi (196,500 MPa) for prestressing strand. Table 12 ference between the two corresponding strains within the
lists the average elastic modulus measured for each of the initial linear-elastic range of the measured strand stress-
strand spools (from three samples each) as well as the aver- strain relationship. For most tests, the two stresses chosen
age elastic modulus for each diameter of strand tested. were 50 ksi (350 MPa) and 150 ksi (1030 MPa), but these
values were altered if irregularities were observed in the
The linear-elastic modulus from each test was calculated as stress-strain curve near these points.
the ratio of the difference between two stresses and the dif-
0.5A 28,028
0.5B 29,474
0.5D 28,276
0.5E 28,926
0.6A 27,872
0.6B 27,827
0.6D 27,308
0.6E 28,328
Note: 1 ksi = 6.895 MPa.
Stress, psi
260,000
appears that 0.5-in.-diameter strand tends to have a slightly
250,000
higher elastic modulus as well. Nonetheless, the average of
the entire sample set is 28,435 ksi (196,050 MPa), which 240,000
is close to the value of 28,500 ksi (196,500 MPa) from the 230,000
PCI Design Handbook.21
220,000
0.00 0.01 0.02 0.03
Strand stress–strain relationship
Strain
The information derived from the free-length fracture PCI Combined test average Proposed combined
experiments is important not only for determining anchor-
independent strand capacities but also for determining PCI Design Handbook, combined test average,
and proposed combined relationship
idealized, approximate strand stress–strain relationships
that can be used to correlate strand strains with stresses for
design purposes. The PCI Design Handbook includes the
280,000
following expression to relate stress fps and strain εps for
strand with nominal 270 ksi (1860 MPa) strength:ksi 270,000
230,000
This relationship gives a strand limit-of-proportionality
stress of fpl equal to 245.1 ksi (1690 MPa) at a strand 220,000
limit-of-proportionality strain of εpl equal to 0.0086. The 0.00 0.01 0.02 0.03
PCI Design Handbook proportionality limit stress is high Strain
compared with the results of this investigation and does not PCI 0.5 in. test average Proposed 0.5 in.
distinguish between 0.5 in. (13 mm) and 0.6 in. (15 mm)
strand diameters. Figure 12 plots the PCI Design Hand- PCI Design Handbook, 0.5-in.diameter strand test
average, and proposed 0.5-in.-diameter relationship
book stress-strain relationship against the combined test
average (that is, average of all 0.5-in.-diameter and 0.6-in.-
diameter strand test results) and a proposed relationship
derived from this combined average as follows: 280,000
270,000
For εps ≤ 0.0082, fps = 28,500εps ksi
Stress, psi
260,000
240,000
This relationship uses a strand limit-of-proportionality
stress of 233.7 ksi (1611 MPa) at a strain of 0.0082. The 230,000
proposed relationship provides a better approximation 220,000
than the PCI Design Handbook relationship does to the 0.00 0.01 0.02 0.03
measured average strand stress–strain curve from the Strain
proportionality limit up to a strain of about 0.02, which is
PCI 0.6 in. test average Proposed 0.6 in.
the most important strain range for most practical design
purposes. Although the differences in the proposed and PCI Design Handbook, 0.6-in.diameter strand test
PCI Design Handbook strand stress–strain relationships in average, and proposed 0.6-in.-diameter relationship
Fig. 12 may look insignificant, small variations in stress
result in rather large differences in strain in this nonlinear
range of behavior. The reason for the deviation of the PCI
Design Handbook strand stress–strain relationship from the
measured data in this paper is unknown. Figure 12. The free-length fracture experiments were used to determine ideal-
ized, approximate strand stress–strain relationships for design purposes.
Note: 1 in. = 25.4 mm; 1 psi = 6.895 kPa.
PCI Design Handbook Proposed combined Proposed 0.5 in. diameter Proposed 0.6 in. diameter
Note: For εps ≤ εpl, fps = Epεps ksi; for εps > εpl, fps = 270α - [a/(εps - b)] ksi; a = (270α - Epεpl)(εpl - b); b = equation variable; Ep = strand linear-elastic
modulus; fpl = strand limit-of-proportionality stress; fps = strand stress; α = ultimate strength factor; εpl = strand limit-of-proportionality strain; εps =
strand strain. 1 in. = 25.4 mm; 1 ksi = 6.895 MPa.
Similar to the PCI Design Handbook relationship, the tion format. These adjustments include lowering the elastic
proposed strand stress–strain relationship does not account modulus to emulate the average elastic modulus from the
for the strain hardening observed in the test data. Using 0.6-in.-diameter strand tests (Table 12) and adjusting the
the same form as the PCI Design Handbook relation- nominal strength by a factor of 0.97 as follows:
ship, the proposed model simply lowers the propor-
tionality limit and then asymptotically approaches the For εps ≤ 0.0081, fps = 28,000εps ksi
nominal 270 ksi (1862 MPa) strength as the PCI Design
Handbook curve does. The proposed proportionality For εps > 0.0081, fps = 0.97 × 270 - [0.060/(εps - 0.0064)] ksi
limit strain of 0.0082 is high enough for the jacking
stress limit of 0.80fpu permitted in American Concrete This relationship provides a proportionality limit stress
Institute’s (ACI’s) Building Code Requirements for of 226.8 ksi (1575 MPa) at a strain of 0.0081. Figure 12
Structural Concrete (ACI 318-08) and Commentary shows the relevant stress-strain curves. Of course, the
(ACI 318R-08)22 for 270 ksi (1860 MPa) strand. The adjustment factor 0.97 on the nominal strength limits the
proportionality limit stress of 233.7 ksi (1611 MPa) is ultimate strength of the strand, but the conservative nature
less than the minimum yield stress of 0.9fpu or 243 ksi of this adjustment and its necessity to accurately model the
(1675 MPa) specified for low-relaxation strand in ASTM test data between the proportionality limit and a strain of
A416; however, the ASTM A416 yield stress require- 0.02 warrant its use.
ment is defined at a strain of 0.01, which is satisfied by
the proposed stress-strain relationship. Table 13 gives a summary of the proposed stress-strain rela-
tionships. The difference in the average maximum free-length
Given the considerable difference in the yield stresses for strengths of the 0.5-in.-diameter (13 mm) strands (286.6 ksi
the 0.5-in.-diameter (13 mm) and 0.6-in.-diameter (15 mm) [1976 MPa]) and the 0.6-in.-diameter (15 mm) strands
strands, separate stress-strain relationships can also be (279.7 ksi [1928 MPa]) tested in this investigation is significant.
proposed for each diameter. The PCI Design Handbook Assuming any imprecision in load or area measurements is too
stress-strain relationship needs to be altered only slightly small to account for the notable strength differences between
to model the average test data from the 0.5-in.-diameter the two strand diameters, these differences may have been
strand samples as follows: caused by material and/or geometrical variations. To determine
steel material strengths, three middle-wire samples were taken
For εps ≤ 0.0083, fps = 28,500εps ksi from each of the two most tested strand spools (0.5A and
0.6A), which were from the same manufacturer. The middle
For εps > 0.0083, fps = 270 - [0.043/(εps - 0.007)] ksi wires are straight and are slightly larger in diameter than the
outer wires, permitting a more accurate measurement of the
This relationship results in a proportionality limit stress wire cross-sectional area and the application of concentric ten-
of 236.6 ksi (1631 MPa) at a strain of 0.0083. Figure 12 sile loading. The middle-wire samples were tested in a smaller
shows the relevant stress-strain curves. The PCI Design 60 kip (270 kN) hydraulic universal testing machine. Standard
Handbook curve is barely visible beneath the proposed wedge grips available with the testing machine were used to
0.5-in.-diameter (13 mm) strand curve. hold the wires at both ends, and all tests resulted in free-length
fracture.
Modeling the 0.6-in.-diameter (15 mm) strand average
test data requires more significant adjustments to the PCI The wire specimens were tested at a consistent strain rate,
Design Handbook relationship to maintain the same equa- resulting in an average maximum strength of 289.6 ksi
(1997 MPa) for the wires from the 0.5-in.-diameter the expected strand strains exceed this limit, anchors
(13 mm) strand samples and 287.8 ksi (1984 MPa) for the that have been proved to consistently achieve higher
wires from the 0.6-in.-diameter (15 mm) strand samples. strand wire fracture strains should be specified.
