28dbb993 4

MARCH–APRIL 2023
VOLUME 68, NUMBER 2
Precast/Prestressed Concrete Institute
Design
30 52 71 88
Gusset plate Repair technique Experimental Influence of
connection design investigation for background structural form
for buckling- deteriorated end behind partially on hydration-
restrained braced regions of bridge debonded strand heat-induced
frames girders AASHTO LRFD temperature rise in
specifications precast concrete
The Team!
A Great Connection for
Great Connections!
Your Connection Connection

7131 North Ridgeway Avenue • Lincolnwood, IL 60712 USA
847-675-1560 • 1-800-742-8127 • www.jvi-inc.com
Table of Contents
Design
Design and Cyclic Testing of a Gusset Plate Connection 30
for Precast Concrete Buckling-Restrained Braced Frames
Hannah D. Kessler, Kaitlynn M. Conway, Laura M. Redmond, and Garrett J. Pataky
Investigation of Repair Techniques for Deteriorated End Regions 52

52 of Prestressed Concrete Bridge Girders
William B. Rich, Christopher S. Williams, and Robert J. Frosch
Experimental Background behind New AASHTO LRFD Specifications 71

for Partially Debonded Strands
Mathew W. Bolduc, Bahram M. Shahrooz, Kent A. Harries, Richard A. Miller, Henry G. Russell,
and William A. Potter
71 Influence of Structural Form on Hydration-Heat-Induced Temperature Rise 88

of Precast Concrete Lining Segments for a Metro Transit Station
Yuzhen Han, Lei Zhang, and Jizhong He
New ACI 440.11 Code Adopted for Design of Concrete Reinforced 22

with Glass-Fiber-Reinforced Polymer Bars
John J. Myers, Douglas Gremel, Alvin Ericson, and Chad Van Kampen
Index of advertisers
CEG............................. Inside Front Cover PCI.................................................4, 6, 12, 15
cegengineers.com pci.org
Hamilton Form .......................................... 6 Prestress Supply Inc ...........Back Cover
hamiltonform.com prestresssupply.com
JVI ....................................................................1 Tucker’s...................... Inside Back Cover
jvi-inc.com tuckerbilt.com
PCI Journal | March–April 2023

2
March–april 2023 • VoluMe 68, NuMber 2
MARCH–APRIL 2023
VOLUME 68, NUMBER 2
Departments
On the cover
Precast/Prestressed Concrete Institute
Chairman’s Message 5
The new 11 Hoyt
condominium tower in Designed to Last
Brooklyn, N.Y., is designed
to provide space for Design
President’s Message 7
nature and community to The Next Class of PCI Titans
thrive, vertically, within the
densifying neighborhood of 30 52 71 88
From PCI Headquarters 8
Gusset plate Repair technique Experimental Influence of
connection design investigation for background structural form
the downtown area. Building

for buckling- deteriorated end behind partially on hydration-
restrained braced regions of bridge debonded strand heat-induced
frames girders AASHTO LRFD temperature rise in
specifications precast concrete
on the high thermal mass 01_Cover_MA23.indd 1 2/14/23 12:47 PM
of concrete, the structure’s PCI Calendar 11

60-to-40 window-to-wall
ratio pairs continuous interior insulation with high- Our Members 13
performance windows to create an energy efficient
thermal envelope. BPDL Beton Prefabrique, of Alma,
QC, Canada, was the precaster. Courtesy of Chris Coe In the News 16
for Optimist Consulting.
Industry Calendar 16
Project Spotlight 17
Research Corner 19
PCI Directories 101

Board of Directors and
Technical Activities Council 101
PCI Staff Directory 102
Regional Offices 103

JOURNAL EDITORIAL ADVISORY COMMITTEE Coming Ahead 103
Chair Richard Alan Miller Adel ElSafty Chungwook Sim
Vice Chair Pinar Okumus Amir Fam Sri Sritharan Meet Nancy Peterson 104
Secretary Collin Moriarty Alexander G. Mihaylov
Staff Liaison Tom Klemens Stephen J. Seguirant
EDITORIAL DESIGN & PRODUCTION

Tom Klemens Editor-in-Chief Lisa Scacco Publications Manager PCI Journal (ISSN 0887-9672) is published bimonthly by the Precast/Prestressed Concrete Institute, 8770 W. Bryn Mawr Ave.,
Chicago, IL 60631. Copyright © 2023, Precast/Prestressed Concrete Institute. The Precast/Prestressed Concrete Institute is
K. Michelle Burgess Managing Editor Walt Furie Senior Production Specialist
not responsible for statements made by authors of papers or claims made by advertisers in PCI Journal. Original manuscripts
Courtney McCormick Technical Editor and letters on published papers are accepted on review by the PCI Technical Review Committee. No payment is offered.
Angela Mueller Technical Editor ADVERTISING SALES Direct all correspondence to PCI Journal at journal@pci.org or Precast/Prestressed Concrete Institute, c/o PCI Journal, 8770
W. Bryn Mawr Ave., Suite 1150, Chicago, IL 60631. For information on advertising rates, send an email to adsales@pci.org.
Carrie Wyrick Technical Editor Trice Turner Business Development Manager
Subscription rates are $80 per year and $200 for three years in the United States, $170 per year and $470 for three years for
Rory Cleveland Copy Editor international, and $80 per year and $200 for three years for electronic-only subscriptions anywhere in the world. A single or
Elizabeth Nishiura Copy Editor back issue is $15. International subscriptions are delivered by an international carrier; allow one to three weeks.
Laura Vidale Copy Editor Postmaster: Please send address changes to PCI Journal, 8770 W. Bryn Mawr Ave., Suite 1150, Chicago, IL 60631.
Laura Bedolla Technical Activities Program Manager Periodicals postage rates paid at Chicago and additional mailing offices.
This paper is milled from a 3rd-party certified source
PCI Journal | March–April 2023

3
CALL FOR ENTRIES
Entries open on April 17, 2023. Join us in our
search for excellence and submit your projects
electronically by July 26, 2023.
The PCI Design Awards is not just looking for design excellence,
but also for projects with outstanding use of precast concrete.
PCI looks for projects that push the envelope and advance the
precast concrete industry.
The PCI Design Awards program showcases winning

projects in multiple ways:
■ PCI Convention Reception ■ Special project video
■ Full coverage in PCI publications ■ Dedicated project profile
■ Opportunity to appear on the on PCI website
front cover and/or as a project ■ Editorial coverage in external,
feature of Ascent local, and national magazines
Sponsored by:
VISIT PCI.ORG/DESIGNAWARDS FOR MORE INFORMATION

AND SUBMISSION DETAILS.
Chairman’s Message
Designed
to last
B readth of design is a tremendous advantage of precast, prestressed concrete over other con-
struction materials. Whether you need an ornate building with a parking structure that can
withstand an earthquake and hurricane at the same time or a normal office building, we can pro-
vide the engineering and an aesthetic design solution.
Design breadth, faster, better, safer, resilient. We do a great job of promoting all those quali-
ties. What are we missing here? Carbon friendly.
Buildings account for 39% to 40% of all global energy-related carbon emissions. About 27%
to 28% of this is from operations, and 11% to 13% is from construction. In the next 40 years, the
global building floor area is expected to increase by 2.4 trillion ft2. That is approximately double
today’s square footage. The energy consumption of construction and operations will be critical,
and the opportunities for our industry are astronomical.
At the same time, carbon is becoming a common topic of discussion and is being integrated
into future building codes. In addition to the previously stated advantages of precast, prestressed
concrete, the carbon consumption advantage of our products needs to be examined, understood,
and promoted. The embodied energy of precast, prestressed concrete is relatively low: 2.0 com-
pared with 2.5 for brick, 12.7 for glass, and 25 for steel.
With normal maintenance, precast concrete structures will last 100 years, which makes the
carbon consumption of these structures divided by their lifespan very desirable. If the need for
building replacement were to be reduced by half, we could eliminate a significant portion of car-
bon emissions worldwide.
Carbon consumption from heating and cooling buildings is a large portion of the operations
usage. The R-value of precast concrete with 3 or 4 in. encapsulated polyisocyanurate insulation
is 20 and 27, respectively. Using the thermal mass effect of the interior concrete wythe yields an
even higher effective R-value of the system. Compare the energy perspective of a glass curtain
wall having a best-case R-value of 5, and the choice is obvious. J
Matt Ballain
2023 PCI Board Chair
Vice President and General Manager
Coreslab Structures (INDIANAPOLIS) Inc.
Indianapolis, Ind.
PCI Journal | March–April 2023 5

SPECIFY PCI CERTIFICATION
THERE IS NO EQUIVALENT
The Precast/Prestressed Concrete Institute (PCI) certifications are the industry’s

most proven, comprehensive, trusted, and specified certification programs.
The PCI Plant Certification Program is accredited by the International Accreditation
Service (IAS), which provides objective evidence that an organization operates
at the highest level of ethical, legal, and technical standards. This accreditation
demonstrates compliance to ISO/IEC 17021-1. PCI Certification offers a complete
regimen covering personnel, plant, and field operations. This assures owners,
specifiers, and designers that precast concrete products are manufactured and
installed by companies who subscribe to nationally accepted standards and are
audited to ensure compliance.
To learn more about PCI Certification, please visit pci.org/certification
CERT21x40057_Certification_Halfpage_Ad_2022.indd 1 4/19/22 2:35 PM
It’s An Art Installation,

a 3-mile Bike/Walking Path and
an Illuminated Twin-Span Bridge
Crossing the Hudson River.
And, It Required Intricate Forms.
Project: Gov. Mario M. Cuomo Bridge

(formerly Tappan Zee Bridge)
Client: Unistress Corporation
Our Role: Hamilton Form created the

forms for the deck panels.
Hamilton Form Company For more than 55 years,

Custom forms Hamilton Form has been helping
Custom equipment
Practical solutions the precast community meet its
www.hamiltonform.com greatest challenges.
sales@hamiltonform.com
817-590-2111
It’s all we do.
President’s Message
The next class

of PCI Titans
“I f [we] have seen further, it is by standing on the shoulders of giants.”
To mark PCI’s 50th anniversary in 2004, the PCI Board of Directors created the Titans
of the Precast/Prestressed Concrete Institute award. The purpose of the award is to recognize
an individual who has demonstrated exceptional contributions to the industry, who has demon-
strated a dramatic impact on industry innovation, and who has shown unique leadership in
advancing and accelerating the growth of the precast concrete industry.
To honor the anniversary, the first class of Titans consisted of 50 individuals who were
cited as pioneers who made extraordinary contributions to the development of prestressed
concrete technology and PCI and its many programs. The group included industry innovators,
practitioners, educators, and leaders who built PCI. The award also called for a new group of
up to 10 new Titans to be chosen every 10 years thereafter, and on PCI’s 60th anniversary
in 2014, eight more Titans were named. A list of all Titans is available on the PCI website at
https://www.pci.org/PCI/About/Awards/PCI-Titans.
PCI will celebrate its 70th anniversary next year, so it is time to start the work of naming the
next group of Titans, which will be honored at the 2024 PCI Convention in Denver, Colo.
We need your help identifying the next group of industry leaders by nominating deserving
PCI and industry members. The criteria call for any individual with a long history of outstanding
service to the industry in any functional area, including education, research, design, production,
quality, erection, marketing, or management, or contributions related to society in general, such
as public service that brings positive attention to the industry. Nominations need to include
at least five letters of recommendation from past PCI Titans, PCI Medal of Honor recipients,
or PCI Fellows. Submissions opened February 14 and will close April 14. A Titan Task Group
made up of five Titans, five current or former PCI Executive Committee members, and me will
make recommendations to the PCI board at its summer meeting in June.
At the 2024 PCI Convention, we plan to have a special Titans ceremony honoring the new
Titans with videos highlighting their contributions.
Nominations can be made by any PCI member, employee of a PCI organizational member, or
employee of PCI or any affiliated chapter or partner organization. Please note per the policy that
PCI employees or employees of affiliated chapters or partners are not eligible to be PCI Titans.
You can find the entire Titan rules and a link to submit nominations on the PCI website under
Awards.
Please help us make this once-in-a-decade celebration special by nominating individuals who
have helped build PCI and the industry. These are the giants on whose shoulders we stand. J
Bob Risser, PE
PCI President and CEO

From PCI Headquarters
PCI cancels ICC-ES Dixon, Fink added to

evaluation reports PCI Mid-Atlantic leadership
I CC-ES evaluation report ESR-3010 for “precast concrete

diaphragms” is canceled effective October 2022. This evalu-
ation report was developed to facilitate the use of the DSDM
P CI Mid-Atlantic has made some additions to its leader-
ship. The regional chapter welcomes Sean Dixon, vice
president of construction services at High Concrete Group
(Diaphragm Seismic Design Methodology) under the 2012 in Denver, Pa., to its board as a director at large. In addition,
and 2015 International Building Code (IBC). Because PCI Evan Fink with Northeast Prestressed Products LLC in
supported adopting the DSDM into the American Society Cressona, Pa., has been named the producer member director,
of Civil Engineers and Structural Engineering Institute’s PCI Mid-Atlantic, on the PCI Board of Directors.
ASCE 7-16, Minimum Design Loads and Associated Criteria
for Buildings and Other Structures, which is referenced by the
2018 and 2021 IBCs, the evaluation report is no longer neces- PCI Foundation Board of
sary to use the DSDM.
ICC-ES evaluation report ESR-1997 for “design of fire-re- Trustees names new members
sistive construction for precast/prestressed concrete, using the
Precast/Prestressed Concrete Institute (PCI) third edition
manual MNL 124-11” is cancelled effective March 2023. As
the description suggests, this evaluation report was developed
P CI Foundation chair Gregory Force of Tindall Corp.
recently announced Ray Clark, executive director of
Georgia/Carolinas PCI, is the newly appointed vice chair and
to facilitate the use of the third edition of MNL 124, Design for Monty Oehrlein of Coreslab Structures (TEXAS) Inc., moved
Fire Resistance of Precast/Prestressed Concrete. Since PCI pub- to secretary.
lished PCI 124-18, Specification for Fire Resistance of Precast/ Joining the PCI Foundation Board of Trustees are Paul
Prestressed Concrete, which is referenced by the 2021 IBC, this Ramsburg of Sika Corp., Chris Kercsmar of CEG, and Matt
evaluation report has also been superseded by the new standard. Shea of the University of Colorado Denver. To learn more
As of these effective dates, any use or reference to these about the PCI Foundation whose role it is to expand precast
evaluation reports must be removed per ICC-ES policy. Any concrete education in colleges and universities, visit PCI-
questions may be directed to technical@pci.org. Foundation.org.
2023 T. Henry Clark Award call for nominations

Nominations for the T. Henry Clark Award, to or enhances the quality of precast concrete
be presented at 2023 PCI Committee Days, Octo- products or processes. T. Henry Clark believed in
ber 4–8, 2023, at the J. W. Marriott, Tampa, Fla., quality and quality processes, and this award is to
should be submitted to qualityprograms@pci.org recognize those who create or promote quality in
by June 1, 2023. The T. Henry Clark Award nomi- a way that would have made him proud.
nation form is available at https://www.pci.org For more information, contact Gary Wil-
/PCI/About/Awards/Clark. dung, the Quality Activities Council chair, at
The T. Henry Clark Award was established to gary.wildung@fdgcolorado.com or Mike Kessel-
recognize an individual, group of individuals, or mayer, PCI managing director of quality programs,
firm that has delivered a resource that improves at mkesselmayer@pci.org.
8 PCI Journal | March–April 2023

NSF awards precast concrete assistant professor of civil engineering at Clemson University
in Clemson, S.C., and Jose Restrepo, professor of structural
buckling-restrained braced engineering at UCSD.
The PCI Research and Development Council has commit-
frame research project ted to fund an additional $400,000 to support this project
over the three-year duration. PCI producer members that have
T he National Science Foundation

(NSF) has awarded a nearly $1.4
million grant for a precast concrete–
expressed commitments to support the project include Clark
Pacific, Concrete Technology Corp., Coreslab Structures
(INDIANAPOLIS) Inc., Metromont Corp., MidState
focused research project titled: “Analysis, Precast, NAPCO Precast, and Tindall Corp.
Design and System-Level Performance of Additional interested contributors are encouraged to con-
Repairable Precast Concrete Buckling- tact technical@pci.org. More information on the NSF award
Restrained Braced Frames under Seismic is available at https://www.nsf.gov/awardsearch/showAward
Loads.” This project seeks to develop a Yahya Kurama ?AWD_ID=2230187&HistoricalAwards=false.
repairable precast concrete buckling-re-
strained braced (BRB) frame structure
with a new type of nonproprietary precast 2023 Foundation studio
concrete diagonal brace. The research plan
includes numerical analyses and testing of grants to be announced April 1
isolated braces and their connections and
culminates in tests on the seismic perfor-
mance and repair of a scaled three-story
building on the University of California Laura Redmond
P reliminary proposals for PCI Foundation studio grants
are due by December 1 each year. On or before April
1, after a process of refining proposals and discussion with
San Diego (UCSD) outdoor shake table in members of the PCI Foundation Board of Trustees, recipi-
La Jolla. ents are announced.
Leading the research team are Yahya For more information and guidelines on how to apply for a
C. Kurama, the principal investigator and PCI Foundation grant, please visit https://www.pci
professor of civil and environmental engi- -foundation.org/proposal-guidelines-checklist-faq. Currently
neering and earth sciences at the University the foundation works with 43 universities in North America.
of Notre Dame in Notre Dame, Ind., with A map of studio locations is available at https://www.pci
coprincipal investigators Laura Redmond, -foundation.org/studio-info.
Jose Restrepo
Welcome the students

As we have grown over the tive to give plenty of time for professors to share
past 10 years or so, one of the best practices and real-world examples of quality
goals that has evolved at the PCI pedagogical outcomes.
Foundation is the idea of working The result has been amazing. Professors want
collaboratively with the profes- to come year after year because of the relation-
sors receiving our grants. Not ships they have formed, not only with industry
only having them work with the friends, but also with each other. Their willingness
industry but also with each other. to share their successes and challenges from day
Greg Force
Research shows that collaborative PCI Foundation one of the PCI grant curriculum grant program
problem-solving leads to bet- Chair has been a major factor in its success.
ter outcomes. This year we are opening up the Professors
People are more likely to take calculated risks Seminar and inviting our friends from the north.
that lead to innovation if they have the support The Canadian Precast/Prestressed Concrete Insti-
of a team behind them. This is why one of our tute is a cosponsor, as is PCI, which will be sharing
most successful programs since the beginning its latest teaching tools with the professors. More
of our educational programs has been the PCI collaboration means more problem solving and
Foundation Professors Seminar. While the sem- even better outcomes all around.
inar does allow time for industry members to If you have a professor you work with now,
share our knowledge and state-of-the-art re- even one that is already teaching precast con-
search with professors, we also have purposely crete, please sponsor them to come to the PCI
designed the three-day program to be collabora- Foundation Professors Seminar.