The difference between these middle-wire strengths, as-
sumed to depict material tensile strengths, does not account • Poor wedge seating may be associated with low strand
for the considerably greater difference between the average wire fracture strains, but low fracture strains are not
strengths of the entire 0.5-in.-diameter and 0.6-in.-diameter entirely dependent on poor wedge seating.
strand sample pools or between the average strengths of
strands 0.5A and 0.6A (Table 14). Thus the reason for the • The measured yield and maximum strengths of the
observed differences in the 0.5-in.- and 0.6-in.-diameter 0.5-in.-diameter (13 mm) strand samples tested in this
strand strengths is not known. research were higher than those of the 0.6-in.-diameter
(15 mm) strand samples.
Conclusion
• The prestressing strand stress–strain relationship
This research investigated the strand fracture strains of provided by the PCI Design Handbook needs to
unbonded post-tensioning strand-anchorage systems under be updated. Given the notable strength differences
different loading conditions. In addition, stress-strain between the 0.5-in.-diameter (13 mm) and 0.6-in.-
relationships were proposed for seven-wire, low-relaxation, diameter (15 mm) strands, separate stress-strain rela-
uncoated 270 ksi (1860 MPa) prestressing strand. tionships are proposed for each diameter, in addition
to a relationship for the combined 0.5-in.-diameter and
The following conclusions and recommendations are made 0.6-in.-diameter strand data pool.
based on the investigation:
Additional research should investigate the following topics
• Postyield cyclic loading is the parameter that most related to this paper:
consistently reduces the average wire fracture strains
of strand-anchorage systems. • The maximum strand strain demands expected in un-
bonded post-tensioned structures under extreme seis-
• Postyield cyclic loading should be used in the valida- mic loading should be researched and compared with
tion testing of unbonded post-tensioning strand-an- the performance data measured in this research. The
chorage systems for extreme seismic applications. investigation of the maximum strand strain demands
should consider the geometry and type of the structure
• The ICC-ES 50-cycle fatigue loading test uses an (for example, walls, frames, coupling beams), the
inadequate upper stress limit to observe the postyield intensity of the earthquake, and the initial strain level.
cyclic behavior that may occur in strands under ex-
treme seismic loading conditions. • Testing procedures that subject strand-anchorage
systems to cyclic end eccentricity should be devised.
• The large standard deviation and coefficient of variation By comparison, the strand end eccentricity used in
of the fracture strain data, with strains almost as large this research maintained the position of the transverse
as 0.04 and as small as 0.01, or slightly less for a small offset without mirroring it about the concentric strand
number of samples, suggests that a larger sample pool axis during the applied loading. A cyclic eccentricity
size should be used in the validation testing of post- condition (that is, full reversal of the transverse offset)
tensioning strand-anchorage systems than the pool size would more realistically model the effects of seismic
of three samples that is currently required by ICC-ES. loading.
• Unbonded post-tensioned structures under extreme • Testing equipment that is capable of subjecting the
seismic loading conditions should be designed such strand specimens to more realistic (that is, faster)
that the maximum strand strains are less than 0.01. If seismic strain rates should be used.
This research was funded by PCI under a Daniel P. Jenny 8. Post-Tensioning Institute (PTI). 2000. Specification for
Fellowship. The support of the members of the PCI Unbonded Single Strand Tendons. Phoenix, AZ: PTI.
Research and Development Committee and the project
advisory members, Tom D’Arcy (chair), Ned Cleland, and 9. Innovation Task Group 5. 2007. Acceptance Criteria
Richard Sause is gratefully acknowledged. In addition, the for Special Unbonded Post-Tensioned Precast Struc-
authors recognize the support provided by: Neil Hawkins tural Walls Based on Validation Testing (ACI 5.1-07)
of the University of Illinois at Urbana-Champaign; Susie and Commentary (ACI 5.1R-07). ITG-5.1. Farmington
Nakaki of the Nakaki Bashaw Group Inc.; John Pearson Hills, MI: ACI.
and Roger Pelletier of Wiss, Janney, Elstner Associ-
ates Inc.; John Reihl and Ken Everett of StresCore Inc.; 10. Walsh, K., and Y. Kurama. 2009. Behavior of Unbond-
Alan Stone of Aston Metallurgical Services; Brent Bach, ed Post-tensioning Monostrand Anchorage Systems
Michael Gekhtman, Kapil Khandelwal, Tracy Kijewski- under Monotonic Tensile Loading. PCI Journal, V. 55,
Correa, Albert Miller, Patrick Murren, and Brad Weldon of No. 1 (Winter): pp. 97–118.
the University of Notre Dame; and a number of material
suppliers and donors. The opinions, findings, and conclu- 11. International Code Council Evaluation Service Inc.
sions expressed in the paper are those of the authors and do (ICC-ES). 2007. Acceptance Criteria for Post-tension-
not necessarily reflect the views of PCI or the individuals ing Anchorages and Couplers of Prestressed Concrete.
and organizations acknowledged. AC 303. Whittier, CA: ICC-ES.
19. Ross, S. M. 2004. Probability and Statistics for Engi- TID = anchor top inside diameter
neers and Scientists. 3rd ed. Burlington, MA: Elsevier
Academic Press. TOD = anchor top outside diameter (equal to BOD in
case of barrel anchors)
20. ASTM International. 2008. Standard Test Methods and
Definitions for Mechanical Testing of Steel Products. TW = wedge top outside width (diameter in case of
Annex A7 Method of Testing Multi-Wire Strand for two-piece wedges)
Prestressed Concrete. ASTM A370. West Conshohock-
en, PA: ASTM International. W = cast anchor width
21. PCI Industry Handbook Committee. 2004. PCI Design α = ultimate strength factor in strand stress–strain
Handbook: Precast and Prestressed Concrete. MNL- equations
120. 6th ed. Chicago, IL: PCI.
εpf,free-length = strand fracture strain from free-length fracture
22. ACI Committee 318. 2008. Building Code Require- tests
ments for Structural Concrete (ACI 318-08) and Com-
mentary (ACI 318R-08). Farmington Hills, MI: ACI. εpl = strand limit-of-proportionality strain
Abstract Keywords
This paper presents an experimental investigation Anchor, cyclic, eccentricity, post-tensioning, seismic,
on the strand wire fracture stresses and strains in strand, stress-strain relationship, wedge, wire fracture.
unbonded post-tensioning strand-anchorage systems
subjected to a variety of loading conditions that can Review policy
affect the performance of the strand inside the anchor.
While unbonded post-tensioned gravity load sys- This paper was reviewed in accordance with the
tems (for example, floor and roof slabs) are common Precast/Prestressed Concrete Institute’s peer-review
throughout the United States, the use of this construc- process.
tion technique for seismic-resistant structures can put
the strands under significantly greater strain demands. Reader comments
Strand wire fractures can occur inside the anchorages,
limiting the seismic performance of the structure. Please address any reader comments to journal@pci
.org or Precast/Prestressed Concrete Institute, c/o PCI
The research focuses on the following loading param- Journal, 200 W. Adams St., Suite 2100, Chicago, IL
eters: loading rate, eccentricity between the strand 60606. J
ends, postyield cyclic loading, and initial strand stress
(that is, prestress). Eight representative anchor-wedge
configurations with 0.5 in. (13 mm) or 0.6 in. (15 mm)
P
recast concrete structures are used worldwide because
of their shorter construction time, cost effectiveness,
and better plant control of structural elements and
materials compared with cast-in-place concrete structures.
■ Experimental tests on the cyclic behavior of column-to-founda- This structural layout reduces construction time and is cost
tion subassemblies were conducted to compare the response effective. However, this effectiveness could be compro-
of grouted sleeve connections with those of cast-in-place and mised for construction in seismic regions, especially if
pocket-foundation connections. the design follows the capacity design rule. The design of
column footings is based on the assumption that a plastic
■ Confinement by the grouted sleeves inhibited buckling of the hinge may develop at the column base in an earthquake
longitudinal reinforcement and increased the compressive event, while capacity design is used to prevent the forma-
strength of the grout. tion of an inelastic mechanism in the footing. The latter
goal is achieved by considering the bending moment asso-
■ The damage was localized to the column base, allowing ciated with the column base flexural capacity acting on the
easier postseismic column repair compared with traditional foundation amplified by an overstrength factor. Following
connections. this design approach, the foundation base becomes large,
This paper investigates the behavior and performance of The horizontal loads and displacements were recorded by
grouted corrugated steel sleeve connections under cyclic a load cell and a string pot transducer, respectively, while
loading. While the behavior of precast concrete socket base three pairs of linear displacement transducers evaluated
connections is well documented,4–6 no published experi- the average column curvature recorded at three consecu-
mental results concerning the cyclic response of grouted tive levels of the column base (Fig. 2). The lower pair of
sleeve column-to-foundation connections have been found. displacement transducers was connected directly to the
foundation. Therefore, the average curvature evaluated
Although temporary bracing is needed for erection comprises the strain penetration of the reinforcing bars into
stability, grouted steel sleeves are an inexpensive type of the foundation block.
connection, especially if the retrofit after a seismic event
is considered. The confinement provided by grouted steel Figure 3 shows the geometry of the tested specimens.
sleeves inhibits buckling of the longitudinal reinforcement Specimen CP represented a typical cast-in-place con-
and increases the strength of the grout. As a result, the crete column-to-foundation connection, and specimen
damage associated with this kind of connection is localized PF represented a typical grouted pocket foundation with
to the column base, allowing easier postseismic column no extra roughening of the surfaces at the column and
repair compared with traditional connections. foundation interface. Specimens GS4 and GS4B both had
four grouted sleeves, but each specimen had a different
Another advantage of grouted steel sleeve connections anchorage length of the reinforcing bars in the foundation.
compared with modern mechanical splice devices is the The specimen GS4 had 90° hooks at the reinforcing bar
ability to include an unbonded portion in the reinforc- ends, whereas GS4B had straight reinforcing bars, which
ing bar inside the sleeves. This distributes the reinforcing represented a connection sometimes adopted to increase
bar deformation over the unbonded length, increasing the construction speed. Specimen GS4U was equal to speci-
ultimate rotation capacity of the subassembly and reducing men GS4B with an additional 300 mm (12 in.) unbonded
the reinforcement strain demand for a given column drift length on the spliced reinforcing bars. Adhesive tape pro-
(defined as the interstory displacement divided by the inter- vided the unbonding for these bars. GS4UR was specimen
story height). This advantage is not typical of mechanical GS4U after retrofitting. The retrofitting sequence consisted
splice devices. The unbonded length of the reinforcement of clearing the grout crushed during the previous tests at
can be taken into account in the design process by consid- the column base, placing collar formwork, and filling it
ering an appropriate plastic hinge length. with high-strength shrinkage-compensating grout.
Electromechanical
screw jack
(1000 kN capacity)
Column-to-
foundation
Self-equilibrated
subassembly
reaction frame
used inside the sleeves, retrofitted connection for GS4UR, for the 2nd order load-displacement effects P-Δ, indi-
and pocket foundation PF was a shrinkage-compensating cated by the dashed lines. Any gain or loss of strength is
cement-based grout with a maximum aggregate size of obtained when the experimental curve in absolute terms
2.5 mm (0.1 in.) and a specified 28-day strength greater lies above or below the P-Δ lines, respectively. Figure 5
than 75 MPa (11 ksi). The ducts used for the grouted sleeve shows the horizontal load versus drift comparison between
connections were typical post-tensioning ducts made with the unbonded grouted sleeve connection before and after
corrugated galvanized strip steel with an 80 mm (3 in.) retrofitting.
internal diameter, an 84 mm (3.3 in.) external diameter, and
a 0.6 mm (24-gauge) thickness. Figure 6 shows pictures of the damaged region at the
base column corresponding to 5% drift for all tests except
Test results GS4UR, which failed at approximately 3.5% drift.
Figure 4 shows the experimental horizontal load versus The three pairs of linear displacement transducers at the
drift curves for the tested specimens. The lateral load was column base side computed the column average curvature
reported as measured by the load cell with no correction at three consecutive levels (low, intermediate, and high) at
220 mm
Stirrups 8 mm diameter
160 mm
50 mm
70 mm
50 mm
70 mm
at 50 mm center to center
18 mm diameter
Precast
concrete
1780 mm
column C
Stirrups 8 mm diameter
at 200 mm center to center
L
3200 mm
C
L 26 mm diameter
240 mm
80 mm
80 mm
Corrugated
steel sleeves
Displacement transducers at column base
900 mm
Stirrups 8 mm diameter
at 100 mm center to center
A A 215 mm High
250 mm
125 mm
125 mm
150 mm
2000 mm
215 mm (8.46 in.) spacing. For each level, the average cur- load due to the P-Δ effect.
vature was computed by dividing the difference between
the measurements of the two displacement transducers by Analysis of results
the level spacing (215 mm) and by the distance between
the transducers (150 mm [5.9 in.]). The curvature com- As expected, all of the columns had almost the same
puted at the lowest level comprises the strain penetration maximum lateral force capacity, though some differences
of the reinforcing bars into the foundation block because arose due to misalignment of the reinforcing bars during
the displacement transducers were connected directly to construction.
the foundation. The moment-curvature curves (Fig. 7) are
the envelope curves for each test, computed as the aver- The cast-in-place concrete column specimen CP failed at
age of both directions of excitation. The abscissa is made 5% drift due to the tensile failure of one of the reinforcing
dimensionless by multiplying the curvature by the column bars. This is likely due to low-cycle fatigue of the bar and
cross-section height B (400 mm [16 in.]). localized strain at the column base where a major crack
opened (Fig. 6). The column cyclic behavior was stable
Figure 8 shows the dimensionless energy for each test in up to 2.5% drift, and no damage other than base concrete
the drift range from 1% to 5%. The dimensionless energy spalling was observed. Significant pinching was present in
is defined as the ratio between the hysteretic energy (area the cycles after 2.5% drift.
inside each complete force-displacement cycle) and the
corresponding elastic energy in a semicycle (defined as one The grouted pocket foundation specimen PF showed a sta-
half of the maximum displacement times the corresponding ble response though a small amount of strength degrada-
lateral force), which accounts for the equivalent horizontal tion was registered in one loading direction after 3% drift.
A A A A
400 mm
400 mm
80 mm 80 mm 80 mm 80 mm
240 mm 240 mm
Grout
A A A A
400 mm
400 mm
80 mm 80 mm 80 mm 80 mm
240 mm 240 mm
Grout Grout
A A A A
300 mm Repair collar
400 mm
400 mm
unbonded
80 mm 80 mm 80 mm 80 mm
240 mm 240 mm
Grout Grout
Horizontal load, kN
0 0 0 0
-50 -50
P-Δ P-Δ
Bar failure
Horizontal load, kN
Horizontal load, kN
-50 -50
P-Δ P-Δ
0 0 0 0
-50 -50
P-Δ P-Δ
Bar failure
Figure 4. Lateral load versus column drift for the different connection types. Note: P-∆ = 2nd order load-displacement effects.
adopted in practice to increase construction speed even a value over which considerable damage is expected in
if full anchorage strength of the splice reinforcement is nonstructural members, such as cladding panels.
not provided. As expected, the limited anchorage length
resulted in strength degradation in the cyclic response. This The use of grouted steel sleeves on the column induced a
degradation is negligible for drift values smaller than 2.5%, major crack opening at the grout layer beneath the sleeves.
The compressive strain concentration at the column base
led to progressive damage at the unconfined grout next to
Horizontal load, kN
GS4 GS4B
Figure 6. Damage at 5% lateral drift. Specimen GS4U was repaired and renamed GS4UR; the latter experienced the failure of one reinforcing bar at a drift of 3.5%.
where
1.5
0.5
0
1.0 1.0 1.0 1.5 2.0 2.5 2.5 2.5 3.0 3.5 4.0 4.5 5.0
Drift, %
Figure 8. Dimensionless energy of the column-to-foundation connections. Note: Eelastic = elastic energy; Ehysteretic = hysteretic energy.
Column H
Mat foundation
LP
LSP
Фy Фp
Figure 9. Column-curvature idealization. Note: H = column height; Lp = plastic hinge length; Lsp = strain penetration length; øp = column plastic curvature;
øy = column yield curvature.
Horizontal load, kN
Drift ultimate (GS4U)
Drift ultimate (GS4B)
where
50 0
φy = column yield curvature
Δu = Δy + Δp (3) 25
where
0 -22.5
Δp = column displacement associated with plastic hinge 0 2 4 6 8
rotation Drift, %
= (φu − φy) LpH (4) Figure 10. Yield and ultimate drift for the grouted sleeve solution with and
without debonding.
where
φu = column ultimate curvature curs in the column outside the base section. Due to the
damage localization observed, the postseismic column
Equations (2) to (4) were applied to find the yield and repair for the grouted sleeve connection is simpler than
ultimate drift associated with specimens GS4B and GS4U, that for cast-in-place concrete or pocket foundation con-
that is, the grouted steel sleeve connection with and with- nections and can be effectively done by replacing the base
out the unbonded region. Figure 10 shows the performance layer of crushed grout with new high-strength grout. To
enhancement in terms of displacement capacity increase. further reduce the column damage and increase the column
The adoption of an unbonded length of reinforcement leads lateral displacement capacity, an unbonded length of steel
to a reduction of the curvature demand at a given imposed reinforcement can be used within the grouted sleeves. The
lateral displacement and therefore to a reduction of the proposed equations make it possible to take into account
column damage that is directly related to the curvature the unbonded length of reinforcement in the design process
demand. and to evaluate the column yield and ultimate displacement
when unbonded reinforcing bars are used.