USC Carapace project
wraps up with ribbon cutting
P CI West and the PCI Foundation hosted a ribbon cutting

of the award-winning Carapace Pavilion on December 11,
2022, at Joshua Tree National Park in California.
Designed and constructed by University of Southern
California students and made of ultra-high-performance
concrete, the Carapace Pavilion construction was a mul-
tiyear, student-driven project made possible by the PCI
Foundation, Clark Pacific, JVI, Walter P. Moore, PCI West,
and PCI. In October 2021, it received a Citation Award for
Installations from the Los Angeles chapter of the American
Institute of Architects.
The Carapace Pavilion was featured in a Project Spotlight
in the January–February 2022 issue of PCI Journal.
The Carapace Pavilion, designed by University of Southern California

students, was celebrated with a grand opening on December 11,
2022, after it was installed at Joshua Tree National Park in Southern
California. Courtesy of Douglas Noble.
2023 Sidney Freedman Craftsmanship Award call for entries

PCI is accepting entries for the 2023 Sidney success in overcoming obstacles to production,
Freedman Craftsmanship Award. Launched in solutions to formwork or finishing challenges,
2012, the award recognizes PCI-certified plants and quality of individual units. Therefore, entries
for excellence in manufacturing and craftsman- should include source documents, shop drawings,
ship of architectural precast or glass-fiber-re- production photos as well as finished project
inforced concrete structures and individual photos to fully demonstrate the complex solutions
components. implemented for the project. For more informa-
Any kind, size, or type of structure and/or tion, visit http://www.pci.org/SFCA. The deadline
element may be entered. Judging is based on for all entries is July 1, 2023.
Sarah Fister Gale

Sarah Fister Gale, a long-time odicals and website.
writer for PCI Journal, died No- Gale was a freelance journalist and ghostwrit-
vember 3, 2022. She was 54. er based in Chicago, Ill., who covered a variety
Gale studied English literature of industries and topics, including precast con-
and journalism at the University crete, blockchain, artificial intelligence, work-
of Wisconsin–Madison, earning a force technology, human capital management,
bachelor of arts degree. project management, finance, and biopharma
In 2007, Gale started writing industry trends.
the Meet personality profiles that Sarah Fister Gale In addition to PCI Journal, Gale’s writing was
appear on the last page of PCI featured in Workforce Management Magazine,
Journal. Over the next decade and a half, she Talent Economy Magazine, PM Network, Chief
also wrote various Project Spotlight and feature Learning Officer, Salon.com, Chicago Parent,
articles for PCI Journal and covered the winners Jezebel, University of Chicago publications, and
of the annual PCI Design Awards for PCI’s peri- other media outlets.

2023 Irwin J. Speyer Young Professional Engineer Award
call for nominations
The Irwin J. Speyer Young Professional Engi- career. The award will be presented in October
neer Award honors the legacy of Irwin J. Speyer at the 2023 PCI Committee Days Conference in
by recognizing young professional engineers Tampa, Fla.
who have made significant contributions to Complete award details and the official nomi-
PCI during their early careers and who demon- nation form are available at http://www.pci.org
strate their intent to continue serving the pre- /PCI/About/Awards/Speyer-Award. Nominations
cast concrete industry as Speyer did during his must be submitted by May 1, 2023.
2023 Norman L. Scott Professional

Engineer Award call for nominations
The Norman L. Scott Professional Engineer The award will be presented in October at the
Award honors the legacy of Norman L. Scott by 2023 PCI Committee Days Conference in Tampa,
recognizing professional engineers who have Fla. Complete award details and the official nomi-
made significant contributions to PCI, the Amer- nation form are available at http://pci.org/PCI
ican Concrete Institute, the precast concrete /About/Awards/Norman_L_Scott. Nominations
industry, and the engineering profession at large. must be submitted by May 1, 2023.
PCI’s Calendar
Events
PCI event details are subject to change. For the most current information, visit
https://www.pci.org/events.
PCI Productivity Tour
May 1–3, 2023
Charlotte, N.C.
2023 PCI Northeast Meeting
May 9–10, 2023
Westbrook, Conn.
PCI West 2023 Summer Board Meeting
May 24, 2023
Woodland, Calif.
Georgia/Carolinas PCI Summer Meeting
June 14–16, 2023
Hilton Head Island, S.C.
2023 PCI Board of Directors Meeting
June 20–23, 2023
Indianapolis, Ind.
PCI of Illinois & Wisconsin Summer Meeting
July 11–12, 2023
Grand Geneva Resort, Lake Geneva, Wis.
Florida Prestressed Concrete Association Summer Meeting
July 27–30, 2023
Charlotte Harbor, Fla.
PCI Gulf South Summer Convention
July 27–30, 2023
Fairhope, Ala.
PCI Mid-Atlantic Summer Membership Meeting
August 3–4, 2023
Annapolis Waterfront Hotel, Annapolis, Md.

2022/23 Big Beam Contest Call For Entries
The PCI Student Education Committee is in- mance in the stated areas. Students must dis-
viting entries from students to participate in the cuss both the structural design and the concrete
Engineering Student Design (Big Beam) Competi- mixture proportions for the beam. For more
tion for the 2022/23 academic year. information, visit https://www.pci.org
Each student team must work with a PCI pro- /BigBeam. All intending to submit a report must
ducer member to build a precast, prestressed submit an application online at http://www.pci
concrete beam that is 20 ft long. The beams will .org/bigbeamapp. Final reports are due to PCI
be tested and prizes awarded for best perfor- by June 12, 2023.
PCI personnel training and certification schools

Quality Control School event details are subject to change. If you have any questions about the Quality Control School sched-
ule or need help completing a registration form, please contact PCI’s education department at education@pci.org. Registration
forms are available at https://www.pci.org/qc_schools.
May 10–12, 2023 Chicago, Ill.
June 26–29, 2003 online
Levels I and II September 18–21, 2023 online
October 25–27, 2023 Nashville, Tenn.
November 13–16, 2023 online
May 9–12, 2023 Chicago, Ill.
August 14–17, 2023 online
Level III
October 24–27, 2023 Nashville, Tenn.
December 11–14, 2023 online
April 10–13, 2023 online
Certified Field Auditor
September 11–14, 2023 online
April 14, 2023 online
Certified Company Auditor
September 15, 2023 online
Compiled by K. Michelle Burgess (mburgess@pci.org) J
Guidelines for the Use of Ultra-High-Performance

Concrete (UHPC) in Precast and Prestressed
Concrete (TR-9-22)
This new publication provides a practical guide for the development and
TR-9- 22
Guid eline s
qualification of UHPC mixtures based on locally available materials. It
presents an overview of UHPC production specific to long-span precast,
Gu id el in es fo r pretensioned UHPC structural elements for buildings and bridges.
8770 W. Bryn Mawr Ave.
| Suite 1150 | Chicago,
Guide lines for
IL 60631-3517 | 312-786-03
00 | www.pci.org
th e Us e of
Ul tr a- Hi gh -P er
fo rm an ce
Co nc re te (U HP Topics discussed include:
the Use of Ultra-
C) in Pr ec as t
an d Pr es tr es se
d Co nc re te • constituent materials and development of mixture proportions
for manc e Concr H igh-P er
• batching and placement considerations for production

• methods for evaluating UHPC materials for mixture qualification
ete (UHPC ) in Preca
and routine quality assurance.

st and Prestr essed
PCI Con cret e M

Now available in the PCI Bookstore (free PDF download for PCI members).
ater ials
Tech nolo gy Com
mitt ee
Concr ete
TR-9 -22

7/25/22 1:15 PM
Our Members
Kubat retires from Wells Before working at ACI, Frank gained a comprehensive
working knowledge of resilience and sustainability as an
S pencer Kubat has retired from his

position as vice president of sales for
Wells Midwest. Kubat started his career
employee at Wiss, Janney, Elstner Associates Inc., the National
Precast Concrete Association, and through his consulting
company. He also has experience in working as an assessor
with Wells in 1980. with ISO standards governing the operations of conformity
Starting in the 1980s, Wells entered assessment bodies. Frank holds a master’s degree in civil engi-
the architectural building solutions mar- neering from the University of Colorado at Boulder and is a
ket. In recent years, Wells has seen signif- licensed professional engineer in Indiana and Colorado.
icant growth and expansion in geographic Spencer Kubat “NEU has been growing quickly and I’m excited about
regions, employees, and revenue with acqui- our opportunities to make a difference in reducing the car-
sitions of Rocky Mountain Prestress in 2019 and Spancrete bon emissions of concrete,” Frank says. “I have enjoyed the
in 2020. Following the acquisitions, Kubat’s role transitioned work I’ve been doing with NEU so far in setting up the val-
from day-to-day client support to strategic endeavors. idation program, which we believe will have a big impact on
Greg Roth, president and chief operating officer of the the industry.”
Midwest Division for Wells, says, “Spencer was instrumental —Source: NEU: An ACI Center for Carbon Neutral
in pushing Wells into the high-end architectural building solu- Concrete
tions market that we are well known for today. He put quality
and service at the core of everything he did internally with
teammates and externally among numerous relationships with Wells promotes McGlothlen
some of our largest repeat clients that we still have the pleasure
of calling partners today.” to senior vice president of HR
—Source: Wells
ACI Center of Excellence

W ells has promoted Amy
McGlothlen to senior vice presi-
dent of Human Resources.
McGlothlen has been part of the
for Carbon Neutral Concrete Wells team for more than nine years,
starting as a human resources generalist
names Frank executive director and advancing to human resources man-
ager for the Great Lakes Division in 2020. Amy McGlothlen
N EU: An ACI Center of Excellence

for Carbon Neutral Concrete has
appointed of Dean A. Frank to serve as its
The new position gives McGlothlen more
opportunities to have an impact on the full
employee experience, from onboarding and training through
executive director. long-term engagement and development.
Frank previously worked as director of With 1200 employees across the nation, employee health
validation for NEU and as program devel- and well-being and diversity in the workforce is paramount to
oper for the American Concrete Institute Wells’ success.
(ACI). He also served as director of qual- Dean A. Frank McGlothlen has spent more than 20 years focused on the
ity programs for PCI from 2007 to 2011 human resources field. She holds a bachelor’s degree in busi-
and then as the director of quality and sustainability programs ness administration and a Professional in Human Resources
from 2011 to 2017. He brings extensive experience in sustain- certification.
ability, International Standards Organization (ISO) standards —Source: Wells
and certification development of personnel, products, and
manufacturing plants.

Welcome to PCI!
Precast Installers Services Associate
Connecticut Mason Services Catalyzer Inc.
Contractors Inc. 804 Hithergreen Court
75 Bysiewicz Drive Lansing, KS 66043
Middletown, CT 06457 (913) 547-0665
ConnecticutMason.com Primary contact: Stephen Ingalls
(860) 296-9984 ingalls@teamcatalyzer.com
Primary contact: Connie Buscema
cbuscema@connecticutmason.com
Supplier Associate
Stonebridge Inc. Esch Construction Supply Inc.
3230 Hamilton Blvd. 561 Phalen Blvd.
South Plainfield, NJ 07080 St. Paul, MN 55130
StonebridgeSteelErection.com EschSupply.com
(908) 753-1100 (651) 487-1880
Primary contact: Mark Hoffa Primary contact: Dan Esch
mhoffa@stonebridgese.com dane@eschsupply.com
Charles E. “Budd” Hilgeman Jr.

Charles E. “Budd” Hilge- in 2004 for his contributions to precast concrete
man Jr., one of the founders of innovation, leadership of industry-altering devel-
Concrete Technology Inc., died opment, and accelerating the growth of precast
December 5, 2022. He was 88. concrete in North America. In 2006, he became a
Hilgeman graduated from the PCI Life Member.
University of Dayton in Dayton, Active in PCI administrative and committee
Ohio, with a bachelor of arts work, Hilgeman served two terms as director of
degree in 1956. In his senior year, Charles E. “Budd” zone 4 in 1977/78 and 1995/96, and he was sec-
Hilgeman was promoted to cadet Hilgeman Jr. retary-treasurer of the PCI Board of Directors in
colonel and student commander of the University 1997. He chaired the Convention Advisory, Educa-
of Dayton Army Reserve Officers’ Training Corps tion, and Student Education Committees and was
unit. Hilgeman served for three years of active a member of the Architectural Precast Concrete,
duty as an officer in the 18th Infantry Regiment, Architectural Precast Concrete Division Manage-
1st Infantry Division of the United States Army. ment, Fellows Nominating, and Program Planning
In 1969, Hilgeman became one of the founders Committees and the Industry Marketing Council
of Concrete Technology Inc. (CTI) in Springboro, and Research and Development Council.
Ohio, where he served as chairman of the board Hilgeman was dedicated and committed to
and chief executive officer. CTI was a leading furthering student education at all levels. He was
manufacturer of architectural precast concrete heavily involved in promoting precast concrete to
cladding panels and beams for commercial, students, professional educators in both architec-
government, manufacturing and retail buildings. tural and engineering disciplines, and practicing
Hilgeman and the other three founders were professional architects, engineers, and owners.
proud that they converted CTI to an employee He also encouraged other PCI members to get
stock ownership plan so that the workers would involved in supporting education for students, ed-
become owners in the company they were work- ucators, and practitioners. Hilgeman was a guest
ing for. He retired from CTI in 2005. lecturer at the Construction Materials and Tech-
Hilgeman became a PCI member in 1977 with nology Institute Workshops, a program sponsored
CTI and was named a PCI Fellow in 1997 and a Ti- by PCI for the Association of Collegiate Schools of
tan of the Precast/Prestressed Concrete Institute Architecture.

PCI’s newly certified erector
PCI recently certified the following erector. For an explanation of the certification designations, visit
http://www.pci.org/Erector_Certification.
• Rocky Mountain Precast in Yigo, Guam: S2
Sylaj named CPCI president Wells promotes Kloos

to Mountain States
T he Canadian Precast/Prestressed
Concrete Institute (CPCI) has
named Val Sylaj its new president effec- president, COO
tive January 1, 2023, succeeding Robert
Burak.
“I have been fortunate to be a part of
a vibrant cement and concrete industry
W ells has promoted Steve Kloos
to president and chief operating
officer of its Mountain States region. He
over the last 35 years, and personally Val Sylaj will be responsible for preconstruction,
honored to be the president of CPCI over sales, engineering, operations, and field
the last 14 years. I am excited to see the next chapter unfolding and installation services throughout the
for CPCI and I am equally excited to ‘pass the torch’ to Val,” mountain states.
says Robert Burak, former CPCI president. “Val leads a strong A 27-year veteran of Wells, Kloos Steve Kloos
team that is certain to evolve over time to respond to the started his career in the former Ready Mix
ever-changing needs of the industry.” Division, ascended to general manager, and most recently served
Sylaj is a member of PCI’s Journal Awards and as senior vice president of quality. In this role, Kloos became
Safeguarding Impartiality Committees. heavily integrated in all of Wells’ manufacturing operations,
—Source: Canadian Precast/Prestressed Concrete Institute leading efforts related to quality control, with an emphasis on
continuous improvement and the client experience.
—Source: Wells
PCI Needs You

PCI relies on a network of robust and engaged committees to develop standards, handbooks,
manuals, guidelines, and other documents for our Body of Knowledge. These committees
regularly need additional members to support ongoing efforts. Current PCI committees
soliciting new members include:
• Structural Design of UHPC (newly formed) • Prestressed Concrete Poles
• FRP Composites • Professional Members
• Prestressed Concrete Piling
To learn more or to join a committee,

scan the code or visit: pci.org/PCI/
Applications/Committee.aspx
15
In The News
Proposed ASTM standard limited in the construction industry. The proposed standard,
WK60666, will give the engineering community the data and
to aid in measuring strength confidence needed to evaluate performance of fibers.
This effort directly relates to the United Nations
of fiber-reinforced concrete Sustainable Development Goal 9 on industry, innovation, and
infrastructure.
A STM International’s Committee C09 on Concrete and

Concrete Aggregates is developing a proposed stan-
dard that will be used to help measure the tensile strength of
“Typically, tensile capacity of concrete is ignored when
designed with reinforced concrete,” Pinkerton says. “Being
able to measure tension accurately would give engineers the
fiber-reinforced concrete. confidence to use the tensile strength rather than ignore it.
ASTM International member Luke Pinkerton says that Concrete with fiber reinforcement is stronger, less brittle,
the use of fiber reinforcement as an alternative to conven- more durable, and resilient.”
tional reinforcement of various concrete structures has been —Source: ASTM International
Industry Calendar
Event details are subject to change.
ACI Concrete Convention
April 2–6, 2023
Hilton San Francisco Union Square, San Francisco, Calif.
PTI Convention
April 30–May 3, 2023
Miami, Fla.
AASHTO Spring Meeting
May 15–18, 2023
Hyatt Regency Seattle, Seattle, Wash.
SEI Structures Congress
May 26, 2023
New Orleans, La.
fib Symposium 2023
June 5–7, 2023
Istanbul, Turkey
SynerCrete ’23
June 15–16, 2023
Milos Island, Greece
fib International Symposium on Conceptual Design of Concrete
Structures June 29–July 1, 2023
Oslo, Norway
BEI-2023 “Sustainability in Bridge Engineering”
July 17–20, 2023
Roma Eventi-Fontana di Fontana Trevi, Rome, Italy
CACRCS 2023
September 12–15, 2023
Parma, Italy
Greenbuild
September 26–29, 2023
Walter E. Washington Convention Center, Washington, D.C.
PTI 2023 Committee Days
October 3–6, 2023
Cancun, Mexico
ACI Concrete Convention October 29–
Boston Convention Center and Westin Boston Waterfront, Boston, Mass. November 2, 2023

Project Spotlight
11 Hoyt makes complex facade

panel design look easy
T he new condominium tower 11 Hoyt in Brooklyn, N.Y.,

is designed to provide space for nature and community
to thrive, vertically, within the densifying neighborhood of the
downtown area.
The building is approximately 770,000 ft2 (71,500 m2),
including 55,000 ft2 (5100 m2) of interior and exterior ameni-
ties, 476 residential units ranging from studios to three bed-
rooms, and a total of 52 stories standing at 620 ft (189 m) tall.
Precast concrete was chosen for 11 Hoyt, and BPDL
Beton Prefabrique of Alma, QC, Canada, was selected as the
precaster.
The system of precast concrete double-donuts for the façade
was selected for its reliability and watertightness, as well as its
efficient, on-site construction, saving time and money while
reducing risk.
The cast-in-place superstructure had to maintain a simple,
flat slab edge to meet cost and construction schedule demands.
Using a strong yet fluid and easily malleable material, the team
was able to fabricate the facade panels in a shop-controlled
environment. The result is a precast concrete facade that is
three-dimensional, as deep as 3 ft 9 in. (1.14 m); inhabitable;
and able to support a live load.
Building on the high thermal mass of concrete, the build-
ing’s 60-to-40 window-to-wall ratio pairs continuous interior
insulation with high-performance windows to create an ener-
gy-efficient thermal envelope that greatly reduces energy loads
imposed on mechanical systems.
The project’s connection between inside and outside is fur-
thered by the tower’s thickened precast concrete facade. The
building pushes out in plan to create expanded living spaces
with built-in window seats framing 8 ft (2.4 m) tall windows,
maximizing views of the neighborhood and waterfront and
offering glimpses of the building’s sculptured exterior.
Migrating across the facade like cusps of a wave in sections, The system of precast concrete double-donuts used in the 11 Hoyt
the scalloped bay windows allow for more than 190 unique condominium tower facade was selected for its reliability and water-
floor plans that accommodate a diverse mix of residents. tightness. The new Brooklyn, N.Y., structure contains 476 residential
The project progressed smoothly for the most part, largely units. Copyright Tom Harris, courtesy of Studio Gang.
because of the cooperation of the various entities involved.
“There was a weekly design assist process that started at the before the production of the first panels and mold, the design
early stage of the project that allowed the design team and team visited BPDL’s precast concrete plant in Canada to
BPDL to go over every design challenge and discuss solutions,” visualize the production process and understand any possible
says Matthieu Gagne, project manager for BPDL. “From limitations.
the start, the cooperative atmosphere and mindset was what To reduce production challenges, BPDL produced multiple
allowed the design and construction to be so successful.” Even internal samples and mock-ups ahead of the production and

drawings process, which allowed BPDL to foresee challeng-
es that this complicated geometry brought. “We needed to
understand our limitations during the design assist and draft-
ing process,” Gagne says.
Transportation and delivery were also tackled using exten-
sive preplanning. “The BPDL transportation team modelized
the loading of our trucks and used short trial runs ahead of
time,” Gagne says. “The atypical geometry required custom
parts to be drafted and built for the panels to be safely secured
on the trucks.”
Installation challenges were also circumvented by preplan-
ning. “The erector, Midwest Steel, was brought in early during
the design process to provide feedback and solutions,” he
says. “The coordination they did with the general contractor, Ransom Everglades School’s new STEM Center was placed to form a
Triton, was instrumental in the project, especially for develop- quad at the entrance to the school in Coconut Grove, Fla. Courtesy of
ing the hoist panels rigging procedure.” Gate Precast Co.
—William Atkinson
awareness of the activities inside and outside the classrooms,
and provides an opportunity to showcase learning and the
Florida STEM center ever-present cinematic activity within the building.
The roof level has an outdoor laboratory designed for modular
built for the future agriculture. The majority of the roof surface also hosts photovol-
taic panels, contributing to the pursuit of the U.S. Green Building
The new STEM Center—which focuses on science, tech- Council’s LEED silver or gold level of certification.
nology, engineering, and math—at Ransom Everglades School Although the project went smoothly for the most part,
in Coconut Grove, Fla., creates a new quad at the entrance there were a few challenges, says Bruce Bartscher, vice pres-
to the school’s historic campus. The thought behind the new ident of operations for Gate Precast. “To handle the design
building’s placement is that it showcases the school’s commit- intent of undulating features on the face, we thickened our
ment to academic excellence, the arts, and next-generation typical 2¼ in. face to 3¼ in. and added high-density foam that
environments for teaching and learning. was run through our CNC [computer numerical control] to
The school envisioned a 45,000 ft2 (4180 m2) center for sci- create sharp contrast,” he says.
ence and math that allows for a multidisciplinary approach to Fortunately, there were no challenges related to production.
education for both faculty and students. Precast concrete was In terms of transportation, though, size had to be considered.
chosen for the project because the owner and designer wanted “Because the panels were 10 ft by 30 ft, we shipped them on
to create a responsibly designed building that is as intelligent as edge, utilizing the frames for support,” he says.
the students, activities, and faculty it houses. —William Atkinson J
Gate Precast in Kissimmee, Fla., was selected for the proj- On a lightweight architectural precast concrete panel system, a steel
ect. Gate and the designer created a lightweight building enve- frame provides the structure that supports the weight of the nonstruc-
lope solution that not only reflects the core mission statement tural concrete skin for the Ransom Everglades School’s STEM Center.
of the school but also provides forward-thinking construction Courtesy of Gate Precast Co.
prefabrication methods.
The STEM building is in one of the most vulnerable areas
of the city’s coastal neighborhoods. As a result, it had to be a
high-performing “smart” building at all levels: its structure, its
systems, and even the way the architect planned it.
The precast concrete facade is an integral part of this
STEM building, which is already gaining attention as one of
the most innovative STEM buildings in the region and possi-
bly the nation.
Flexibility and adaptability, from moving walls to movable
furniture and services, make this facility one of the leading
STEM high school centers in the country and showcase a for-
ward-thinking faculty and award-winning student body.
Transparency not only allows for passive security best
practices but also enhances a daylit environment, creating

Research Corner
Evolution of the
L-shaped spandrel beam
Andrew Osborn and Gary Klein
N umerous PCI-sponsored research projects over the years

have endeavored to improve the design of precast con-
crete components. One such component is the beam with an
integral ledge.
Beams with ledges may be inverted-T shaped or L shaped.
This Research Corner will describe research funded by PCI
pertaining to the L-shaped beam, also called an L beam. The
PCI-funded research had a profound impact on the evolution
of design procedures for such beams.1–3
L beams are typically placed at the perimeter of a precast
concrete structure and support double-tee-beam floor com-
ponents. As used in parking structures, the web of the L often
extends above the floor surface to function also as a barrier
wall. Such beams are often referred to as deep or slender span-
drels. The beams may be prestressed or mild reinforced. Both
styles are common.
The double-tee beam stems bear on the ledge, creating a
series of eccentric forces.These impose a torsional load on the
L beam that is usually restrained by connections to the col-
umns at both ends. The top flanges of the double-tee beams,
typically at each double-tee stem, are fitted with an embed Configuration of typical L-beam. Source: Reproduced by
plate to provide a welded connection to the L-beam web. At permission from Lucier (2010), Fig. 2-4, p. 7.
the ledge surface, double-tee beams are supported on bearing
pads to spread concentrated load effects and reduce undue design method. Notably, for hanger reinforcement, a footnote in
restraint to volume change forces. section 6.14 of the third edition of the PCI Design Handbook6
The first edition of the PCI Design Handbook: Precast and states, “A consensus on the design procedure for hanger rein-
Prestressed Concrete4 mentions L-shaped beams and provides forcement, Ash, has not been reached as of publication, and
a design chart for standard sections, but no design procedure future recommendations may differ from that shown here.”
was given. In the second edition of the PCI Design Handbook5 Several references that also discuss the issue are noted.8,9
a ledge design is provided but a complete design for the entire PCI realized that the L-beam design procedures relied on
L beam is not presented. The third edition of the PCI Design certain assumptions of behavior rather than physical test data.
Handbook6 presented a complete L-beam design procedure and PCI also sought to reduce the conservative design assumptions
references several PCI Journal papers. In this procedure, assump- of the compression field theory approach. Accordingly, PCI
tions are made regarding the design of the hanger steel and han- underwrote specially funded research project 5. The research1
dling of the torsional reaction at the L-beam ends. End torsion involved physical testing, finite element analysis, and a lit-
reinforcement and hanger reinforcement designs are based on erature review. The focus of the report was a more rational
the recently developed compression field theory from Collins approach for the design of hanger steel and end-of-span rein-
and Mitchell.7 That design procedure resulted in increased end- forcement. One of the findings of that research was that the
of-beam reinforcement compared with the previous torsion PCI design method for punching shear may be unconservative.

Subsequently, PCI funded two relevant Jenny Fellowship Additional research was conducted at NCSU and WJE3 to
projects. Yazdani10 performed finite element studies of address the potentially unconservative punching shear design
L beams, pocketed spandrels, and discrete corbel spandrels. procedure presented in PCI Design Handbook through the
The research reviewed the ACI 318-9511 design proce- seventh edition.13 It was found that design for punching shear
dures, which, similar to the third edition of the PCI Design must consider global flexural and shear stress in the portion of
Handbook,6 were based on the compression field theory. the L beam below the ledge.
Similar to Klein,1 Yazdani found that the design procedure The PCI design procedures for L beams were substantially
resulted in highly congested steel reinforcement near the changed in the eighth edition of the PCI Design Handbook14
L-beam ends. Lini12 performed similar research focused on as a result of the Lucier2 and Rizkalla3 research and are being
thin spandrels and had similar findings. further modified for the ninth edition of the PCI Design
PCI funded further research at North Carolina State Handbook, which is under development. The alternative tor-
University (NCSU) and Wiss, Janney, Elstner Associates Inc. sion design procedure for slender beams, those with a web
(WJE).2 Alternative torsion design procedures were devel- height–to–depth greater than 4.5, has also been recognized by
oped for slender spandrels based on separate consideration of ACI 318-1915 section 9.5.4.7. Further code change proposals
the plate bending and twist components of torsion across an are pending.
inclined failure plane. As a result, closed transverse reinforce-
ment was no longer required, and reinforcement congestion at References
L-beam ends was greatly reduced. 1. Klein, G. J. 1986. Design of Spandrel Beams. PCI specially
funded research and development program research proj-
ect 5. Chicago, IL: PCI.
2. Lucier, G., C. Walter, S. Rizkalla, P. Zia, and G. Klein.
2010. Development of a Rational Design Methodology
for Precast Concrete Slender Spandrel Beams. Technical
report IS-09-10. Raleigh, NC: North Carolina State
University.
3. Rizkalla, S., M. Nafadi, G. Lucier, P. Zia, and G. Klein.
2016. Behavior and Design of Directly-Loaded L-Shaped
Beam Ledges. Technical report RD-16-03. Raleigh, NC:
North Carolina State University.
4. PCI Industry Handbook Committee. 1972. PCI Design
Handbook: Precast and Prestressed Concrete. 1st ed.
Chicago, IL: PCI.
Handbook: Precast and Prestressed Concrete. 2nd ed.
Chicago, IL: PCI.
Handbook: Precast and Prestressed Concrete. 3rd ed.
Failure mode near beam end. Source: Reproduced by per- Chicago, IL: PCI.
mission from Lucier (2010), Fig. 3-47, p. 63.
Punching shear failure mode. Source: Reproduced by permission from Rizkalla (2016), Fig. 4-15b, p. 96.

7. Collins, Michael P., and Denis Mitchell, 1980 “Shear
and Torsion Design of Prestressed and Non-prestressed
Concrete Beams.” PCI Journal 25 (5): 85–86.
8. Zia, P., and W. D. McGee. 1974. “Torsion Design of Pre-
stressed Concrete.” PCI Journal 19 (2): 46–65. About the authors
9. Zia, Paul, and Thomas Hsu. 1978. “Design for Torsion
and Shear in Prestressed Concrete.” Preprint 3424, ASCE Andrew Osborn is senior
Chicago Exposition, October 1978. principal at Wiss, Janney,
10. Yazdani, Nur, and Jennifer Ach. 2000. Behavior of Thin Elstner Associates Inc. in
Precast Spandrels in Torsion. Tallahassee, FL: Florida Boston, Mass., and chair of the
A&M University and Florida State University. PCI Research and Develop-
11. ACI (American Concrete Institute) Committee 318. ment Council.
1985. Building Code Requirements for Structural Concrete
(ACI 318-85) and Commentary (ACI 318R-85). Gary Klein is executive vice
Farmington Hills, MI: ACI. president and senior principal at
12. Lini, Carlo R., and Julio A. Ramirez. 2004. On the Design Wiss, Janney, Elstner Associates
for Torsion of Precast/Prestressed Concrete Spandrel Girder. Inc. in Northbrook, Ill.
West Lafayette, IN: Purdue University.
Handbook: Precast and Prestressed Concrete. 7th ed.
Chicago, IL: PCI. Keywords
14. PCI Industry Handbook Committee. 2017. PCI Design Beam, deep spandrel, design, L beam, ledge, slender
Handbook: Precast and Prestressed Concrete. 8th ed. spandrel, spandrel.
Chicago, IL: PCI.
15. ACI Committee 318. 2019. Building Code Requirements Publishing details
for Structural Concrete (ACI 318-19) and Commentary This paper appears in PCI Journal (ISSN 0887-9672)
(ACI 318R-19). Farmington Hills, MI: ACI. V. 68, No. 2, March–April 2023, and can be found at
https://doi.org/10.15554/pcij68.2-06. PCI Journal
is published bimonthly by the Precast/Prestressed
Concrete Institute, 8770 W. Bryn Mawr Ave.,
Suite 1150, Chicago, IL 60631. Copyright © 2023,
Precast/Prestressed Concrete Institute.
Have a research idea?