In the case of debonding, the ultimate reinforcing bar
capacity is ensured, providing an appropriate develop- Acknowledgments
ment length in addition to the unbonded length. The
use of grouted steel sleeves contributes to reducing the The experimental program reported in this paper was part
development length due to the sleeves’ ability to restrain of a research program on precast concrete column-to-foun-
the splitting cracks in the grout rather than providing pas- dation connections financed by Moretti SpA in Erbusco,
sive confinement. Although the aim of the research is not Italy.
the determination of the bond strength associated with
grouted steel sleeves, the use of the sleeves, especially with The cooperation of Cristian Ratti, Andrea Zini, and Nicola
debonding, resulted in significant column lateral displace- Barzi in setting up the reaction frame and conducting the
ment without reinforcement bar failure. experimental tests is acknowledged.
Conclusion References
The experimental program showed that grouted steel 1. fib (International Federation for Structural Concrete)
sleeves are suitable as column-to-foundation connections Commission 7. 2003. Seismic Design of Precast
in seismic regions. The high ductility of the grouted steel Concrete Building Structures, International Federa-
sleeve connections is related to the confining effect of the tion for Structural Concrete. Bulletin 27. Lausanne,
corrugated steel sleeves on the grout. Furthermore, the Switzerland: fib.
presence of a highly confined grout prevents buckling of
the longitudinal reinforcement. 2. PCI Connection Details Committee. 2008. PCI Con-
nections Manual for Precast and Prestressed Concrete
In such connections, the damage was localized to the Construction. 1st ed. Chicago, IL: PCI.
20 mm (0.8 in.) grout layer between the precast concrete
column and the foundation. As a result, little damage oc-
Notation
H = column height
Keywords
T
he performance of a structure is mainly linked to its
strength, stiffness, and ductility, although function-
ality, aesthetics, speed of construction, and econo-
my are also significant.
■ To validate the quality and performance of this type of structure, It is well known that the economic performance of a struc-
an extensive program of theoretical analyses and structural tural system greatly depends on the quantities of materi-
tests (including shake table tests) was conducted. als per square meter (for example, concrete, reinforcing
bars, and strands), but the real construction cost can also
■ Due to its functional and technical advantages, high productiv- depend on speed of construction, local market conditions,
ity, and low cost, the building system has been used in more availability of labor and specialized equipment, and other
than 500 flats in Transylvania, Romania. factors.
Precast concrete shear walls panels spanning between adjacent columns. Through their
horizontal connections, the panels provide the structural
The shear wall comprises large reinforced precast concrete wall continuity along its height. Through their vertical
Figure 4. Precast concrete flat-slab panel and connections between the panels.
connections (column-panel joints) they make up a dual between precast concrete members should be determined
structural system, combining frames with structural walls by analyses or by tests.”7
(Fig. 2 and Fig. 5).
In addition, section 21.2.1.5 says that a precast concrete
Experimental program structural system “may be used for earthquake resistance if
it is demonstrated by experimental evidence and analysis
Objectives of the program that the proposed system will have strength and toughness
equal to or exceeding those provided by a comparable
The structural behavior of a system that incorporates pre- monolithic reinforced concrete structure.”7
cast concrete members may differ substantially from that
of a comparable monolithic cast-in-place system. The pre- To assess the structural quality and the performance of this
cast concrete members should be assembled and connected structural system designed for seismic zones, the research
to produce a structural system capable of resisting vertical team conducted an extensive analytical and experimental
and lateral forces due to shrinkage, creep, temperature study. The research program2 included tests on column-
changes, elastic deformation, wind, or earthquake. to-column joints, experimental tests on a precast concrete
structural wall subassemblage, postelastic computer analy-
The design and detailing of the connections to transmit sis of precast concrete structural walls, tests on a full-scale
such forces are especially important to performance. assemblage of columns and precast concrete flat slabs
Connections are essential to the integrity of the structural subjected to gravity forces, shake table tests on a struc-
system. ture prototype (scale 1:4), and in-place dynamic tests on
a real five-story building. In this large scientific program,
Taking into consideration the novelty of the proposed researchers from three universities and four institutes
system and the overall behavior of the precast concrete and structural designers from two design institutes were
structure, “the adequacy of connections to transfer forces involved.
Flat-slab structure ally good behavior of the structure under vertical loads and
under gravity loads underlines the large load-bearing capacity of the structure.
The full-scale test of an assemblage of precast concrete The shape of the gravity load–vertical deflection P–Δv
columns and slabs was performed to observe the behavior diagrams reveals an elastic behavior of the structure with
of the precast concrete flat-slab panels, the connections 26.80% to 39.70% permanent irreversible deformations
between panels, and the column-to-slab connections under and 60.30% to 73.20% elastic recovery (Fig. 7).
gravity loads. The behavior up to the failure of the entire
floor system and the problem of slab punching were also The maximum vertical displacement Dvmax of 7.3 mm
studied. (0.29 in.) does not exceed the allowable vertical displace-
ment Dva of L/350 equal to 9.3 mm (0.36 in.) given by
The assemblage comprised eight square precast concrete the Romanian design code for concrete structures STAS
slab panels of 3.60 m (12 ft) each and 15 square columns 10107/0-90.11 The monolithic connections between the
of 350 mm (14 in.) (Fig. 6). slab precast concrete panels, which have 8 mm loops at
150 mm (0.3 in. loops at 6 in.) spacing and longitudinal
The authors used the equivalent frame method accepted reinforcing bars, ensure complete integrity of the precast
by many design codes, including the American Concrete concrete floor, including its flexural and shear integrity.
Institute (ACI 318-08),7 European (EN 1992-1-1:2004),8
and Romanian standards (SR EN 1992-1-1:20049 and Dual precast concrete
STAS 10107/1-90).10 The designers calculated theoretical structural walls’ subassemblage
values for cracking load P crd of 5.50 kN/m2 (115 lb/ft2) and under lateral loading
ultimate gravity load P ud of 7.70 kN/m2 (160 lb/ft2).
The experimental program focused on the general behav-
The strains in the concrete and reinforcing bars, the verti- ior of the precast concrete structural walls under static
cal displacement of the precast concrete panels, and the alternate-reversed loadings in all ranges—elastic, crack-
displacements of the connections between the slab panels ing, postelastic, and failure—as well as the behavior of the
were recorded. Figure 7 shows a typical applied load-dis- vertical connections, horizontal connections, and joints
placement diagram recorded during the test (Fig. 6) at the between precast concrete column members.
midspan of the panel.
The tested subassemblage was designed to model at scale
The first cracks developed at 4.40 kN/m2 (92 lb/ft2) in the 1:2.5 the structural behavior of a real precast concrete
longitudinal edge connections and at 9.90 kN/m2 (210 lb/ structural wall with eight stories. The unit’s components
ft2) in the middle of the panel. The experimental ultimate are monolithic concrete foundations under each shear wall,
gravity load P uexp of 15.90 kN/m2 (332 lb/ft2) was reached precast concrete columns of two or three stories for each
in the last loading stage and corresponded to the initiation member, precast concrete shear wall panels on the height
of slab punching in the vicinity of columns. The ratio of of a floor, and precast concrete slab panels (Fig. 8).
2.10 between the experimental and the theoretical (design)
values of the ultimate load ( P uexp /P ud ) confirms the gener- The loading of the model was accomplished by gravity
loads Ng (dead load) so that the same compressive stress The large subassemblage, measuring 2.02 m × 1.62 m ×
as in a real eight-story shear wall could be reached at each 5.95 m (6.62 ft × 5.31 ft × 19.50 ft), was loaded vertically
level. and horizontally (Fig. 8). The principal observations and
conclusions include the following:
The vertically applied forces N1 (Fig. 8) compensate in
the model for the effect of the reduction of the upper three • The cracking extended only along the height of three
floors of the real structural wall. These adjustments from stories. Although it was all–precast concrete, the
the real structure to the model did not alter the results structural wall subassemblage (column–shear wall
of the experiment. On the lower third of the shear wall, panels–column) behaved at cracking as a monolithic
the part of maximum interest, the bending moments are structural wall, the cracks of the columns being contin-
practically equal and the shear forces are generally slightly uous through connections and through precast concrete
higher than in the real structure. panels. The cracks are typical of bending.
Previous tests (Maniu et al.)12 had shown that the cracking • The main cracks that finally led to failure developed
does not extend beyond the lower third of the structural only after the imposed 3Δy cycle (Fig. 9). In the
wall. The horizontal load H was applied in alternate cycles monolithic area (vertical connections), the cracks were
following the principle of imposed displacements, where narrow, having a width under 0.20 mm (0.0078 in.)