We urge readers to send in their research ideas to
Jared Brewe, PCI’s vice president of technical ser-
vices, at jbrewe@pci.org. J

John J. Myers, Douglas Gremel, Alvin Ericson, and Chad Van Kampen
New ACI 440.11 code adopted

for design of concrete reinforced with
glass-fiber-reinforced polymer bars
T
he American Concrete Institute (ACI), through the
work of ACI Committee 440, Fiber-Reinforced Poly-
mer Reinforcement, has published ACI 440.11-22,
Building Code Requirements for Structural Concrete Rein-
forced with Glass Fiber-Reinforced Polymer (GFRP) Bars—
Code and Commentary.1 This new code was developed by an
American National Standards Institute–approved consensus
process and addresses structural systems, members, and con-
nections, including cast-in-place, precast, nonprestressed,
and composite concrete construction.
This is the first comprehensive building code covering the

use of nonmetallic, GFRP reinforcing bars in structural con-
crete applications. GFRP reinforcement has been in use for
decades as an alternative to steel reinforcement because of
its noncorrosive, nonmagnetic, and lightweight properties.
Scope and organization of ACI 440.11
The new ACI 440.11-22 code includes 27 chapters with

provisions for designing GFRP-reinforced concrete beams,
one-way and two-way slabs, columns, walls, connections,
and foundations. Other model codes and standards can
directly reference ACI 440.11-22 to allow for widespread,
■ A new code has been published on the design of responsible use of this important technology.
concrete reinforced with glass-fiber-reinforced-poly-
mer bars. ACI 440.11-22 mirrors ACI’s Building Code Requirements
for Structural Concrete (ACI 318-19) and Commentary
■ This article provides background on this new code (ACI 318R-19)2 with the same layout and chapters but also
and discusses potential uses for precast concrete uses existing provisions where possible. An equals sign
components and structures. is used to indicate where provisions in ACI 440.11-22 are

identical to provisions in ACI 318-19 (see section 7.7.1.1 in for FRP design an over-reinforced design permits a higher
Fig. 1). Where a section in ACI 318-19 does not apply, sec- strength reduction factor.
tions are noted as “Intentionally left blank” in ACI 440.11-22
(see section 7.7.1.4 in Fig. 1). The consistency was intentional Although some FRP products, such as carbon-fiber-reinforced
to help design professionals and practitioners familiar with polymer, have similar stiffness to steel, other FRP prod-
ACI 318-19 to become familiar with and use ACI 440.11-22 ucts, such as GFRP, have much lower stiffness. This means
more efficiently. that many designs may be governed by serviceability limits
rather than strength. On an equal reinforcement area basis,
There are a few topics in ACI 318-19 that are not addressed FRP design will often produce larger deflections and wider
or not applicable to ACI 440.11-22. Not addressed in ACI cracks. Larger crack widths may be accepted since FRP is
440.11-22 are chapter 12, “Diaphragms;” chapter 14, “Plain noncorrosive; however, there is a perception of failure with
Concrete;” chapter 17, “Anchoring to Concrete;” chapter 18, larger visible crack width. Commentary section R24.3.2 in
“Earthquake-Resistant Structures;” and chapter 23, “Strut- ACI 440.11-22 discusses the crack control provision differ-
and-Tie Models.” Diaphragms are expected to be included ences between ACI 440.11-22 and ACI 318-19. Notably, the
in the next edition. ACI 440.11-22 also does not cover maximum bar spacing limits in ACI 318-19 correspond to a
lightweight concrete, prestressed concrete, deep beams, and maximum crack width of approximately 0.018 in. (0.457 mm)
shotcrete. ACI 440.11-22 does not permit GFRP- whereas the maximum bar spacing limits in ACI 440.11-22
reinforced concrete members to be designed as part of a are based on a crack width of 0.028 in. (0.711 mm). This larg-
seismic-force-resisting system in seismic design categories er crack spacing is to prevent deterioration due to freezing and
B and C nor does it permit any GFRP reinforced concrete thawing rather than reinforcement corrosion for steel.
member in structures assigned to seismic design categories
D, E, and F. For designing the bond and development length for GFRP
bars, ACI 440.11-22 specifies the length based on the required
Research is ongoing with respect to anchoring fiber-rein- stress in the bar to develop the full nominal section capacity
forced-polymer (FRP) bars into existing concrete using and not ffu. This is different from ACI 318-19, which specifies
epoxy or adhesive systems along with expanding the current lengths to develop fy of the steel reinforcement. There are dif-
anchoring database with short anchoring lengths using FRP
bars. Although the linear elastic nature of FRP products
proves to be less ductile than steel reinforcement, more
work is needed to specifically develop and detail more duc-
tile connections should it be used for seismic application.
To date, some work has been done to examine hybrid bars
that provide a combination of higher strength and stiff-
ness with improved strain capacity at failure. In addition,
although FRP has been used in prestressed applications, it
is less mature than traditional prestressed steel applications
and design processes; however, advances are being made in
the area of anchor and grip FRP bars, which had been one
of the major challenges for use of high-strength/high-
stiffness FRP products.
Design differences from ACI 318
Some of the key design differences for FRP bars include

using guaranteed bar properties provided by the manufactur-
er. These properties often vary not only by bar manufacturer
but also by bar size. The properties do not align similarly to
specific steel grades that are familiar to many designers for
reinforced concrete or prestressed concrete. In addition, the
FRP design approach uses an environmental reduction factor
that considers the long-term service life of the material. In the
case of steel reinforcement, no considerations are included
for long-term environmental degradation due to corrosion
and loss of cross section, for example. ACI 440.11-22 allows
for both over-reinforced designs (concrete crushing) as well
as under-reinforced designs (bar rupture). The failure mode Figure 1. Example use of existing ACI 318-19 provision in ACI
is affiliated with a specific strength reduction factor similar 440.11-22. Note: An equals sign is used before provisions in ACI
to steel reinforcement design; however, unlike ACI 318-19, 440.11-22 that are identical to provisions in ACI 318-19.

ferences in shear design as well since GFRP has lower dowel • where machinery will consume the reinforced member
resistance, lower modulus of elasticity, lower tensile strength (such as in mining and tunneling)
around a bend compared with the straight part of a bar, higher
tensile strength, and no yield point. The larger crack widths • applications that require thermal nonconductivity
relate to less aggregate interlock, and the smaller compression
zone depth results in less concrete resistance in the compres- Material requirements
sion zone. Therefore, the contribution from the concrete Vc in ASTM D7957-22 for GFRP bars
varies for FRP compared with steel. ACI 440.11-22 ignores
the contribution of GFRP bars in compression. For beams in ACI 440.11-22 makes substantial references to ASTM D7957,
flexure, the designer replaces the FRP area with an equivalent Standard Specification for Solid Round Glass Fiber Rein-
area of concrete. For columns, a limit tensile strain of 1% is forced Polymer Bars for Concrete Reinforcement.4 Analogous
set to ensure that failure in the GFRP bar will not occur. to ASTM A615 for steel,5 D7957 provides physical and me-
chanical property limits used by the designer and referenced
Key differences from conventional in ACI 440.11-22. In addition to providing consensus design
steel reinforcement values, D7957 describes a variety of ASTM test methods
used to qualify a given GFRP bar. The standard is somewhat
Some key differences between using GFRP and conventional prescriptive in nature in that it limits constituent materials to
steel reinforcement in concrete that may be considered limita- those that have been extensively tested and proved to provide
tions include no yielding before failure, low transverse strength, excellent long-term performance. A voluminous amount of
susceptibility to fire and smoke production, a high coefficient research, testing, and validation has gone into evaluating all
of thermal expansion perpendicular to the fiber direction, and aspects of FRP bar performance and in following a prescrip-
the inability to field bend bars; however, these limitations are tive method with testing of certain parameters, the ASTM
offset by some key advantages, including a high longitudinal committee has distilled this body of work so the designer will
strength–to–weight ratio, corrosion resistance, electromagnetic have the best assurance of good long-term performance.
neutrality, high fatigue resistance, low thermal and electrical
conductivity, a light weight, and ease of cutting on-site. The D7957 material standard goes beyond providing design
material limits by standardizing a series of qualifying char-
Some desirable applications for using GFRP reinforced con- acterization tests and a series of quality assurance tests to be
crete include the following:3 performed on a given production lot as shown in Fig. 2. It also
recommends sampling frequency for qualification and quality
• any concrete member susceptible to corrosion by chloride control as shown in Fig. 3.
ions or chemicals
GFRP bars from varying bar producers have different form
• any concrete member requiring nonferrous reinforcement factors and means of enhancing the surface of the bar to effect
because of electromagnetic considerations bond with the concrete. Thus, the concept of a nominal bar area
that is the same as that of A615 is used to calculate all proper-
• as an alternative to epoxy, galvanized, or stainless steel ties. Because there is a wide variation in bond enhancements
reinforcing bars on GFRP bars (sand coatings, helical wrap surfaces, helical
Figure 2. ASTM D7957 table of property limits and test methods for qualification. Source: Reprinted, with permission, from ASTM
D7957/D7957M-22 Standard Specification for Solid Round Glass Fiber Reinforced Polymer Bars for Concrete Reinforcement,
copyright ASTM International. A copy of the complete standard may be obtained from www.astm.org.

Figure 3. ASTM D7957 table of property limits and test methods for quality control and certification. Source: Reprinted, with per-
mission, from ASTM D7957/D7957M-22 Standard Specification for Solid Round Glass Fiber Reinforced Polymer Bars for Concrete
Reinforcement, copyright ASTM International. A copy of the complete standard may be obtained from www.astm.org.
wrapped and sand coated surfaces, ribbed lugs, and machined 1901.2.1 Structural Concrete with GFRP reinforcement.
lugs), D7957 provides a tolerance for deviation from nominal. Cast-in-place structural concrete internally reinforced
Using the Archimedes method, the measured cross-sectional with glass fiber reinforced polymer (GFRP) reinforce-
area of a given bar is determined and the measured area of the ment conforming to ASTM D7957 and designed in
bar must fall within the tolerances shown in Fig. 4. accordance with ACI CODE 440.11 shall be permitted
where fire resistance ratings are not required and only for
Characterization tests tend to be more elaborate in nature and structures assigned to seismic design category A.
take longer to perform. For example, ASTM D7705 testing
requires the bar to be subjected to an elevated-temperature The justification for the proposed change was described as
alkaline-solution bath for up to 90 days and residual tensile follows:
strength is measured to screen constituent materials for suit-
able long-term durability. The addition of this new standard allows the design
and construction of cast-in-place reinforced concrete
In addition to tests performed on straight lengths of GFRP bar, using non-metallic reinforcement bars. Currently the
D7957 describes testing and limits on fabricated bent shapes. design and construct requirements contained in the
The use of D7957 has standardized the GFRP bar industry standard are limited to use in Seismic Design Catego-
and allowed designers to implement bars without having to ry A. ACI Committee 440 developed this standard to
commit to a specific proprietary supplier. If designers follow provide for public health and safety by establishing
the values and limits in D7957 along with the design provi- minimum requirements for strength, stability, service-
sions of ACI 440.11-22, they will be following consensus ability, durability, and integrity of GFRP reinforced
standards that have been met by multiple suppliers and can be concrete structures.
validated and traceable to production lot certifications ensur-
ing a long lasting and safe implementation. The standard not only provides a means of establishing
minimum requirements for the design and construction
2024 IBC adoption and reference of GFRP reinforced concrete, but for acceptance of
design and construction of GFRP reinforced concrete
To this end, the 2024 International Building Code6 adopted a structures by the building officials or their designated
proposed change submitted by ACI as follows. representatives.
Figure 4. ASTM D7957 table of geometric properties. Source: Reprinted, with permission, from ASTM D7957/D7957M-22 Stan-
dard Specification for Solid Round Glass Fiber Reinforced Polymer Bars for Concrete Reinforcement, copyright ASTM Internation-
al. A copy of the complete standard may be obtained from www.astm.org.

The standard applies to GFRP reinforced concrete struc- • For flexural design, do not simply use a same-size or one-
tures designed and constructed under the requirements of to-one bar substitution in structural members. Compres-
the general building code. sion often controls design in flexural members reinforced
with GFRP reinforcement and is the favored response.
GFRP reinforced concrete is especially beneficial for
satisfying a demand for improved resistance to corrosion • For column design, ignore the presence of GFRP bars in
in highly corrosive environments, such as reinforced con- compression because they do not contribute anything and
crete exposed to salt water, salt air, or deicing salts. just use the overall area of concrete and its compressive
strength.
This standard establishes minimum requirements for
GFRP reinforced concrete in a similar fashion as ACI • For seismic design, use GFRP only to resist dead and live
318 Building Code Requirements for Structural Concrete loads.
establishes minimum requirements for structural concrete
reinforced with steel reinforcement. A separate standard • GFRP bars should not be used in lateral-load-resisting
is needed, as GFRP reinforcement behaves differently systems.
than steel reinforcement.
In detailing, note that 90-degree bends are the only practical
Currently GFRP is accepted for use to reinforce highway solution when fabricating stirrups because all bends are made
bridge decks. Acceptance is primarily in areas where at the factory during initial production, never in the field.
deicing salts are used on the roads and cause severe GFRP uses thermoset resins that cannot be reheated and only
corrosion to conventional steel reinforcement. This pro- no. 2 through no. 8 bars are allowed for bending.
posed change provides minimum requirements for other
applications where GFRP reinforced concrete is being Minimum GFRP bar development lengths are 12 in.
considered, such as marine and coastal structures, park- (305 mm) or 20 bar diameters. Mechanical splices must meet
ing garages, water tanks, and structures supporting MRI 1.25 of the guaranteed minimum ultimate tensile strength of
machines. Design reasons to use GFRP bars in structures the bar. Unlike steel, GFRP bars vary in strength by size, with
are resistance to corrosion in the presence of chloride the smallest being stronger per unit area.
ions, lack of interference with electromagnetic fields, and
low thermal conductivity. Durability of GFRP reinforced
concrete
Currently the standard prohibits the use concrete inter-
nally reinforced with GFRP for applications where fire Long-term durability performance of GFRP reinforced
resistance ratings are required. Chapter 6 of the Interna- concrete in field applications is questioned because of the
tional Building code cites applications for floors, roofs, material's linear elastic behavior and the use of separate
walls, partitions, and primary and secondary structural environmental reduction factors in the design process, even
frames where fire resistance ratings are not required. after knowing the materials are noncorrosive. Recent long-
term field studies8,9 evaluating GFRP reinforced concrete
structures constructed 15 to 20 years ago continue to demon-
Precast concrete design applications strate excellent long-term performance. Additional studies on
in-place performance have indicated no significant change in
ACI 440.11-22 may be applied to the design of precast concrete the properties of GFRP.10
members in similar fashion to ACI 318-19. ACI 440.11-22 al-
lows the use of nonmetallic corrosion-resistant reinforcement in Additional resources
the design of precast concrete beams, slabs, columns, and walls.
GFRP reinforcement has been used in major applications, such Additional information is available on the ACI web site at
as in civil structures including bridge decks and marine appli- Concrete.org.
cations including piers and seawalls. Precast concrete products
used in these areas will also benefit from the corrosion resis- References
tance of nonmetallic reinforcement. AASHTO LRFD Bridge
Design Guide Specifications for GFRP-Reinforced Concrete7 is 1. ACI (American Concrete Institute) Committee 440. 2022.
applicable to precast concrete bridge deck panels. Building Code Requirements for Structural Concrete
Reinforced with Glass Fiber-Reinforced Polymer (GFRP)
The key aspect to remember about designing with GFRP Bars—Code and Commentary. ACI 440.11-22. Farming-
compared with steel is that although strength usually governs ton Hills, MI: ACI.
with steel, it is most often deflection and crack control that
govern GFRP design because GFRP bars have a lower modu- 2. ACI Committee 318. 2019. Building Code Requirements
lus of elasticity. Additional recommendations for the design of for Structural Concrete (ACI 318-19) and Commentary
specific precast concrete components include the following: (ACI 318R-19). Farmington Hills, MI: ACI.

3. NEx Workshop on Designing Concrete Structures
Reinforced with GFRP Bars Using the New ACI Code
440.11-22 presented in Dallas, Texas, Oct. 26, 2022.
4. ASTM International Subcommittee D30.10. 2022. Stan-

dard Specification for Solid Round Glass Fiber Rein-
forced Polymer Bars for Concrete Reinforcement. ASTM
D7957. West Conshohocken, PA: ASTM International.
5. ASTM International Subcommittee A01.05. 2022. Stan-

dard Specification for Deformed and Plain Carbon-Steel
Bars for Concrete Reinforcement. ASTM A615. West
Conshohocken, PA: ASTM International.
6. International Code Council. 2024 International Building

Code. Country Club Hills, IL: International Code Coun-
cil, forthcoming.
7. AASHTO (American Association of Highway and Trans-

portation Officials). 2018. AASHTO LRFD Bridge Design
Guide Specifications for GFRP-Reinforced Concrete. 2nd
ed. Washington, DC: AASHTO.
8. Al-Khafaji, A. F., R. T. Haluza, V. Benzecry, J. J. Myers,

C. E. Bakis, and A. Nanni. 2021. “Durability Assessment
of 15- to 20-Year-Old GFRP Bars Extracted from Bridges
in the US—Part II: GFRP Bar Assessment.” Journal of
Composites for Construction 25 (2). https://doi.org/10
.1061/(ASCE)CC.1943-5614.0001112.
9. Benzecry, V., A. F. Al-Khafaji, R. T. Haluza, C. E. Bakis,

J. J. Myers, and A. Nanni. 2021. “Durability Assessment
of 15- to 20-Year-Old GFRP Bars Extracted from Bridges
in the US—Part I: Selected Bridges, Bar Extraction,
and Concrete Assessment.” Journal of Composites for
Construction 25 (2). https://doi.org/10.1061/(ASCE)
CC.1943-5614.0001110.
10. Benmokrane, B., C. Nazair, M. A. Loranger, and A.

Manalo. 2018. “Field durability study of vinyl-es-
ter-based GFRP rebars in concrete bridge barriers.” Jour-
nal of Bridge Engineering 23 (12). https://doi.org/10
.1061/(ASCE)BE.1943-5592.0001315.
Notation
ffu = design ultimate tensile strength of fiber-reinforced

polymer
fy = specified yield strength of reinforcement
Vc = nominal shear strength provided by concrete

About the authors Abstract
John J. Myers, PhD,

<Body>Mohamed K.PE,
Nafadi,
FACI, The American
Body text Concrete Institute has published ACI
PhD, is an
FASCE, FIAAM,
assistantFTMS,
professor
FVEST,
of 440.11-22, Building Code Requirements for Structur-
structural
is a professor
engineering
in the Civil,
at Assiut
Archi- Keywords
al Concrete Reinforced with Glass Fiber-Reinforced
University
tectural andinEnvironmental
Assiut, Egypt. He is Polymer (GFRP) Bars—Code and Commentary. This
a former graduate
Engineering Department
research and Bodycode
new text was developed by an American National
assistantof
director inthe
theMissouri
Department Center
of for Standards Institute–approved consensus process and
Civil, Construction,
Transportation Innovation
and Environ-
at the Review policy
addresses structural systems, members, and connec-
Missouri University of
mental
Science
Engineering
and Technology
at Northin tions, including cast-in-place, precast, nonprestressed,
Rolla. Carolina State University (NCSU) Bodycomposite
and text concrete construction. ACI 440.11-22
in Raleigh. is the first comprehensive building code covering the
Douglas Gremel is the director of Reader comments
use of nonmetallic, GFRP reinforcing bars in structural
Omar M. Khalafalla,
engineering at OwensisCorning
a graduate concrete applications. This article provides background
research and teaching
Infrastructure Solutions assistant
LLC in Body
on thistext
new code and discusses potential uses for pre-
and PhDNeb.
Seward, candidate
He has inathe
BSDepart-
in cast concrete components and structures.
ment of Civil,
engineering science,
Construction,
electrical and
Environmental
engineering, andEngineering
business admin-at Keywords
NCSU.
istration from Colorado State
University in Fort Collins. Composite construction, GFRP, glass-fiber-reinforced
Gregory W. Lucier, PhD, is a polymer, reinforcing bar.
research
Alvin assistant
Ericson, professor
FACI, FPCI, in is the
an
Department of Civil, Construction,
independent technical marketing Publishing details
and Environmental
consultant Engineering
specializing in fiber-
and
reinforced-polymer Constructed
manager of the composite and This paper appears in PCI Journal (ISSN 0887-9672)
Facilities Laboratory
precast concrete at NCSU.
construction V. 68, No. 2, March–April 2023, and can be found
systems. He has a BS in architec- at https://doi.org/10.15554/pcij68.2-05. PCI Journal
Sami H. Rizkalla,
ture from PhD, FPCI,
the Massachusetts is published bimonthly by the Precast/Prestressed
FACI, FASCE, FIIFC,
Institute of Technology in Cambridge and an FEIC,
MBA Concrete Institute, 8770 W. Bryn Mawr Ave., Suite
from Northeastern University in Boston, Mass.Profes-
FCSCE, is Distinguished 1150, Chicago, IL 60631. Copyright © 2023, Precast/
sor of Civil Engineering and Prestressed Concrete Institute.
Construction,
Chad Van Kampen,
directorMASCE,
of the PE,
Constructed
LEED AP BD+C,
Facilities
LPCI, Laboratory,
is a Reader comments
and director
licensed professional
of the National
engineer in
Science Foundation
multiple states and manager
Center on of Please address any reader comments to PCI Journal
preconstruction with Fabcon
Integration of Composites into editor-in-chief Tom Klemens at tklemens@pci.org or
Precast in Grandville,
Infrastructure at NCSU. Mich. He Precast/Prestressed Concrete Institute, c/o PCI Jour-
has a BS in civil engineering from nal, 8770 W. Bryn Mawr Ave., Suite 1150, Chicago, IL
Calvin College in Grand Rapids, Mich., and extensive 60631. J
senior-level management experience with large
complex projects centered on the precast concrete
construction industry. An active member of PCI and
AltusGroup, Van Kampen serves as chair of the PCI
FRP Composites Committee and Total Precast Systems
Committee and vice chair of the PCI Hollow Core
Committee.

Design
Design and Cyclic Testing of a Gusset Plate Connection for Precast
Concrete Buckling-Restrained Braced Frames 30
Investigation of Repair Techniques for Deteriorated End Regions

of Prestressed Concrete Bridge Girders 52
Experimental Background behind New AASHTO LRFD Specifications

for Partially Debonded Strands 71
Mathew W. Bolduc, Bahram M. Shahrooz, Kent A. Harries, Richard A. Miller, Henry G. Russell, and William A. Potter
Influence of Structural Form on Hydration-Heat-Induced Temperature

Rise of Precast Concrete Lining Segments for a Metro Transit Station 88
Design and cyclic testing of a gusset
plate connection for precast concrete
buckling-restrained braced frames
B
uckling-restrained braced frames (BRBFs) have
become a well-established lateral-force-resisting
system for steel construction. Buckling-restrained
braces implemented in steel structures have been shown
to possess enough rigidity to satisfy structural drift limits,
provide significant energy absorption, and reduce forces on
foundations and adjacent members.1–4 Testing procedures
for buckling-restrained brace subassemblages and design
procedures for steel structures using buckling-restrained
braces have been codified in the American Institute of
Steel Construction’s Seismic Provisions for Steel Buildings
(AISC 341) since the 2005 edition.5,6 In contrast, there is an
insufficient amount of laboratory experiments from which
a codified method of design for BRBFs for precast concrete
systems can be developed and, before the study described in
this paper, no U.S. laboratory experiments had been con-
ducted on the seismic performance of these systems.
■ A test specimen was designed and constructed to
represent a scaled partial model of a precast con- Presently, the most common types of lateral-force-resisting
crete buckling-restrained braced frame. A gusset system for precast concrete structures are shear walls and
plate connection was designed using the uniform moment frames.7 Because both shear walls and moment
force method (UFM). frames can include cast-in-place elements, post-tensioned
connections, and grouted connections, some of the inherent
■ The specimen was tested under representative seis- benefits of selecting a precast concrete system—such as
mic loads. quick erection time, improved quality control, and lower
project costs—are limited.8
■ Study objectives were to test a partial system un-
der representative seismic loads and to determine Buckling-restrained braces were recently used as the lat-
the applicability of the force distribution assumed eral-force-resisting system for a precast concrete structure
by UFM. in the New Madrid seismic zone of the United States.9 The

braces were selected because they eliminated the need for Seismic Performance Factors (FEMA P695)14 methodology
moment connections and shear walls and provided sufficient could be given a response modification factor of 8 for seismic
load capacity and seismic drift levels. Prior to approval, local design. This model did not explicitly capture damage within
building authorities required that the project team justify the the joints, but the system behavior of the model was validated
use of this novel system. If the use of precast concrete BRBFs against the experimental results of Guerrero et. al.10 Given the
could be codified, it could reduce erection times and project promising experimental and numerical results for system-level
costs for precast concrete construction in seismic zones. performance of precast concrete BRBF systems, investigation
of the behavior of connections from steel BRBFs to precast
Precast concrete BRBFs have only been examined via labo- concrete frame systems is warranted.
ratory experiments by Guerrero et al.10 Although the BRBFs
showed promising seismic performance, the detailing of the The most common design procedure for steel gusset plate
system reflected the standards used in Mexico and the results connections is the UFM, which was developed by Thornton15
may not directly translate to structures in the United States. in conjunction with a joint American Society of Civil
Numerical studies conducted by Oh et al.11 indicated excellent Engineers (ASCE) and AISC task group and first appeared in
seismic performance of a precast concrete frame system with the AISC Manual of Steel Construction16 in 1992. The UFM
steel buckling-restrained braces but did not explicitly account was created to satisfy static equilibrium and assigns dimen-
for joint failure modes. sions of the gusset’s connected edges such that no moments
occur on any of the three connection interfaces (gusset-to-
The study presented in this paper aimed to fill a knowledge beam, gusset-to-column, and beam-to-column). Because the
gap by experimentally evaluating a proposed connection from load path through the gusset plate depends on the stiffness of
a buckling-restrained brace to a precast concrete beam and the connections and the members that they attach, the analysis
column. Design options for buckling-restrained brace connec- is indeterminate. This indeterminacy was neglected in the
tions and their impact on precast concrete beam and column derivation of the UFM in the interest of creating an easily
designs were first examined. Given concerns about high usable design method for distribution of interface forces due
gravity loads being transferred from the beam into the gusset to brace action.
plate, several statically determinate connection types that pre-
vented gravity load transfer were first examined.12 However, When Thornton derived the UFM,15 he acknowledged that
even the most promising of these options, a lug connection distortion (referred to as frame action in this paper) causes
that connected only to the corbel, would have required uncon- additional positive and negative forces on the gusset due to
ventional design procedures. Ultimately for these reasons, a the opening and closing of the angle between the beam and
gusset plate connection designed by the uniform force method the column; however, the UFM recommends that frame action
(UFM) was selected for physical testing. be ignored for concentric connections. In the AISC manual,16
this recommendation was justified by the agreement between
The objective of this study was to determine the applicabil- the UFM’s predictions and the experimental forces from ide-
ity of the force distribution assumed by the UFM to precast alized experimental tests by Gross and Cheok17 and Bjorhovde
concrete BRBFs via a cyclic test on a gusset plate connection and Chakrabarti18 (described later). Figure 1 shows variable
designed by UFM. Using a test approach adapted from buck- and interface force locations for the UFM.
ling-restrained brace–to–steel frame connection tests executed
by Coy,13 investigators used two servo-controlled hydraulic In the UFM, the following equations are used.
actuators to simulate the behavior of a partial precast concrete
BRBF under load due to a seismic event. To assign the dimensions of the gusset connected edges based
on the beam, column, and brace geometry, the following
Literature review equation was used:
The design of gusset-type connections to precast concrete α = tan θ ( β + eb ) − ec

systems is of interest to researchers because recent exper-
imental and numerical research has indicated promising where
performance of precast concrete BRBF systems. Guerrero et.
al.10 compared the performance of a one-third-scale precast α = distance from the face of the column to the centroid
concrete BRBF and a precast concrete frame of the same scale of the gusset-to-column connection
constructed without BRBFs and found that damage in the
beams, columns, and joints was reduced in the BRBF speci- θ = angle between the centroid of the column and the
men. However, as noted previously, this study was performed centroid of the brace
in Mexico and the detailing was different from that specified
in U.S. concrete design standards. Numerical research by Oh β = distance from the face of the beam to the centroid
et. al.11 demonstrated that a precast concrete frame system of the gusset-to-column connection
with buckling-restrained braces evaluated by the Federal
Emergency Management Agency’s Quantification of Building eb = one-half the depth of the beam

Figure 1. Variable definitions and location of gusset plate interface forces for the uniform force method. Note: eb = one-half
the depth of the beam; ec = one-half the depth of the column; Hb = required shear force on the gusset-to-beam connection; Hc
= required axial force on the gusset-to-column connection; Vb = required axial force on the gusset-to-beam connection; Vc =
required shear force on the gusset-to-column connection; W.P. = work point; α = distance from the face of the column to the
centroid of the gusset-to-column connection; β = distance from the face of the beam to the centroid of the gusset-to-column
connection; θ = angle between the centroid of the column and the centroid of the brace.
ec = one-half the depth of the column P = required brace axial force

eb
To derive the distance between the work point and the cen- Vb = P
troid of the gusset plate, the following equation was used: r
where
r= (α + ec )2 + ( β + eb )2
Vb = required axial force on the gusset-to-beam
where connection
ec
r = distance between the work point and the gusset Hc = P
centroid r
where
To distribute the brace load to the beam and column connect-
ed interfaces as shear and axial forces, the following equa- Hc = required axial force on the gusset-to-column
tions were used: connection
α β
Hb = P Vc = P
r r
where where
Hb = required shear force on the gusset-to-beam Vc = required shear force on the gusset-to-column
connection connection

The UFM was validated based on two different sets of ideal- buckling-restrained braces. Maheri and Yazdani23 created
ized steel gusset connection tests. The first set of tests, com- finite element models tuned to the benchmark experimental
pleted by Gross and Cheok,17 were executed on subassem- tests of Maheri and Hadjipour24 on cast-in-place concrete
blages consisting of a stub of a continuous column between joints with steel gusset connections. The stub beam and
two floors, two stub braces, and one stub beam subjected to column ends were allowed to freely translate and rotate,
lateral loading. These stub members were all pinned at their the beam-to-column joint was pinned, and pure tension
midspans, the theoretical points of zero internal moment, was applied along the stub brace’s longitudinal axis. Once
during the tests. The gusset was connected to the beam using sufficient agreement between the experimental results and
a fillet weld and to the column using clip angles. The clip initial finite element models was achieved, Maheri and
angle was extended along to the beam web to create a pinned Yazdani undertook a parametric study that varied the gusset
connection between the beam and the column. Although geometry and brace angle. They concluded that the UFM
effort was made to simulate full-frame behavior, the boundary could be applied conservatively for the design of steel brace
conditions of these tests were highly idealized. The UFM con- connections to reinforced concrete structures. This work was
servatively predicted a lower capacity than the actual tested highly idealized, much like the tests of Gross and Cheok17
capacities for all of the experiments by Gross and Cheok, with and Bjorhovde and Chakrabarti.18 Tsai et al.25 investigated the
an average difference of 5%. applicability of the UFM to gusset plates, attaching buckling-
restrained braces to fixed cast-in-place concrete full frames.
The second set of tests, completed by Bjorhovde and They concluded that the generalized UFM, a more general
Chakrabarti,18 were executed on subassemblages consisting of version of the UFM proposed by Muir,26 was acceptable for
a stub continuous column between two floors, one stub brace, prediction of force distribution at gussets in foundation-col-
and one stub beam loaded in tension at the free end of the stub umn-brace connections because frame action is minimized by
brace. The stub column was pinned at both ends, and the ends the rigid foundation. Tsai et al. did not find generalized UFM
of the stub beam and brace were left to rotate and translate alone to be acceptable for beam-column-brace connections.
freely. As in the tests by Gross and Cheok, the gusset was
connected to the beam using a fillet weld and to the column Experimental test program
using clip angles. This clip angle was extended to create the
same type of pinned beam-to-column connection. Although The work initiated in the study described in this paper aims to
effort was made to simulate full-frame behavior, the bound- determine whether a steel gusset plate connection designed by
ary conditions of these tests were highly idealized. The UFM the UFM would be sufficient for seismic design of a precast
conservatively predicted a lower capacity than the actual concrete system with buckling-restrained braces. Although
tested capacities for all of the experiments by Bjorhovde and the shortcomings of the UFM have been noted in the studies
Chakrabarti, with an average difference of 52%. with fixed connections, it remains the most common method
for sizing and design of steel gusset plates and distributing the
As stated, Thornton proposed that frame action can be ignored brace force to the connected components. In addition, precast
for concentric connections in his initial derivation of the concrete systems are much closer to a pinned condition than
UFM;15 however, more recent studies investigating the UFM’s the fixed frames investigated by Tsai et al.25 For these reasons,
application to full steel frames with fixed connections19–22 the UFM was selected as the design method for the gusset
have shown that frame action can alter the gusset interface plate in this first experimental test on this system.
forces significantly. Chou et al.19 concluded that the forces
developed by frame action and brace action are similar at low Prototype structure and design loads
frame displacements; however, at high frame displacements,
the forces developed by frame action exceed those developed A four-story parking structure (three elevated levels) was
by brace action. Chou et al. also proposed an equation to selected as the prototype structure for the experiment. BRBFs
predict the normal and shear forces developed by frame action were a replacement for the shear walls typically used as the
based on the geometry of the connection, the stiffness of the lateral-force-resisting system in parking structures. Two bays
gusset plate, the stiffness of the beam, and the stiffness of of BRBFs per story were required to replace one shear wall.
the column. Lin et al.20,21 concluded that forces due to frame The story height was 9 ft 11⅞ in. (3.045 m), and the bay
action cause the experimental interface force distribution to width was 16 ft 0 in. (4.877 m). These dimensions represent
deviate significantly from the UFM’s predictions. Lin and the average story height and average bay width derived from
colleagues found that the forces due to frame action were three example parking structure projects provided by a local
additive to the interface shear forces and subtractive to the precast concrete producer. The resulting angle between the
interface normal forces. Cui et al.22 concluded similarly that brace and the beam was approximately 32 degrees. Figure 2
gusset connection design should account for both brace action illustrates the full-scale prototype bay.
(force distribution predicted by the UFM or a similar model)
and frame action. The test specimen was scaled to one-third for area, 0.577 for
length, and one-third for force. This ensured equivalent stress-
Similarly, limited research has been conducted on braced es between the test and prototype structures. The maximum
frames consisting of cast-in-place concrete members and steel 90 kip (400 kN) brace force in the test structure corresponded