Δ is the top lateral displacement of the experimental model throughout the test (Fig. 10).
and Δy is the top lateral displacement at the first yield of the
steel reinforcement (Fig. 9). • The column-to-column joint behaved well. The cracks
were insignificant in two of the joints and totally absent
in the other two joints.
• The lateral stiffness of the experimental specimen The dynamic response of this precast concrete dual system
was remarkable. After 23 cycles of loading, the lateral was established during four static tests, five tests to deter-
interstory drift measured at the uppermost level in the mine the dynamic characteristics of the structure, and 17
4Δy cycle was only 1/205. dynamic tests.
• The ultimate displacement ductility factor Δu/Δy (where The dynamic tests took 900 seconds in all and had about
Δu is ultimate top lateral displacement) attained dur- 70 acceleration peaks a0max of more than 1 m/sec2 (3 ft/
ing the test was 5.25. The structural subassemblage sec2) from those 1200 peak accelerations applied. The real
behaved well from this point of view, too. precast concrete structure and the experimental prototype
were designed for a seismic intensity of VIII on the modi-
Shake table tests fied Mercalli intensity scale (MM scale).
on a five-story prototype
The experimental program included dynamic tests with an
The most comprehensive experimental program was the artificial Housner-Jennings type earthquake (a0max of 1.108
testing of a 1:4 scale model of a five-story building on a ÷ 2.632 m/sec2 [3.635 ÷ 8.635 ft/sec2]); 1977 Vrancea,
140 kN (315 kip) shake table (Fig. 12). The model struc- Romania, earthquake (a0max of 2.775 ÷ 3.885 m/sec2 [9.104
ture was tested in the transverse direction. ÷ 12.75 ft/sec2]); 1940 El Centro, Calif., earthquake (a0max
of 2.775 m/sec2 [9.104 ft/sec2]); artificial Jassy, Roma-
nia, earthquake (a0max of 4.163 ÷ 6.383 m/sec2 [13.66 ÷
20.94 ft/sec2]); and harmonic type earthquake (a0max of The most significant results and conclusions are as follows:
1.943 m/sec2 [6.375 ft/sec2]). The shake table tests in-
creased in intensity from V to X on the MM scale. • The static tests intermingled among the dynamic tests
did not produce damage in the experimental model.
Figure 13 presents lateral displacements recorded for
each floor, where A corresponds to the elastic stage,
B was conducted after cracking, C corresponds to
the elastic-plastic stage when the columns and shear
panels have open cracks, and D was performed in the
presence of plastic zones in the structure. By compar-
ing the static deformations recorded at different stages
of structural behavior (A, B, C, and D), a lateral stiff-
ness degradation of 12.50% was observed in the final
failure stage (Fig. 13).
the ultimate limit state, where the allowable interstory spans of 3.60 m (11.8 ft), ten bays of 3.90 m (12.8 ft), and
drift is 2.5%, even for earthquakes of X on the MM five stories of 3.00 m (9.84 ft) high.
scale, where the measured maximum interstory drifts
are of 0.82% to 1.25%. Comparisons have been made between the dynamic char-
acteristics of a precast concrete structure comprising only
• The tests to determine the dynamic characteristics structural columns, panels, and walls (the first stage) and
within five different stages of structural behavior, from
the elastic to the plastic stage, showed that the natural
frequency of the model diminished from 6.50 Hz to
5.00 Hz in the transverse direction and from 6.00 Hz
to 5.03 Hz in the longitudinal direction. At the same
time, the critical damping ratio increased from 1.38 to
3.10 in the transverse direction and from 1.00 to 3.02
in the longitudinal direction. Despite reductions of fre-
quency by 23% transversally, 16% longitudinally, and
29% in torsion, the stiffness degradation after a series
of high-intensity seismic actions was only 12.50%,
which shows that the structure can be consolidated and
its structural integrity can be restored.
To assess the quality and performance of this system, an 1. Crainic, L. 2003. Reinforced Concrete Structure. Cluj-
extensive program of theoretical analyses, structural tests, Napoca, Romania: Napoca Star.
and applied studies to improve the construction technology
were conducted. These studies maximized the effect on 2. Tripa, E. 2002. Contributions to the Study, Design and
construction productivity and minimized the total cost. Erection of Buildings with Precast Concrete Mixed
Structure. [In Romanian.] PhD thesis. University Po-
A full-scale test on an assemblage of precast concrete litehnica from Timis̹oara (UPT), Timis̹oara, Romania.
columns and precast concrete flat slabs confirms the gener-
ally good behavior of the structure under gravity loads and 3. Lungu, D., C. Arion, S. Demetriu, and A. Aldea. 2003.
underlines the large load-bearing capacity of the structure. Probabilistic Zonation of Vrancea Seismic Hazard.
The ratio between the experimental and the theoretical Eurocode 8, Representation of Design Actions. In Pro-
values of the ultimate load is 2.10. ceedings of the International Conference: Construc-
tions 2003, July 16–17, 2003. Cluj-Napoca, Romania:
The large subassemblage of two precast concrete, five- Napoca Star and Argonaut.
story structural walls; four precast concrete columns; and
five precast concrete flat slabs was tested up to failure. The 4. Ministry of Public Works and Territory Management
loading consisted of constant vertical forces and horizontal (MLPAT). 1992. Seismic Design Code (P100-92). [In
forces applied in alternate cycles following the principle Romanian.] Bucharest, Romania: MLPAT.
of imposed displacements. The test on this model (scale
1:2.5) confirmed its structural performance with respect to 5. Ministry of Transportation, Construction and Tourism
the load-bearing capacity, stiffness, ductility, and cracking. (MTCT). 2006. Seismic Design Code—Part I: Provi-
The experimental values regarding the ultimate capacity and sions for Buildings (P 100-1/2006). [In Romanian.]
lateral displacements are in good agreement with the theo- Bucharest, Romania: MTCT.
retical ones. The hysteretic loops (H–Δ) are stable, and the
prototype demonstrated a large energy absorption capacity. 6. Federal Emergency Management Agency (FEMA).
2000. Prestandard and Commentary for the Seismic
The types of connections proved suitable, resulting in a Rehabilitation of Buildings. FEMA 356. Washington,
behavior for the precast concrete subassemblage that was DC: FEMA.
similar to that of the monolithic solution.
7. American Concrete Institute (ACI) Committee 318.
The dynamic tests carried out on the 140 kN (315 kip) 2008. Building Code Requirements for Structural
shake table at INCERC-Jassy, where a five-story precast Concrete (ACI 318-08) and Commentary (ACI 318R-
concrete structure (scale 1:4) was tested, showed that 08). Farmington Hills, MI: ACI.
the lateral stiffness degradation of the structure was only
12.5% in the final failure stage. The tests took 900 seconds 8. European Committee of Standardization (CEN). 2004.
and had 70 acceleration peaks of more than 1 m/sec2 EN 1992-1-1: 2004. In Eurocode 2: Design of Con-
(3 ft/sec2). The interstory drifts recorded during the tests crete Structures—Part 1-1: General Rules and Rules
satisfy the requirements of the Romanian Seismic Design for Buildings. Brussels, Belgium: CEN.
Code P 100-1/2006.5
9. Romanian Standards Association (ASRO). 2004. SR
In-place tests on a real structure erected in the city of Deva EN 1992-1-1: 2004. In Eurocode 2: Design of Con-
indicated that the natural periods of vibration of this new crete Structures—Part 1-1: General Rules and Rules
type of precast concrete structure are similar to the values for Buildings. [In Romanian.] Bucharest, Romania:
determined in Bucharest on reinforced concrete frame ASRO.
structures with infill masonry panels. These structures ex-
11. IRS. 1990. Design and Detailing of Concrete, Rein- Dvmax = experimental maximum vertical deflection
forced and Prestressed Concrete Structural Members.
In Civil and Industrial Buildings. STAS 10107/0-90. Δ c = cracking displacement
[In Romanian.] Bucharest, Romania: IRS.
Δu = ultimate top lateral displacement
12. Maniu, H., A. M. Ioani, and E. Tripa. 1992. Structural
Investigation on a Seismic Performance of Precast Δu/Δy = displacement ductility factor
Mixed Shear Walls. In Proceedings of the Tenth World
Conference on Earthquake Engineering, Madrid, July Δy = top lateral displacement at first yield
19–24, 1992. Rotterdam, Netherlands: Balkema.
Notation
P ud
= design (calculated) ultimate gravity load for the
full-scale assemblage
Dexp
max
= e xperimental maximum (ultimate) top lateral dis-
placement
Keywords
Eusebiu Tripa, PhD, P.E., FA-
ICPS-Romania, is a senior Connection, dual system, flat-slab panel, loading,
structural engineer at Tripexpert seismic.