16'-0" 16'-0"
32°
9'-117
8"
PORTION OF PROTOTYPE
SCALED AND CONSTRUCTED
FOR THIS TEST
Figure 2. Full-scale prototype bay diagram, including partial frame constructed for the test program. Note: 1” = 1 in. = 25.4 mm;
1’ = 1 ft = 0.305 m.
to a 270 kip (1200 kN) brace force in the full-scale prototype. beam-to-column bearing and a difference between the pre-
This 270 kip load was greater than the required brace loads scribed and as-built conditions caused a ¾ in. (19 mm) gap
of all floors of a six-story seismic design category B example between the bottom of the beam and the top of the embedded
parking structure and all five stories of a seismic design cate- plate at the corbel. This gap was accounted for by assigning
gory C example parking structure. load to the beam and column using the full length of the
gusset (assuming no gap) at the beam and column interfaces
The buckling-restrained brace manufacturer, CoreBrace, used but applying the column vertical and horizontal force com-
a frame model designed using structural analysis software to ponents at the center of the connected interface (accounting
determine a full-scale prototype brace size and overstrength for the gap) at the column. To maintain equilibrium, a small
factors that would induce a maximum adjusted brace strength moment was developed at the column interface. The design
of 270 kip (1200 kN) in compression at the elongation of the gusset plate and its connections accounted for this
caused by 2% story drift, Δh2%. The maximum adjusted brace moment, which is described by the following equation and
strength, overstrength factors, and frame displacements were shown in Fig. 3.
scaled appropriately for design of the test specimen. All test l −l ⎞
l ⎛
specimen brace connections and members adjoining the brace M c = Vc ec + Vb ⎛⎜ ec + b ⎞⎟ − H b eb − H c ⎜ eb + lg + c g ⎟
⎝ 2⎠ ⎝ 2 ⎠
were designed to resist this maximum adjusted brace strength
per AISC 341-16 section F4.27 Gravity and live loads were where
determined in accordance with ASCE 7-16, Minimum Design
Loads and Associated Criteria for Buildings and Other Mc = required moment on the gusset-to-column interface
Structures.28
lb = length of gusset interface connected to beam
Test specimen design and construction
lg = length of gap between bottom of beam and top of
Because the test frame was designed as pinned, a partial frame corbel
could be tested in lieu of the whole frame. The members of
the partial frame were scaled but were of complete scaled lc = length of gusset interface connected to column
length; this design meant that a realistic pinned column base
connection could be tested and realistic beam and column All relevant gusset, bolt, and weld failure modes were
curvature could be monitored. The prototype precast concrete checked in accordance with the corresponding sections J2
specimen in context of the full frame is shown within the and J3 of ANSI/AISC 360-16, Specification for Structural
dashed outline in Fig. 2. Steel Buildings30 using directions from AISC’s Design Guide
29: Vertical Bracing Connections—Analysis and Design.31
The gusset plate dimensions were sized using the test loads Although 3/16 in. (4.8 mm) fillet welds using 80 ksi (552 MPa)
(neglecting gravity loads) and the UFM as it is outlined in electrodes were specified along the gusset’s connected edges,
chapter 13 of the 15th edition of the AISC Steel Construction ⅜ in. (9.5 mm) fillet welds using 70 ksi (483 MPa) electrodes
Manual.29 The randomly oriented fiber bearing pad at the were fabricated. This difference in the as-built condition

Figure 3. Interface force distribution adjusted for gap between the beam and corbel. Note: eb = one-half the depth of the beam;
ec = one-half the depth of the column; Hb = required shear force on the gusset-to-beam connection; Hc = required axial force on
the gusset-to-column connection; lb = length of gusset interface connected to beam; lc = length of gusset interface connected to
column; lg = length of gap between bottom of beam and top of corbel; Mc = required moment on the gusset-to-column interface;
Vb = required axial force on the gusset-to-beam connection; Vc = required shear force on the gusset-to-column connection.
added additional conservatism to the design. Yield of em- F11 and G4.30 Although ⅞ in. (22.2 mm) fillet welds using
bedded plates at the beam and column were checked using 80 ksi (552 MPa) electrodes were specified along the jumper
finite element models.12 Bars welded to embedded plates were plate edges, fillet welds with a 5/16 in. (7.9 mm) root that used
designed to resist combined shear and tensile forces in accor- 110 ksi (758 MPa) electrodes had to be fabricated due to
dance with chapter 22 of the American Concrete Institute’s warping of the embedded plates. The strength of the as-built
Building Code Requirements for Structural Concrete (ACI condition was sufficient when rechecked in accordance with
318-19) and Commentary (ACI 318R-19).32 ANSI/AISC 360-16 section J2. The reinforcing bars welded to
the embedded plates of the jumper connection were designed
The jumper plate connection was modified from a typical in the same manner as those incorporated into the gusset plate
precast concrete beam-to-column connection provided by connection. Figure 4 shows the jumper plate and gusset plate
a local precast concrete producer. Typical connections use connections.
headed studs that do not transfer uplift. Because the connec-
tion for the test specimen was required to transfer uplift at Although this test frame should have, in theory, been
this location, deformed bar anchors with headed terminators pinned—meaning it should not have been required to transmit
were necessary. Shear and flexure of the jumper plates were large moments—the beam and column were detailed to resist
checked in accordance with ANSI/AISC 360-16 sections moment because it provided the most ductility and conserva-

7"
A572 JUMPER PLATES,
8
7"
FILLET WELD WITH 80 KSI ELECTRODES
8
(4) NO. 4 AND (3) NO. 6 EMBEDDED

(4) NO. 5 AND (3) NO. 6 BARS FOR JUMPER CONNECTION
EMBEDDED BARS FOR
JUMPER CONNECTION BEAM
(8) NO. 4 EMBEDDED
BARS FOR GUSSET
ATTACHMENT
3"
16
COLUMN
1"
A572 GUSSET PLATE
2
3"
16
(8) NO. 6 AND (6) NO. 5 EMBEDDED

BARS FOR GUSSET ATTACHMENT
Figure 4. Jumper plate and gusset plate connections. Note: no. 4 = 13M; no. 5 = 16M; no. 6 = 19M; 1” = 1 in. = 25.4 mm; 1 ksi =
6.895 MPa.
tism for this precast concrete BRBF design. The beam was where factored normal load is less than factored shear load.
designed to meet the requirements of ACI 318-1932 section The cantilever beam method (ACI 318-19 section 16.5) was
18.6, “Beams of Special Moment Frames,” and detailed in selected for this frame because it is more commonly used
accordance with ACI 318-19 chapter 25. The column was than the strut-and-tie method (ACI 318-19 chapter 23) and
designed to meet the requirements of ACI 318-19 section because it can provide a potential data point on the perfor-
18.7, “Columns of Special Moment Frames,” and detailed mance of corbels designed by the cantilever beam method for
in accordance with ACI 318-19 chapter 25. The resulting cases where the factored normal load is less than the factored
scaled beam design used Grade 60 (414 MPa) reinforcing bars shear load. The resulting scaled corbel design used Grade 60
and 6000 psi (41 MPa) concrete. The full beam reinforcing (414 MPa) reinforcing bars and 6000 psi (41 MPa) concrete.
schedule is shown in Fig. 5, and the full column reinforcing The corbel reinforcing schedule is shown in Fig. 6.
schedule is shown in Fig. 6.
A widely used pinned precast concrete column base connec-
The corbel was designed to meet the requirements of ACI tion was chosen for this frame. This connection consisted of a
318-1932 section 16.5, “Brackets and Corbels,” and detailed in base plate that was embedded in the column base using rein-
accordance with ACI 318-19 chapter 25. The factored normal forcing bars welded to the plate. The base plate was flush with
load on the corbel was greater than the factored shear load, in the outside dimensions of the column, so the four anchor bolts
violation of the requirement in ACI 318-19 section 16.5.1.1 were inset in pockets at the plate corners. The base plate was
for the use of the cantilever beam method of corbel design. designed according to section 6.11 of the eighth edition of the
The section 16.5.1.1 requirement is stipulated because the PCI Design Handbook.7 The lapped splices between the rein-
cantilever beam method has only been validated for cases forcing bars welded to the base plate and the longitudinal bars

RECESSED
POCKETS AT BEAM
CORNERS FOR
BOLT ACCESS
1"
7'-11 4 1"
12
A B C
1'-2"
43
4"
22" MINIMUM LAP LENGTH
1"
A B 22 C
SECTION A-A SECTION B-B SECTION C-C

13" HOLE,
1'-2" 1'-2" (4) NO. 6 PERIMETER
1-1/4" A572
Ø16
3" 3" 3" 3" LONGITUDINAL BARS, TYPICAL TYPICAL AT
15
8" 58 58 15
8" 15
8" 58 58 15
8"
BASE PLATE
15
8" 15
8" 1"
CORNERS
15
38
8"
1" 17
8"
38
53
8"
3"
58 17
8"
(1) NO. 4 CLOSED LOOP TIE
1'-2" 1'-2" AT 2-1/2" O.C. BETWEEN
5 BOLT POCKETS (START
16"
53
8" (2) NO. 3 SINGLE-LEG
3"
58 CENTER OF FIRST LOOP
TYPICAL AT 1-1/4" OFF OF BASE PLATE)
STIRRUPS AT 3" ON NO. 5 BARS
CENTER
(2) NO. 5 x 2'-4" LONG BARS
15
8" 47 1" 47 15
8"
WELDED TO PLATE AT
15
8" 44 8" EACH CORNER, TYPICAL 15 8"
8"
3
(8) NO. 6 LONGITUDINAL BARS (2) NO. 3 SINGLE-LEG 8"
STIRRUPS AT 3" ON
TYPICAL AT
(8) NO. 6 CENTER
NO. 6 BARS
LONGITUDINAL BARS (1) NO. 3 CLOSED LOOP
TIES WITH 135° HOOK AT 3" (1) NO. 3 CLOSED LOOP
ON CENTER TIES WITH 135° HOOK AT 3"
ON CENTER
Figure 5. Beam reinforcing schedule. Note: no. 3 = 10M; no. 4 = 13M; no. 5 = 16M; no. 6 = 19M; 1” = 1 in. = 25.4 mm;
1’ = 1 ft = 0.305 m.
that were terminated above the bolt pockets were designed previously. The warping of the base plate at the column
to meet the requirements of ACI 318-19 section 25.5.2.32 caused the two bolts farthest from the strong floor to not be in
The base plate used ASTM A57233 Grade 50 (345 MPa) steel contact with the reaction frame. Washers were added between
and was 1.25 in. (31.8 mm) thick, the reinforcing bars were the base plate and reaction frame at these locations for better
Grade 60 (414 MPa), and 1.25 in. diameter Grade A490 bolts contact.
were used. The same connection was used to attach the beam
to its actuator. These connections can be viewed in section Reaction frame and loading protocol
C-C of Fig. 5 and section C-C of Fig. 6.
A reaction frame with a typical section size of HP18x135 was
Typically, bolt pockets would be grouted solid after the used for testing this specimen horizontally on the floor of the
column is installed and the bolts or threaded rods have been testing facility (Fig. 7–8). Three out-of-plane restraints—a
tightened sufficiently; however, in this study, the column base corner brace against the precast concrete specimen itself, a
was designed to provide sufficient shear capacity without brace on the beam actuator, and a brace on the brace actu-
grouting the bolt pockets so that the test specimen could be ator—were provided to ensure that the specimen could not
easily removed from the frame. move more than 0.05 in. (1.3 mm) in the vertical direction
during testing. These restraints were in contact with the
Several challenges were encountered during the design component they were restraining, but the component could
process. First, it was difficult to find space to place the large slide using plastic sheets with low coefficients of friction and
amounts of reinforcing that were required both physically and grease. The precast concrete specimen itself was supported
by the special moment frame detailing requirements inside using greased steel rollers.
the beam and column. Second, there were several tolerance
issues due to warping of the steel embedded plates prior to Gravity load was neglected on the beam and column in the ex-
testing. Warping occurred at the jumper embedded plates perimental setup. It was assumed that the buckling-restrained
and at the column base plate. The modification of the jumper braces and gusset plates would be installed in the prototype
plate welded connection to adjust for warping was discussed structure after the floor topping had been poured and cured,

SECTION A-A
TIES IN THIS DETAIL APPLY TO FULL
DEPTH OF CORBEL
NO. 3 FRAMING BAR
1'-9"
158"
1'-2"
15 67
8" 13
4"
8" 3"
18
3"
58
1" 1'-2" 1"

11 4
1'-24
3"
58
7"
3"
A
A
18
15
8"
(8) NO. 6 (1) NO. 3 CLOSED NO. 3 FRAMING BAR
LONGITUDINAL BARS LOOP TIE AT 3" ON
CENTER (1) NO. CLOSED
LOOP TIE WITH 135°
(1) NO. 3 CLOSED LOOP HOOK AT 3" ON
1'-55 TIES WITH 135° HOOK AT 3" CENTER
8"
ON CENTER
SECTION B-B
1'-2"
1" 15 3"
58 3"
58 15
14
15 8" 8"
8"
B
1"
6'-44
3"
58
1'-2"
3"
58
15
8" (1) NO. 3 CLOSED
LOOP TIE AT 3" ON
(8) NO. 6 CENTER
LONGITUDINAL BARS
(1) NO. 3 CLOSED LOOP
TIES WITH 135° HOOK AT 3"
22" MINIMUM LAP LENGTH ON CENTER
SECTION C-C
(4) NO. 6 PERIMETER Ø13
16" HOLE,
1-1/4" A572
LONGITUDINAL BARS, TYPICAL TYPICAL AT
BASE PLATE
CORNERS
1" 1"
15
38
22 RECESSED POCKETS
8"
AT BEAM CORNERS 17
1"
38 8"
FOR BOLT ACCESS
C
17
8"
(1) NO. 4 CLOSED LOOP TIE AT
2-1/2" O.C. BETWEEN BOLT
1" 5 POCKETS (START CENTER OF
12 16"
FIRST LOOP 1-1/4" OFF OF BASE
TYPICAL AT PLATE)
NO. 5 BARS
(2) NO. 5 x 2'-4" LONG BARS

WELDED TO PLATE AT EACH 15
8"
CORNER, TYPICAL 15
8"
3
8"
TYPICAL AT
NO. 6 BARS
Figure 6. Column and corbel reinforcing schedules. Note: no. 3 = 10M; no. 4 = 13M; no. 5 = 16M; no. 6 = 19M; 1” = 1 in. = 25.4 mm;
1’ = 1 ft = 0.305 m.
preventing as much gravity load transfer as possible. The the corbel (20.8 kip [92.5 kN]). Although column gravity
gravity load that would pass through the corbel was 1.15 kip loads would have been present in the prototype structure, the
(5.12 kN), which was less than 6% of the UFM estimate of primary focus of the experimental work was to characterize
the highest vertical portion of the brace load to pass through load path in the connection, which would be largely unaffect-

Figure 7. Precast concrete specimen (white) installed in the reaction frame.
CONNECTION TO STRONG CONNECTION TO STRONG

FLOOR, OOP RESTRAINT ONLY, FLOOR, OOP RESTRAINT ONLY,
FREE TRANSLATION IN PLANE SP3 SP2 FREE TRANSLATION IN PLANE
ACTUATOR PINNED
ATTACHMENT TO
REACTION FRAME
SP9 DG6
SP1
SP8
FIG. A
FIG. A FIG. B
DG8
2" 1"
92 1"
92 2"
2"
GAUGES
SP7 1-4
83
4"
FIG. B 1
83
4" DIC ON 2
TOPSIDE OF 3
4
PLATE
SP6 2"
VIEW FROM TOP
RECTANGULAR STRAIN
OF CORBEL
ROSETTE ON
SP10 UNDERSIDE OF PLATE GAUGES 5-8
DG7
ACTUATOR PINNED
PINNED COLUMN BASE ATTACHMENT TO
ATTACHMENT TO FRAME REACTION FRAME
DG1 DG5
STRING POTENTIOMETER
DG2 DG3 DG4
DIAL GAUGE MEASURING IN-PLANE MOTION
DIAL GAUGE MEASURING OUT-OF-PLANE MOTION
Figure 8. Instrumentation plan, including gusset plate instrumentation (Fig. A) and corbel reinforcing bar instrumentation (Fig. B).
Note: DG = dial gauge; DIC = digital image correlation; OOP = out-of-plane; SP = string potentiometer. 1” = 1 in. = 25.4 mm.

ed by gravity load in the column. Only test loads were used to prescribed and unadjusted experimental hysteresis divided into
design the gusset plate connection. the seven procedures executed are shown in Fig. 9.
The behavior of a buckling-restrained brace was simulated Instrumentation

with a servo-controlled hydraulic actuator. The hysteresis of a
representative buckling-restrained brace with a peak com- The instrumentation plan is shown in Fig. 8. Rectangular
pression force of 270 kip (1200 kN) was generated using the rosette strain gauges were applied along the connected edges
backbone of hysteresis provided by CoreBrace and a re-cre- to validate the force distribution predicted by the UFM.
ated version of the hysteresis generation procedure proposed These rectangular rosette gauges allowed for the deter-
by Coy.13 This process involved the derivation of multiple mination of normal and shear strains along the connected
stress-strain relationships (consisting of portions of linear and interfaces in a manner similar to that used in other gusset
nonlinear behavior) for several strain levels. These stress- plate tests in the literature.17,18,22 The strain gauges were
strain relationships were validated with existing experimental applied to the underside of the gusset plate (Fig. 8). Linear
data and comparison between the experimental hysteresis and strain gauges were applied to each of the eight corbel bars at
the hysteresis predicted using the derived stress-strain rela- the critical section of the corbel (Fig. 8) because the corbel
tionships yielded a satisfactory level of accuracy.12 primary tension reinforcing bars were determined to be
the weakest link if the assumed UFM load distribution was
The hysteresis was then scaled appropriately for the one- correct. The gauges on the corbel bars were applied at the
third-length scale specimen shown in white in Fig. 7. The full theoretical point of highest stress.
hysteresis was derived for the simulated brace at the following
fractions of the strain, strain levels, in the brace at 2% story Both of the servo-controlled hydraulic actuators were
drift: 1/8, ¼, ½, ¾, and 1. In addition, this full hysteresis includ- equipped with integrated load cells and linear variable dis-
ed strain levels corresponding to one-half yield strain and yield placement transducers (LVDTs). These allowed for monitor-
strain. Each strain level, hereafter referred to as procedure, of ing of internal force and displacement in the precast concrete
this full hysteresis was broken into multiple discrete steps for beam and the simulated brace. Curvature of the beam and
testing. Each procedure was planned to have two cycles, that column were monitored using string potentiometers and dial
is, each procedure was planned to be executed twice. The full gauges. String potentiometers 1 and 10 were placed along
Figure 9. Prescribed and experimental hysteresis for procedures 1 through 7 (P1-P7). Note: 1 in. = 25.4 mm; 1 kip = 4.448 kN.

the longitudinal axes of the beam and column and monitored correlation measurements were not taken for procedure 7.
axial deformation (Fig. 8). String potentiometers 6 through 9 Manual measurements for string potentiometers 1, 2, and
were placed perpendicular to the column longitudinal axis at 3 and all dial gauges were taken at points where the proce-
quarter points along the column height to measure curvature. dure changed slope. The precast concrete beam and column
String potentiometers 2 and 3 were placed perpendicular to were whitewashed to improve the visibility of cracks that
the beam longitudinal axis at the first two quarter points along formed during testing. After each cycle was completed,
the beam length to measure curvature. the accessible sides of the beam and column were visually
inspected with the aid of a flashlight. The length and width
One dial gauge was placed at the column base to monitor the of any cracks found during this inspection were measured.
slip of the column base. Another dial gauge was placed at the After the cracks were measured, the cracks were traced with
five-eighths point of the column (measured from its base) to marker and the procedure and cycle number were written
monitor the out-of-plane movement of the precast concrete next to the crack.
specimen. The in-plane movement of the reaction frame was
also monitored through use of dial gauges. These gauges were After all measurements were taken and cracks were inspected
placed at the free corners of the reaction frame and at the for a given cycle, the next cycle was completed in the same
actuator-to-frame attachment points. fashion. This process continued until two cycles of each
procedure had been completed consecutively for procedures
Digital image correlation was used along with the physical 1 through 6. Procedure 7 was terminated early because the
strain gauges to provide a complementary set of measure- beam actuator capacity was reached.
ments. Digital image correlation was used to measure the
deformations on the top of the plate in the areas not covered Results and discussion
by the lug connector. A two-camera setup in a stereo system
was used to capture the full three-dimensional (3-D) defor- Frame force versus displacement
mation of the plate to understand bending and out-of-plane behavior
deformation of the plate occurring during loading. The plate
was painted white and stamped black with a random speckle Figure 10 shows a comparison of the prescribed hystere-
pattern. Images were captured with Flir Grasshopper3 digital sis and the load and frame displacement at the cycle peaks
cameras and Schneider Kreuznach Xenoplan 1.9/35 lenses. adjusted to account for slip at the column base. The slip was
The stereo setup was 70 in. (1778 mm) above the specimen, accounted for by subtracting the slip at each peak from the
with the cameras spaced 16 in. (406 mm) apart. Strain calcu- prescribed displacement at each peak. Note that procedure 7
lations were performed in Vic 3D from Correlated Solutions could not be completed in full because the beam actuator hit
Inc., with Gaussian weights, a step size of 5 pixels, a strain the safety limit of 98 kip (436 kN) on its tension capacity.
window of 5 subsets, and a subset of 55 pixels to create a
virtual strain gauge of 21 pixels and a spatial resolution of The column base bolt holes were 0.125 in (3.175 mm) over-
75 pixels. The maximum sigma, a measure of uncertainty sized, so some slip occurred at the column base. Because con-
in the digital image correlation procedure, was 0.075 mil tinuous measurements could not be taken at the column base
(0.002 mm). during the test, it was not possible to actively compensate
for the slip. Thus, only procedures 5, 6, and 7, which had the
Test procedure least error due to slip, were used for the posttest analysis. The
maximum slips as a percentage of the beam LVDT readings in
The servo-controlled hydraulic actuator that simulated the tension were 13% for procedure 5, 10% for procedure 6, and
buckling-restrained brace is referred to as the brace actuator. 9% for procedure 7. In addition, the entire hysteresis could
The second actuator, referred to as the beam actuator, induced not be adjusted for the column base slip because slip mea-
frame horizontal displacement. A fraction of the maximum surements were taken only at the peaks with a mechanical dial
displacement of a given procedure was applied with the beam gauge. For this reason, Fig. 10 shows the prescribed hysteresis
actuator first, and then the brace actuator was adjusted to the and the peak points adjusted for the maximum slip observed
corresponding brace force of each procedure. These steps at the column base at these peak hold points.
were repeated until two stair-stepped versions of each of the
seven procedures had been completed (Fig. 9). The specimen The displacements at the beam were also calculated from
was brought back to its zero displacement and force position interpolations of the column curvature from the string poten-
at the end of each cycle of each procedure. tiometer readings for procedures 4, 5, 6, and 7. The displace-
ments measured during procedures 1, 2, and 3 were close to
The test data collected were a mixture of continuous and or below the minimum increment that could be measured by
discrete measurements. Continuous measurements were the string potentiometers and could not be reliably used for
taken at a rate of 1 Hz for all strain gauges and string poten- calculating slip compensation during the test. The displace-
tiometers 6 through 10. Continuous digital image correla- ments interpolated from the string potentiometer readings
tion measurements were taken at a rate of 1 Hz (procedures had an average 12% error (compared with the beam actuator
1, 3, 5, and 6) and 0.25 Hz (procedure 4); digital image LVDT measurements for cycles 5, 6, and 7) and a maximum

Figure 10. Prescribed hysteresis compared with the load and adjusted frame displacement at the cycle peaks during procedures
1 through 7 (P1-P7). Note: 1 in. = 25.4 mm; 1 kip = 4.448 kN.
of 15% error. Because procedures 5, 6, and 7 were the most for procedure 7, two cycles of each procedure were executed
relevant cycles, they were used to examine the gusset plate and achieved behavior near the desired hysteresis unadjusted
strains and force distributions. for slip (Fig. 9). Because the distribution of force between
the members did not follow the pinned design assumption,
Out-of-plane motion the beam actuator hit the safety limit on its tension capacity
(098 kip [436 kN]), which was just a bit lower than peak load
Data from the string potentiometers and dial gauges helped and displacement in tension for procedure 7.
quantify the incidental out-of-plane motion and were com-
pared with the out-of-plane motion detected by the digital Gusset plate strains and interface forces
image correlation at the gusset plate. The maximum out-of-
plane motions in compression recorded by dial gauge 8 (DG8) Virtual extensometers were added in the post-processed
on the column and by digital image correlation on the gusset digital image correlation data at 16 locations along the beam
plate were 0.011 and 0.031 in. (0.279 and 0.787 mm), respec- interface and 16 locations along the column interface, for a
tively. The maximum out-of-plane motions in tension record- total of 96 virtual extensometers. The virtual extensometers
ed by DG8 on the column and by digital image correlation on were placed approximately 1 in. (25.4 mm) from the beam or
the gusset plate were both 0.022 in. (0.559 mm). column face to allow for direct comparison with the values
measured by the physical strain gauges. The virtual extensom-
Forces in the beam actuator eters were 0.4 in. (10.2 mm) long and measured the engineer-
ing strain along the gauge. Because the strains measured on
Although the precast concrete BRBF was assumed to be fully the gusset plate remained elastic throughout the duration of
pinned for design purposes, the experimental results did not the test, the experimental stresses were determined along the
match this assumption. The difference between the measured gusset’s connected edges using the elastic and shear moduli
beam force and the theoretical (fully pinned) force in- (E and G, respectively) of the plate and Hooke’s law. Out-of-
creased as the magnitude of prescribed brace force increased plane effects were neglected when calculating plate stresses
(Fig. 11). This deviation begins to be particularly significant because the magnitude of out-of-plane motion of the speci-
in procedure 5 (maximum brace load of 70.4 kip [313 kN]), men was small in general and similar between the tension and
with a frame displacement corresponding to 0.5Δh2%. Except compression cycles.

Figure 11. Brace internal force versus beam internal force, both theoretical (fully pinned) and experimental. Note: 1 kip = 4.448 kN.
The strains measured by digital image correlation for pro- the column connected interface for procedure 5 cycle 1.
cedures 1 through 4 were near the noise floor of the digital Figure 14 shows the strain gauge and virtual extensometer
image correlation system and not used in the posttest analy- normal and shear stresses along the beam connected interface
sis. The strains measured by the digital image correlation for for procedure 5 cycle 1.
procedures 5 and 6 were sufficiently greater than the experi-
mental noise level and were therefore more meaningful. In ad- Trends are included in Fig. 13 and 14. A linear trend was
dition, procedures 5, 6, and 7 had significantly less influence chosen for the normal stresses, and a second-order polynomial
from the slip at the base of the column. The five virtual ex- trend was selected for the shear stresses. The average root
tensometers along the beam interface that were closest to the mean square deviations between the trends fit to the strain
free edge of the gusset plate were in an area of heavy shadow, gauge data and the virtual extensometer data were 3.40 and
which caused higher values of sigma, a measure uncertainty, 2.83, respectively, for the normal stress trends and the shear
in that region.34 Results from these five virtual extensometers stress trends.
were discarded in the analysis. Figure 12 shows the location
of these five virtual extensometers and a map of sigma across Contrary to what would be expected from the UFM, the
the entire gusset plate for procedure 5 cycle 1, which are also experimental normal stress trends were not constant along
representative of procedures 5 and 6. the beam- and column-connected edges. This finding indi-
cates that some moment developed along both the beam and
As a second source of data for the gusset plate strains and column interfaces. The assumption that only the column-con-
interface forces, data from the strain gauges on the backside nected interface would see a moment because of the gap at
of the gusset plate were compared with the results of the the beam-corbel bearing is likely incorrect, as shown by the
digital image correlation. Figure 13 shows the strain gauge linear but not constant normal stress trends in Fig. 13 and 14.
and virtual extensometer normal and shear stresses along Although this gap likely contributes to the development of

Figure 12. Map of sigma, a representation of uncertainty, across the gusset plate for procedure 5 cycle 1 and the location of the
five virtual extensometers that were discarded for analysis.
moments along these interfaces, it cannot be isolated as the τb(x) = experimental normal stress distribution along beam
only cause. connected interface
Experimental interface forces for procedure 5 cycle 1 and Vc = t g lc

0 c ( y ) dy (3)
procedure 6 cycle 1 were determined using the trends of the
strain gauge and virtual extensometer data while the brace H c = tg lc
0 c ( y ) dy (4)
was in tension. The integral under each of these trends over
the length of the connected edge was taken and multiplied where
by the thickness of the gusset plate to determine the experi-
mental interface forces. This process is described in Eq. (1) σc(y) = experimental normal stress distribution along col-
through (4). umn connected interface
Vb = t g lb
0 b ( x ) dx (1) The resulting experimental interface forces as determined
from strain gauge and virtual extensometer data and their
where percentage differences for procedures 5 and 6 are presented
in Table 1. There was less than 20% difference between the
tg = thickness of the gusset plate values determined for Hb and Vc and less than 100% differ-
ence for Vb and Hc. The experimental values of Hc, as calcu-
σb(x) = experimental normal stress distribution along lated from the strain gauge data, were consistently positive
beam-connected interface and increased throughout the cycles, as would be expected,
whereas the sign of these same values as calculated from the
H b = tg lb
0 b ( x ) dx (2) virtual extensometer data flipped sign between procedure 5
and procedure 6. Because of the agreement between the strain
where gauge and virtual extensometer trends and the expected signs

Figure 13. Stress along the column connected interface. Note: SG = strain gauge; VX = virtual extensometer; σxx = experimental
normal stress on beam-connected interface; τxy = experimental shear stress on beam- and column-connected interfaces. 1 in. =
25.4 mm; 1 ksi = 6.895 MPa.
and increase of the strain gauge data, it was deemed accept- forces, generally had a negative percent error, and Hb and Vc,
able to determine the experimental interface forces for all the shear interface forces, generally had a positive percent
procedures using only this data set and the same process from error. This finding suggests that, like similar fixed-frame tests
Eq. (1) through (4). from the literature21,22,25 where force due to frame action was
subtractive to interface normal force and additive to interface
Table 2 displays the theoretical interface forces (as derived shear force, this system was significantly affected by frame
from the UFM), experimental interface forces determined action, although its connections were not fully fixed.
by integrating under the strain gauge trend, the percent error
between the two, and the sign of this error for procedures 5, The theoretical ratios of Vc/Vb and Hc/Hb were 0.77 and 2.43,
6, and 7 at the brace’s peak tension load. For procedures 5 respectively, for all procedures. The average experimental
and 6, the values from the procedure’s first and second cycles ratios (taken over procedures 5, 6, and 7) of Vc/Vb and Hc/Hb
were averaged. An average for procedure 7 was not necessary were 4.69 and -0.39, respectively. The minimum experimental
because only one cycle was completed. The error in proce- ratio of Vc/Vb, 3.01, occurred in procedure 5 and the minimum
dures 1 through 4 was not included because it was established experimental ratio of Hc/Hb, -0.48, occurred in procedure 7.
that the slip at the column base affected the results of these The maximum experimental ratio of Vc/Vb, 7.29, occurred in
procedures more substantially. procedure 6 and the maximum experimental ratio of Hc/Hb,
-0.29, occurred in procedure 5. According to the UFM, Vc and
The percent error between the theoretical predictions of the Vb should act in the same direction and Hc and Hb should act
UFM and experimental values for all interface forces was in the same direction. In the experiment, negative ratios of Hc/
generally high (>51%) in all procedures. Generally, the Hb occurred when this was not the case. Generally, the magni-
percent error increased in magnitude as the procedure number tude of the ratios of Vc/Vb and Hc/Hb increased with increasing
increased (that is, as a larger brace force was applied to the brace load and frame displacement; this finding implies that
system). Through all cycles, the beam had higher forces along the column saw more load than the beam increasingly as the
its longitudinal axis (shear) than the column (Hb versus Vc). procedure number increased.
The beam initially had larger normal interface forces than
the column, but by the latter procedures, the column had Strains in the reinforcement
larger normal forces than the beam (Vb versus Hc). The most
important observation is that Vb and Hc, the normal interface Of the eight linear gauges attached to the corbel primary

Table 1. Experimental interface forces as determined
from strain gauge and virtual extensometer data and
their percent difference for procedures 5 and 6
Interface force
Procedure Parameter
Vb Hb Vc Hc
SG trend,
-10.23 -46.61 -30.78 13.38
kip
5 VX trend,
-5.71 -38.80 -29.08 -10.42
kip
% difference -57 -18 -6 n/a
SG trend,
-5.22 -62.84 -38.04 25.11
kip
6 VX trend,
-1.85 -69.43 -33.64 15.67
kip
% difference -95 -10 -12 46
Note: Hb = required shear force on the gusset-to-beam connection; Hc

= required axial force on the gusset-to-column connection; n/a = not
applicable (this percent difference was not calculated due to the unex-
pected sign difference between the strain gauge and virtual extensom-
eter data); SG = strain guage; Vb = required axial force on the gusset-
to-beam connection; Vc = required shear force on the gusset-to-column
connection; VX = virtual extensometer. 1 kip = 4.448 kN.
only effectively measured tension or compression in this

direction. As discussed previously, the corbel interface
axial loads were generally lower than expected from
the UFM force distribution, which would reduce bar
longitudinal strains. Larger shear interface forces than
Figure 14. Stress along the beam connected interface. Note:
SG = strain gauge; VX = virtual extensometer; σyy = experi-
predicted from the UFM were observed on the corbel, but
mental normal stress on column-connected interface; τxy = these forces are primarily carried in the noninstrumented
experimental shear stress on beam- and column-connected stirrups or through dowel action in the longitudinal bars,
interfaces. 1 in. = 25.4 mm; 1 ksi = 6.895 MPa which would have been difficult to observe using the
strain gauges as oriented in this test.
reinforcing, two gauges on the top row of reinforcing and two
gauges on the bottom row of reinforcing were responsive. Variation of force distribution
All measured strains on the corbel reinforcing remained in with increasing brace load
the elastic range, with the largest strain recorded at any bar
equal to 6% of yield strain (0.00207). Strains in the gauged One possible cause of the increase in percent error between
corbel reinforcing bars were significantly lower than what was the theoretical and experimental beam force and the increase
expected. There are two likely causes for these very low strain in magnitude of Vc/Vb and Hc/Hb ratios would be an increase
gauge readings: in column base fixity with increasing brace load. A change in
column fixity is a reasonable explanation, given the column
• First, eight no. 6 (19M) corbel primary tension reinforc- connection was not an idealized true pin, but reflective of
ing bars and six no. 5 (16M) supplemental bars were typical “pinned” precast concrete connections.
provided to prevent undesirable plate bending at the
corbel embedded plate. Although it was assumed only the Another possibility is that the interface force distribution
eight no. 6 bars would carry interface force to the corbel changed with the change in relative damage between the beam
in design, it is likely that the supplemental six no. 5 bars and column. However, the relative amount of cracking ob-
also carried some of the interface forces. served in the beam compared with the column did not follow
a consistent trend through the procedures. The damage was
• Second, the strain gauges were adhered along the longi- initially similar, then occurred more on the beam, and finally
tudinal axis of the bars, and the gauges would therefore approached more similar crack spacings and locations.12