Design Office in Deva, Romania.
Review policy
E
xterior enclosure systems for buildings often
consist of elements that are supported by the build-
ing structure. For high-rise buildings, support of
the exterior enclosure by the building structural frame is
necessary, but for low-rise buildings, self-support of the
exterior enclosure may be possible and could result in cost
savings in the structural frame due to the reduced demand
on its members. For self-supported systems, the structural
frame typically provides lateral support at each floor and
the foundation system supports the enclosure gravity loads
directly.
It is best to have redundancy in the SSLFRE system. Typi- When designing a precast concrete SSLFRE system, it is
cally, all or nearly all of the precast concrete enclosure is desirable to avoid or limit structural connections between
part of the SSLFRE system. This way stresses in individual adjacent horizontal panels (at vertical joints). If enough
components are low and the system is not compromised if panels are used for the lateral-force-resisting system, de-
one or two panels are damaged. Also, it is best if most of sign stresses will be limited such that the vertical stacks of
the panels are similar to provide uniform force distribution panels can act independently. This independence signifi-
and improve the redundancy performance. If an SSLFRE cantly reduces temperature performance effects, compared
system is proposed, the architect and the owner or user with connecting all of the panels together, and simplifies
should be aware and in agreement because changes during the design of the panels for vertical and lateral loads. If ad-
the design or in the future may be restricted. jacent panel connections are necessary to achieve appropri-
ate stress limits in the panels, one should consider connect-
Also, if a panel needs to be removed at some point in an ing only two or three vertical stacks so that temperature
SSLFRE system, the removal is more difficult than for deformations have regular relief along the length of the
systems in which panels are individually connected to the wall. Stresses should be limited to avoid cracking of the
structural frame. precast concrete under normal load conditions. Depending
on the project, cracking under seismic conditions may be
Structural design considerations acceptable.
For gravity design of precast concrete SSLFRE systems, For temperature performance reasons, it is suggested that
the weight of the system will be carried by the vertical each vertical stack (or stacks, if connected together later-
elements within the panels; therefore, the vertical elements ally) should possess only one lateral shear force connec-
Consequently, ordinary precast concrete shear walls do not • foundation support design assumptions
have to comply with seismic specifications of the American
Concrete Institute’s (ACI’s) Building Code Requirements • forces (or alternately the reinforcement) at the panel joints
Column
column
Double
Brace Brace
Rigid
Rigid
Rigid
Beam Beam
Rigid
Rigid
Beam Beam
Rigid
Rigid
Rigid
Column
Column
column
Double
Rigid
Rigid
Rigid
Beam Beam
Figure 2. Detail of the structural analysis model showing supports for the panel-to-panel connections.
Example analysis and design The panels spanned one bay horizontally (between two
building columns) and were one story tall. For the panels
The following description represents a recently completed to perform as a lateral-force-resisting system, the inter-
project that used a precast concrete SSLFRE system. The nal panel forces were transferred vertically between the
exterior enclosure for this multibuilding project consisted stacked panels, and each panel was connected to the struc-
of precast concrete panels with punched windows. The ture so that diaphragm forces were transferred to the pan-
buildings ranged from four to six stories tall and had con- els. Therefore, the panel-to-panel and panel-to-structure
siderable repetition in the exterior enclosure. For confi- connections were critical. Because the panels did not carry
dentiality purposes, the location and name of the project is gravity load from the structure, uplift forces were present.
withheld at the request of the owner.
The performance of the panels under service wind loads
General description of the panels was important. Because the panels served as the facade of
the building, cracking of the precast concrete under service
The precast concrete panels were typically 30 ft wide × loads was not allowed.
14 ft tall (9.1 m × 4.3 m). The windows were about 13 ft
wide × 8 ft tall (4 m × 2.4 m) with two windows in each In addition to the typical panels (Fig. 1), there were solid
panel. The panel profile reflects the architectural require- panels without windows, panels with louvers, and panels
ments (Fig. 1). with doors. Panels that were not suitable for use as SSL-
FRE elements were hung from the structural frame.
Structurally, the panel was considered as a weak pier-
strong spandrel type of wall, with three columns and two The shear wall panels were self-supported and designed to
beams. The edge columns were about 12 in. × 12 in. stack on top of each other. Only two supports were consid-
(300 mm × 300 mm), while the center column was about ered, one at each end of the panel. The tension or compres-
24 in. (610 mm) wide × 12 in. thick. The spandrel beams sion force at the panel piers resulted from the combination
were 8½ in. (220 mm) thick with ½-in.-deep (13 mm) of the weight of the panels and the vertical seismic or wind
reveals and a border element 12 in. deep and 12 in. thick. loads.
Structurally, the spandrels were modeled with a constant
cross-sectional thickness of 8 in. (200 mm). The design High-density plastic shims placed between panels trans-
compressive strength of the concrete was 5000 psi (35 MPa). ferred the compression load to the panel below, while a
vertical-shear connection transferred the tension load. The IBC 2003 does not specify seismic design coefficients and
numerical analysis model reflected the lack of moment factors for a lateral-force-resisting system based on precast
transfer between panels at the mechanical connection. concrete shear walls. However, UFC 3-310-04 lists a response
modification factor R of 4, an overstrength factor Ω of 2.5,
Just one connection at one end of each panel transferred and a deflection amplification factor Cd of 4 for ordinary
the shear force from the diaphragm to the panel (the panel- precast concrete shear walls in the building frame system cat-
to-structure connection). egory. An ordinary precast concrete shear wall in the bearing
wall systems category has the following design coefficients: R
This configuration allowed the panels to expand under equal to 3, Ω equal to 2.5, and Cd equal to 3. The provisions
temperature increases without being constrained by the of UFC 3-310-04 are shared by ASCE/SEI 7-02.
structure. This single panel-to-structure shear connection
resulted in a concentration of stresses in the diaphragm at The project is located in an area with a seismic design
the connection location. category of B. The buildings are four to six stories high,
and the total seismic base shear in the applicable direction
Seismic properties is about 4% of the weight of the building.
of precast concrete shear walls
The buildings have, in general, a rectangular shape with
The project was designed in accordance with Interna- dimensions of about 120 ft wide × about 210 ft long (36 m
tional Building Code 20033 and Department of Defense’s × 64 m) (seven bays at 30 ft [9 m]). Precast concrete shear
Seismic Design for Buildings (Unified Facilities Criteria walls were used along both longitudinal sides of each
[UFC] 3-310-04).4 building except at loading docks and entrances.
The precast concrete pier-spandrel panel system was mod- The models were first subjected to the lateral forces from
eled in a structural analysis program with frame elements. service wind load (designed wind velocity of 90 mph
[150 kph]) combined with the self-weight of the panels
Minimum reinforcement for columns and deep beams (about 1 kip/ft [15 kN/m]). All of the panels were sub-
was required for these elements. The frame elements were jected to a combination of axial compression and moment.
modeled at their center of gravity. Rigid pinned elements Tension forces were not present under wind loads alone.
connected two panels at the end columns, and rigid pinned Cracking was limited by the following equation:
braces were added for lateral stability. Shear and tension
or compression forces could then be transferred with this fr = 7.5 fcl = 7.5 5000 = 530 psi (3660 kPa)
configuration while moments would not (Fig. 2). ACI 318-05 Eq. (9-10)
Cast-in-place concrete
Interior finish slab to be blocked out at
connection locations and
concrete filled after
precast concrete
erection
Top of slab
Metal decking
Sheet metal
pour stop
blocked out at
W-beam
connection
locations
with the height and ranged from 25 kip to 40 kip (110 kN walls were used at panels with punched windows, while
to 180 kN). provisions for minimum reinforcement of walls—includ-
ing boundary elements—were used at panels with no
The horizontal shear connection between each panel, as windows. The required reinforcement was specified on the
well as at the base, was established for each building. For structural drawings.
the upper two floors, shear values ranged from 20 kip to
35 kip (90 kN to 160 kN). For the lower floors, maximum Although each panel behaves like a strong beam–weak
base shear values ranged from 45 kip to 75 kip (200 kN to column moment frame, the global system (collection of
330 kN). stacked panels) does not. There is no moment continuity
at the columns along the height of the structure or at the
Finally, the panel reinforcement was checked. Mini- beams along the length of the building. The overturning
mum axial and shear reinforcement were determined to moment from seismic and wind demand is resisted through
be adequate to resist the imposed loads. Provisions for axial forces only at the ends of each panel, like boundary
minimum reinforcement of beams and columns rather than elements in shear wall systems.
Interior finish
Slab on grade
Cast-in-place concrete
foundation
When using an SSLFRE system, a few points should be Cd = deflection amplification factor
considered during construction. Because the SSLFRE sys-
tem only provides lateral-force resistance for the structure fcl = compressive concrete strength
in the completed state, temporary bracing requirements for
the structural frame may be increased. Temporary bracing fr = cracking stress of concrete
will be required until completion of the exterior enclosure
installation. R = response modification factor
Acknowledgments
T
o evaluate the structural performance of prestressed
concrete members, it is often necessary to experi-
mentally determine the deformations in the member
due to applied forces. One typical example of this is the
determination of the transfer length in prestressed concrete
members. The transfer length is defined as the distance
required to transfer the fully effective prestress force in the
prestressing strand to the concrete.1
Transfer lengths are determined by measuring concrete sur- Optical speckle techniques have evolved into powerful
face strains at the ends of actual members or prismatic test tools for the measurement of surface strain since digital
specimens. Metal discs called gauge points are typically image recording and processing have become widely avail-
secured to the surface of the specimens at 2 in. (50 mm) able. They have the advantage of minimal surface prepa-
spacing before releasing the tension in the strands. These ration, work with almost all kinds of rough surfaces, and
points are typically mounted using epoxy or are directly have high resolution.11–13
embedded into the concrete. They are located at the struc-
tural depth of the prestressing steel (Fig. 1). Speckle is generated by illuminating a rough surface with
coherent light (Fig. 3). The random reflected waves inter-
Distances are measured between the gauge points using a fere with each other, resulting in a grainy image (Fig. 4,
mechanical gauge with a typical resolution of about 20 μє. lower right). The speckle pattern could be thought of as
Surface strain readings are usually taken before tension a fingerprint of the illuminated area in the sense that the
release, immediately after release, and then periodically speckle pattern produced by every surface area is unique.
during the first few months after tension release. Furthermore, when the surface undergoes movement or
deformation, the speckle pattern in the image plane will
The determination of transfer lengths in pretensioned also move or deform accordingly. Thus the displacement
concrete members has been done using the current proce- or deformation information of the object surface can be
dure for more than 40 years. This method is time consum- extracted by measuring the speckle pattern movement.
ing and is subject to considerable human judgment and
possible errors because the mechanical readings are taken At the present time, no speckle strain technique had been
and manually recorded. In addition, the labor-intensive applied to prestressed concrete surface measurement, ex-
process of installing the gauge points and taking repetitive cept possibly in a very controlled laboratory setting. This
is because prestressed concrete surface strain measurement disrupts the in-plane displacement measurement.
has some characteristics that make it difficult for a regular
optical sensor to be used: Therefore, an optical strain sensor for the prestressed con-
crete surface strain measurement must be removable from
• The tension release process during the production of the specimen, have large dynamic range, and be insensitive
the prestressed concrete members involves the sudden to surface tilt.
release of large forces into the member. These forces
produce intense vibrations that disrupt the relative An optical strain sensor based on a five-axis freedom
position of the sensor and the concrete surface. To
protect the optical sensor from damage, the sensor
must be removed from the concrete surface after the Coherent light
source
initial reading and before the tension release process.
This prevents some optical strain sensors, designed
to be attached to the specimen surface throughout the In phase-
bright speckle
measurement process, from being used.14
The pretensioned concrete members were each 9 ft 6 in. Because the LSI technique relies on the optical pattern
(2.9 m) long. The transfer lengths were measured on one recognition of images before and after movement, changes
side of each member using both the traditional Whittemore or weathering to the concrete surface could limit the ability
strain gauge and the noncontact LSI method. To facilitate of this technique to measure long-term effects. However,
the laser-speckle measurements, an aluminum rail was
mounted to the side of the member (Fig. 7 and 8). The rail
was attached to the members using small ¼-in.-diameter
(6 mm) inserts that were cast into the sides of the preten- 8 in.
sioned concrete members.
Results
41/2 in.
Surface strain measurements for the trapezoidal specimens 6 /2 in.
1
were obtained using both the standard (Whittemore) strain-
gauge technique and LSI (optical) technique. Figure 9
shows that the LSI technique results in a smoother data
line with less scatter than that generated using the existing
surface strain measurement technique with the Whittemore 3 in.
strain gauge. The LSI technique has been validated on
members cast in both indoor and outdoor operations. Ther-
mal effects, due to temperature differences between the 21/2 in. 21/2 in.
measuring instrument and measured object, are similar to
those exhibited by the traditional Whittemore strain gauge
Figure 6. Trapezoidal cross section of the pretensioned members used to verify
method, as the coupling of the gauge points in the optical the laser-speckle transfer length measurement technique. Note: 1 in. =
25.4 mm.
the LSI method works well during the first month after The LSI method offers a significant improvement in the
tension release. The peak strains varied along the length of reliability of estimating the transfer length.
the member, producing an asymmetric shape. This was due
to a slight horizontal eccentricity of the strand in the small The researchers are working to automate the process of
trapezoidal cross section, which produced biaxial bending traversing along a concrete member and capturing the cor-
in the member. responding LSIs in increments of ¼ in. (6 mm). This will
enable a near-real-time determination of transfer lengths
Conclusion through computerized postprocessing of the digital im-
ages in the field. It is envisioned that LSI will become an
The LSI technique is a viable method to measure transfer effective quality-control technique to screen out deleteri-
lengths in pretensioned concrete members. The LSI tech- ous combinations of strand and concrete mixtures and to
nique, using the current optical arrangement, is more ac- determine the effect of changes in these parameters on
curate than the existing Whittemore strain gauge technique strand bond.
because it eliminates human bias and improves repeat-
ability of measurement. Furthermore, the technique does Acknowledgments
not require extensive operator training to achieve reliable
measurements, as does the Whittemore strain gauge tech- This project was funded through the Kansas Department
nique. The accuracy of the LSI technique has been shown of Transportation (KDOT) and a PCI Daniel P. Jenny Fel-
to be less than 10 μє, compared with a 25 μє accuracy for lowship. Specifically, the authors would like to thank Brian
the Whittemore strain gauge technique. Much of the error Miller and Paul Johal of PCI and David Meggers of KDOT
associated with the use of the Whittemore strain gauge for helping to make this a successful project.
technique is due to poor repeatability, resulting in a large
random error contribution. Thus, without operator training,
it is difficult to get reliable transfer length measurements
with the traditional Whittemore strain gauge technique.
The authors would also like to thank Ed Volkmer, former 4. Larson, K., R. Peterman, and A. Esmaeily. 2007. Bond
graduate student at Kansas State University, who assisted Characteristics of Self-Consolidating Concrete for
with the fabrication of the prestressed concrete members Prestressed Concrete Bridge Girders. PCI Journal, V.
and the initial data collection and reduction. 52, No. 4 (July–August): pp. 44–57.
600
500
400
300
200
Figure 9. Comparison of raw (unsmoothed) strain measurements immediately after detensioning of a pretensioned specimen.
9. Oh, B. H., and E. S. Kim. 2000. Realistic Evalua- Systems for Industrial Inspection, ed. Wolfgang Osten,
tion of Transfer Lengths in Pretensioned, Prestressed pp. 30–36. Munich, Germany: Society of Photo-Opti-
Concrete Members. ACI Structural Journal, V. 97, No. cal Instrumentation Engineers.
6 (November–December): pp. 821–830.
15. Zhao, W., B. T. Beck, and J. Wu. 2004. A Novel Opti-
10. Kaar, P., and D. Magura. 1965. Effect of Strand Blan- cal Technique for Measuring 5-Axis Surface Move-
keting on Performance of Pretensioned Girders. PCI ment. In Proceedings of SPIE Optics East: Two- and
Journal, V. 10, No. 6 (December): pp. 20–34. Three-Dimensional Vision Systems for Inspection,
Control, and Metrology II, ed. Kevin G. Harding, pp.
11. Yamaguchi, I. 1981. A Laser-Speckle Strain Gauge. 66–73. Philadelphia, PA: Society of Photo-Optical
Journal of Physics E: Scientific Instruments, V. 14, No. Instrumentation Engineers.
11 (November): pp. 1270–1273.
16. Wu, C.-H., W. Zhao, T. Beck, and R. Peterman. 2009.
12. Johnson, Peder. 1998. Strain Field Measurements Optical Sensor Developments for Measuring the Sur-
in Industrial Applications Using Dual-Beam Digital face Strains in Pretensioned Concrete Beams. Strain,
Speckle Photography. Optics and Laser in Engineer- International Journal for Experimental Mechanics.
ing, V. 30, No. 5 (November): pp. 421–431. Published electronically March 27, 2009.
13. Sjödahl, M. 1999. A Whole Field Speckle Strain Sen- 17. Peterman, R. J. 2009. A Simple Quality Assurance
sor. In Proceedings of SPIE: Optical Engineering for Test for Strand Bond. PCI Journal, V. 54, No. 2
Sensing and Nanotechnology (ICOSN ’99), ed. Ichirou (Spring): pp. 143–161.
Yamaguchi, pp. 84–87. Yokohama, Japan: Society of
Photo-Optical Instrumentation Engineers.
Weixin Zhao, PhD, is a postdoc- The current experimental method to determine the
toral research associate in the transfer length in prestressed concrete members
Department of Mechanical and consists of measuring concrete surface strains with
Nuclear Engineering at Kansas a mechanical strain gauge before and after releasing
State University in Manhattan, tension. Because this is a time-consuming and tedious
Kan. process, transfer lengths are seldom measured on a
production basis. Furthermore, when transfer lengths
Kyle Larson, PhD, was a postdoc- are determined using the current method, the times to
toral research associate in the release tension of the members being measured are
department of Civil Engineering often delayed, thereby resulting in artificially higher
at Kansas State University. release strengths for the members evaluated.
for the precast concrete structures industry and as such, PCI Certification
is an integrated and ongoing part of the industry’s body of knowledge.
Check One:
Professional Member: Any person who is a licensed professional engineer or
architect, or who meets the education and experience requirements deemed by the LICENSED ENGINEER
Board of Directors to be equivalent to those of professional licensure. LICENSED ARCHITECT
PHoNE fAX
CITY, ToWN or PosT offICE, P.o. NUMBEr If APPLICABLE, AND ZIP CoDE
WEB ADDrEss
ANNUAL FEE = $110 (NON-U.S. = $140) (U.S. ACADEMIC EMPLOYEE = $50) (NON-U.S. ACADEMIC EMPLOYEE = $65).
PAYMENT IN U.S. DOLLARS DRAWN ON A U.S BANK MUST ACCOMPANY THIS APPLICATION.
Fee is not deductible as charitable contributions for federal income tax purposes but may be deductible by members as a business expense.
This certifies that the company with which I am affiliated does NOT qualify for Producer membership or
Associate membership as described in the Bylaws of the Precast/Prestressed Concrete Institute. Applicant’s initials
RECENT GRADUATES:
Your first year of PCI membership is FREE!
You’ve invested a great deal in getting to where you are now—contributing thousands of hours, financing tuition fees, and
Now it’s time to start reaping the benefits of your studies and hard work. When you become an Associate Profes-
sional Member of PCI, you’ll gain valuable opportunities to develop your career and network with more than 1,800
design professionals. And you’ll become part of a vital and growing industry that’s looking for talented professionals
like you.
From these humble beginnings emerged the spark for a "grand plan"
to provide for the Industry's future success by helping ensure the continued
availability of talented people and new technology...
HOW TO PURCHASE:
Contact Skip Francies at sfrancies@tampabay.rr.com or
call at 813-230-8266. If there is no answer, please leave a
detailed voice mail stating the following:
Skip Francies
6136 Kestrelridge Drive
Lithia, FL 33547
1 to 49 ..................................................$5.00 each
50 to 99 ................................................$4.50 each
100 plus ................................................$4.00 each
Coming ahead Board of Directors
Durability Chairman Gregory F. Force
Vice chairman Tom H. Kelley
• Transverse Post-tensioning Arrangement for Side-by-
Secretary-treasurer Dean Gwin
Side Box-Beam Bridges
Immediate past chairman Donna S. Reuter
• Simplified Transverse Post-tensioning Construction
PCI president James G. Toscas
and Maintenance of Adjacent Box Girders
• Spalling of Partial-Depth Precast, Prestressed Bridge Keith G. Allen Mason Hardaway Lampton
Deck Panel Investigation Using Ground Penetrating
Michael Baez Paul J. Lautenbach
Radar
• Proposed Evaluation and Repair Procedures for Jay M. Cariveau Charles R. Lowe
Precast, Prestressed Concrete Girders with End Zone
Todd Culp Marianne Methven
Reinforcement
Francesco J. Genoese Alexander G. Mihaylov
Also
Harry A. Gleich Christopher J. Pastorius
• Torsion Tests on Spun-Cast Prestressed Concrete
Poles Leon H. Grant Chuck Prussack
• Effect of Environmental Conditions on Field Welding
Dean Gwin Larbi Sennour
of Precast Concrete Connections
• The MnDOT Inverted-T System: Parametric Studies Dan Juntunen Rita Seraderian
for Preliminary Design
Daniel L. Kennedy C. Douglas Sutton
Edward S. Knowles Keith E. Wallis Jr.
Technical Activities Daniel P. Kolb Peter Yurkiw
Council Bohdan Kusznir
Kirstin Osgood, sales and member development manager kosgood@pci.org (312) 360-3206
You can find the contact information for the PCI Journal and PCI Publications department staff in
the table of contents.
When Doug Sutton sits down in a PCI meeting, it is He received his PhD in 1972,
rare that he doesn’t recognize at least one of his former along with a National Science
students. Sutton, who retired in 2007, taught structural Foundation Fellowship, and has been
engineering at Purdue University for 46 years and intro- formally recognized many times with
duced many of today’s industry leaders to the structural awards for his teaching, advising, and
and aesthetic benefits of precast concrete. service.
“My focus has always been on teaching,” Sutton says, “It was a career I obviously had
though he admits that he didn’t set out to be a professor. something of an aptitude for,” he says.
Sutton received his BSCE degree in 1957 from the University of Maine Along with teaching, Sutton
and took a position at Charles A. McGuire, an engineering consulting firm became an active member of PCI
in Rhode Island. He says that when he took that job he was certain of only in the early 1970s. He was an active
one thing, “I was never going to work in structural design.” member of the Bridge Committee,
The firm’s owners had other plans. On his first day, instead of assigning and in 1989 he became the Bridge
him to highway design, his area of experience, they put him in the bridge Committee’s liaison to the Technical
group. “I’ve been a structural engineer ever since,” he says, laughing. Activities Council (TAC). He was
At McGuire, Sutton helped shape the U.S. interstate highway system, design- chair of TAC from 1999 through
ing bridges that are still in use today. However, in 1958, at the urging of mentor 2004, chaired the Research and
Gordon Archibald, he took a leave of absence to pursue his master’s degree at Development Committee from 2004
Purdue. to 2009, and became chair of the
While there, he took a position as a teaching assistant to help pay the Education Activity Council in 2010.
bills, filling in for a semester for a full-time instructor who left to pursue his He was recently elected to a second
PhD. At the end of that semester, Sutton was ready to get back to the world term, and will hold this role until
of engineering practice. In 1960, he received his master’s degree, returned to 2013.
McGuire, and married his wife Ellie. For all his hard work and commit-
“It was a very competitive environment in those days, and my graduate ment, Sutton was named one of the
education gave me many opportunities,” he says. 50 Titans of the Precast/Prestressed
Precast concrete was growing in popularity, and Sutton was asked to Concrete Industry in 2004, and
design and even proof test many of the precast concrete beams that were in 2009 he was given the Medal of
coming into fashion. Honor, PCI’s highest award.
After only a couple of years, academia called to him again. In 1961, Purdue “PCI has been a hugely positive
invited Sutton to return as a full-time faculty member while pursuing his own thing in my life,” Sutton says. “The
PhD. familylike structure of this organiza-
“It wasn’t an easy decision,” he says. He liked designing bridges, but he tion has helped move the industry
also recognized that the world was changing. Sputnik had gone up in 1957, forward.”
spurring renewed interest in engineering. It also helped him stay connected
Sutton says, “We realized we had to get moving as a nation.” to the world of precast concrete and
So Sutton took his wife and newborn baby back to Indiana and began a made him a better educator.
new career as an academic that lasted until May 2007. “Sitting on committees, I had a
During his time at Purdue, Sutton taught at least 10 different courses at view of the engineering world that
all levels in structural analysis, structural design, mechanics, and experimen- I couldn’t get in academia,” he says.
tal mechanics, including a course on prestressed concrete, which he taught “This connectivity to the real world
from the early 1980s until he retired. has been very important to my
career.” J
The “Jumbo Tee” has sixty inch stems with a highly stressed bulb
configuration at the stem bottoms. Hamilton Form built the self-stressing
formwork in sections for both shipping and stripping considerations.
The stems have minimal draft requiring the sideforms to be pull back for
stripping. The 118-foot long, twelve foot wide product is casting efficiently
and strips easily, keeping productivity high and delivery on time.
The next time you need formwork, from the simple to the complex,
call Hamilton Form to deliver: 817 590-2111 or sales@hamiltonform.com