Table 2. Theoretical interface forces, experimental interface forces, percent errors, and error signs for proce-
dures 5, 6, and 7 at the brace’s peak tension load
Interface force
Procedure Parameter
Vb Hb Vc Hc
Theoretical, kip -21.16 -17.38 -16.24 -42.31
Experimental, kip -10.23 -46.61 -30.78 13.38

5
% error -51.6 168.2 89.5 -131.6
Error sign − + + −
Theoretical, kip -22.21 -18.25 -17.06 -44.42

6
% error -76.5 244.4 123.0 -156.5
Theoretical, kip -22.74 -18.68 -17.46 -45.47

7
% error -51.1 265.2 139.4 -172.6
Note: The values from the first and second cycles for procedures 5 and 6 were averaged; values for procedure 7 are from the one cycle completed. Hb =
required shear force on the gusset-to-beam connection; Hc = required axial force on the gusset-to-column connection; Vb = required axial force on the
gusset-to-beam connection; Vc = required shear force on the gusset-to-column connection. 1 kip = 4.448 kN.
Conclusion the capacity required by the UFM alone would likely not have
resulted in a sufficient design.
From the results presented, it is concluded that the partial
precast concrete BRBF tested was not similar enough to the Although the experiment described in this paper is a first step
idealized pinned connections from Bjorhovde and Chakrabarti18 toward codifying the use of precast concrete BRBFs, a robust
and Gross and Cheok17 that were used to verify the UFM design methodology must be informed by more than the
to assume this force distribution. During the test, a positive findings from a singular specimen. Based on the results of this
percent error was observed between the theoretical UFM and research, two possibilities are proposed for future work.
experimental shear interface forces and a negative percent
error between the theoretical UFM and experimental normal First, the most promising statically determinate but unconven-
interface forces. These findings align with Lin’s conclusions21 tional connection, the lug connected only to the corbel, could
from tests on fixed frames that shear force induced by frame be revisited. Preliminary analyses12 showed that this connec-
action is additive to interface brace force and normal force tion is viable, but deeper assessment regarding strength and
induced by frame action is subtractive. This research implies constructibility is needed before a connection can be designed
that frame action significantly affects the distribution of forces and validated through testing. If such a connection is feasi-
at the gusset interfaces, even though the tested frame has more ble, this would likely be the quickest and lowest-risk path to
flexibility than the fixed frames from the literature.20–22,25 the adoption of precast concrete BRBFs into building codes
because it would require less testing than the second proposed
Despite the force distribution varying from the UFM assumed option to fully validate.
for design, the connection was robust enough to prevent
failure. At the maximum brace force tested, the beam and Second, a more complex testing program on precast concrete
column saw 265% and 139% more shear load than was pre- frames using steel buckling-resistaned braces could be execut-
dicted by the UFM, respectively. It is important to note that ed to understand how forces distribute through this system and
a failure may have occurred at the beam interface if sources develop an entirely new design procedure. This path would have
of conservatism had not been present in the as-built speci- two objectives: first, to determine a methodology for distribution
men. It seems that typical design practices, such as using the of forces to the precast concrete member according to member
minimum recommended fillet weld size in the AISC’s Steel and connection stiffnesses and second, to quantify the effects of
Construction Manual29 and limiting the embedded bar ca- frame action on gusset plate interface force distribution in precast
pacity by the shear friction factor, µ, from ACI 318-1932 may concrete BRBFs. If this path is taken, designs that leverage stan-
permit enough robustness in the design; however, designing to dard cross sections, member sizes, and connections could be cod-

ified. This path would likely require more experimental tests and 9. Viano, J. D., and T. C. Schaeffer. 2017. “Novel Use of
time than development of the lug connection, but it would allow Buckling-Restrained Braces in Precast Concrete Frames.”
the use of common construction methods and material grades. PCI Journal 62 (5): 28–34. https://doi.org/10.15554
/pcij62.5-03.
Acknowledgments
10. Guerrero, H., T. Ji, J. Alberto Escobar, and A. Teran-
The authors would like to acknowledge the PCI, which Gilmore. 2018. “Effects of Buckling-Restrained Braces
provided funding for this project through a Daniel P. Jenny on Reinforced Concrete Precast Models Subjected to
Fellowship, along with an advisory board. Advisory board Shaking Table Excitation.” Engineering Structures 163:
members included Jared Brewe, Harry Gleich, Suzanne 294–310. https://doi.org/10.1016/j.engstruct.2018
Aultman, Brandon Ross, Jon Mohle, Kim Seeber, Kevin .02.055.
Kirkley, and Greg Force. Tom Schaeffer, Jeff Viano, Gino
Kurama, and Brandt Saxey also provided their advice and ex- 11. Oh, S., Y. C. Kurama, J. Mohle, and B. Saxey. 2021.
pertise. Metromont Corp. fabricated the precast concrete speci- “Seismic Design and Analysis of Precast Concrete
men and lent the expertise of Fanfu Fan and Corbin Martin. Buckling-Restrained Braced Frames.” PCI Journal 66
(5): 54–83. https://doi.org/10.15554/pcij66.5-03.
References
12. Kessler, H. D. 2021. “Initial Investigation of Connection
1. Christopulos, A. S. 2005. “Improved Seismic Performance Behavior for Buckling-Restrained Braces (BRBs) in
of Buckling Restrained Braced Frames.” Master of science Precast Concrete Frames.” Master of science thesis,
thesis, University of Washington, Seattle. Clemson University, Clemson, SC.
2. Black, C. J., N. Makris, and I. D. Aiken. 2004. 13. Coy, B. B. 2007. “Buckling-Restrained Braced Frame
“Component Testing, Seismic Evaluation and Connection Design and Testing.” Master of Science
Characterization of Buckling-Restrained Braces.” Journal thesis. Brigham Young University, Provo, UT. https://
of Structural Engineering 130 (6): 880–894. https://doi scholarsarchive.byu.edu/etd/992.
.org/10.1061/(ASCE)0733-9445(2004)130:6(880).
14. Applied Technology Council. 2009. Quantification
3. Tremblay , R., L. Poncet, P. Bolduc, R. Neville, and of Building Seismic Performance Factors. FEMA
R. DeVall. 2004. “Testing and Design of Buckling P695. Washington, DC: FEMA (Federal Emergency
Restrained Braces for Canadian Application.” Presented Management Agency).
at 13th World Conference on Earthquake Engineering,
Vancouver, BC, August 1–6, 2004. 15. Thornton, W. 1991. “On the Analysis and Design of
Bracing Connections.” In Proceedings, National Steel
4. Newell, J., C.-M. Uang, and G. Benzoni. 2006. Conference. Chicago, IL: AISC. https://www.aisc.org/
Subassemblage Testing of CoreBrace Buckling- globalassets/aisc/manual/15th-ed-ref-list/on-the-analysis
Restrained Braces (G Series): Final Report to CoreBrace, -and-design-of-bracing-connections.pdf.
LLC. Structural Systems Research Project Report
TR-06/01. San Diego, CA: Department of Structural 16. AISC. Manual of Steel Construction. 1992. Chicago, IL:
Engineering, University of California, San Diego. AISC.
5. AISC (American Institute of Steel Construction). 2005. 17. Gross, J. L., and G. Cheok. 1988. Experimental Study
Seismic Provisions for Structural Steel Buildings. ANSI/ of Gusseted Connections for Laterally Braced Steel
AISC 341-05. Chicago, IL: AISC. Buildings. NISTIR 88-3849. Gaithersburg, MD: National
Institute of Standards and Technology. https://www.
6. Della Corte, G., M. D’Aniello, R. Landolfo, and F. M. govinfo.gov/app/details/GOVPUB-C13-805dba97612cbf
Mazzolani. 2011. “Review of Steel Buckling-Restrained c1761701de8dab3446.
Braces.” Steel Construction 4 (2): 85–93. https://doi.org
/10.1002/stco.201110012. 18. Bjorhovde, R., and S. K. Chakrabarti. 1985. “Tests
of Full-Size Gusset Plate Connections.” Journal of
7. PCI. 2017. PCI Design Handbook: Precast and Structural Engineering 111 (3): 667–684.
Prestressed Concrete. 8th ed. Chicago, IL: PCI.
19. Chou, C.-C., J.-H. Liu, and D.-H. Pham. 2012. “Steel
8. Polat, G. 2008. “Factors Affecting the Use of Precast Buckling-Restrained Braced Frames with Single and
Concrete Systems in the United States.” Journal of Dual Corner Gusset Connections: Seismic Tests and
Construction Engineering and Management 134 (3): Analyses.” Earthquake Engineering and Structural
169–178. https://doi.org/10.1061/(ASCE)0733-9364 Dynamics 41 (7): 1137–1156. https://doi.org/10.1002
(2008)134:3(169). /eqe.1176.

20. Lin, P.-C., K.-C. Tsai, A.-C. Wu, and M.-C. Chuang. Requirements for Structural Concrete (ACI 318-19) and
2014. “Seismic Design and Test of Gusset Connections Commentary (ACI 318R-19). Farmington Hills, MI: ACI.
for Buckling-Restrained Braced Frames.” Earthquake
Engineering and Structural Dynamics 43 (4): 565–587. 33. ASTM International. 2021. Standard Specification
https://doi.org/10.1002/eqe.2360. for High-Strength Low-Alloy Columbium-Vanadium
Structural Steel. ASTM A572/A572M-21e1. West
21. Lin, P.-C., K.-C. Tsai, A.-C. Wu, M.-C. Chuang, C.-H. Li, Conshohocken, PA: ASTM International. https://doi
and K.-J. Wang. 2015. “Seismic Design and Experiment .org/10.1520/A0572_A0572M-21E01.
of Single and Coupled Corner Gusset Connections in a
Full-Scale Two-Story Buckling-Restrained Braced Frame.” 34. Sutton, M. A., J. J. Orteu, H. W. Schreier, and P. Reu. 2012.
Earthquake Engineering and Structural Dynamics 44 (13): “Introduction to Digital Image Correlation: Best Practices
2177–2198. https://doi.org/10.1002/eqe.2577. and Applications.” Experimental Techniques 36 (1): 3–4.
https://doi.org/10.1111/j.1747-1567.2011.00798.x.
22. Cui, Y., X. Xu, T. C. Becker, and W. Zhang. 2021.
“Incorporating Frame Action into Seismic Design of Notation
Gusset Plates.” Journal of Structural Engineering 147
(3). https://doi.org/10.1061/(ASCE)ST.1943-541X eb = one-half the depth of the beam
.0002959.
ec = one-half the depth of the column
23. Maheri, M. R., and S. Yazdani. 2016. “Design of Steel
Brace Connection to an RC Frame Using Uniform Force E = elastic modulus
Method.” Journal of Constructional Steel Research 116:
131–140. https://doi.org/10.1016/j.jcsr.2015.09.010. G = shear modulus
24. Maheri, M. R., and A. Hadjipour. 2003. “Experimental Hb = required shear force on the gusset-to-beam connec-
Investigation and Design of Steel Brace Connection to tion
RC Frame.” Engineering Structures 25 (13): 1707–1714.
https://doi.org/10.1016/S0141-0296(03)00162-7. Hc = required axial force on the gusset-to-column con-
nection
25. Tsai, C.-Y., K.-C. Tsai, L.-W. Chen, and A.-C. Wu. 2018.
“Seismic Performance Analysis of BRBs and Gussets in lb = length of gusset interface connected to beam
a Full-Scale 2-Story BRB-RCF Specimen.” Earthquake
Engineering and Structural Dynamics 47 (12): 2366– lc = length of gusset interface connected to column
2389. https://doi.org/10.1002/eqe.3073.
lg = length of gap between bottom of beam and top of
26. Muir, L. S. 2008. “Designing Compact Gussets with the corbel
Uniform Force Method.” AISC Engineering Journal 45
(1): 13–20. Mc = required moment on the gusset-to-column interface
27. AISC. 2016. Seismic Provisions for Structural Steel P = required axial force
Buildings. ANSI/AISC 341-16. Chicago, IL: AISC.
Pbeam = beam actuator force
28. ASCE (American Society of Civil Engineers). 2017.
Minimum Design Loads and Associated Criteria for Pbrace = brace actuator force
Buildings and Other Structures. ASCE/SEI 7-16. Reston,
VA: ASCE. r = distance between the work point and the gusset
centroid
29. AISC. 2017. Steel Construction Manual. 15th ed.
Chicago, IL: AISC. tg = thickness of the gusset plate
30. AISC. 2016. Specification for Structural Steel Buildings. Vb = required axial force on the gusset-to-beam connec-
ANSI/AISC 360-16. Chicago, IL: AISC. tion
31. Muir, L. S., and W. Thornton. 2014. Steel Design Guide Vc = required shear force on the gusset-to-column con-
29: Vertical Bracing Connections—Analysis and Design. nection
Chicago, IL: AISC.
α = distance from the face of the column to the centroid
32. ACI (American Concrete Institute). Building Code of the gusset-to-column connection

β = distance from the face of the beam to the centroid
of the gusset-to-column connection
Δh2% = 2% story drift
θ = angle between the centroid of the column and the

centroid of the brace
µ = shear friction factor from ACI 318
σb(x) = experimental normal stress distribution along

beam-connected interface
σc(y) = experimental normal stress distribution along col-

umn-connected interface
σxx = experimental normal stress on beam-connected

interface
σyy = experimental normal stress on column-connected

interface
τb(x) = experimental shear stress distribution along

beam-connected interface
τc(y) = experimental shear stress distribution along col-

umn-connected interface
τxy = experimental shear stress on beam- and col-

umn-connected interfaces

About the authors •Abstract
Design and experimentally test a partial system
under representative seismic loads.
Hannah Kessler, MS,
<Body>Mohamed K. is
Nafadi,
a PCI Body text
PhD, is member
student an assistant
andprofessor
PhD student
of • Determine the applicability of the uniform force
structural
in civil engineering
engineering at Georgia
at Assiut Keywords
method (UFM) for connection interface force
University
Institute of in
Technology
Assiut, Egypt.
in Atlanta.
He is distribution.
a former
She holdsgraduate
bachelor’s
research
and master’s Body text
assistantinincivil
degrees the Department
engineering offrom A quasi-static cyclic test was performed on a scaled,
Clemson University and
in Clemson, Review
partial system.policy
Experimental results showed that the
S.C. Her primary research
Environmental
interests are
Engineering
large- andat UFM alone does not accurately predict interface forces
full-scale structural testing
North Carolina
that informs
State
theUniversity
design and Body
for thistext
system because the method does not account
(NCSU) in Raleigh.
codification of novel building and connection systems. for frame action. Results also showed that there is
Reader
some changecomments
in column base fixity as the frame un-
Kaitlynn
Omar M. Conway,
Khalafalla, PhD,
is acompleted
graduate dergoes larger horizontal displacements. This was the
research
her PhD inandmechanical
teaching assistant
engineer- BodyU.S.
first textlaboratory test on a scaled, partial version of
and at
ing PhD
Clemson
candidate
University
in the and is this system. Development of a codified design method
Department
currently in aofpostdoctoral
Civil, Construction, would require further testing to determine an appropri-
and Environmental
position at Sandia National
Engineering at ate interface force distribution and quantify the change
NCSU.
Laboratories in Albuquerque, in column base fixity.
N.Mex. Her primary research
interests are in solid Gregory
mechanics W.and Lucier, PhD, is and
the fracture a Keywords
failure of materials. research assistant professor in the
Department of Civil, Construction, Buckling-restrained brace, connection, uniform force
and
LauraEnvironmental
Redmond, PhD, Engineering
is a PCI method.
and manager of
member and assistantthe Constructed
professor in
Facilities
the GlennLaboratory
DepartmentatofNCSU.Civil Review policy
Engineering at Clemson
Sami H. Rizkalla,
University. PhD, FPCI,
Her research interests This paper was reviewed in accordance with the
FACI, FASCE, FIIFC,
include advanced simulation FEIC, for Precast/Prestressed Concrete Institute’s peer-review
FCSCE,
structuralisdesign
Distinguished
and health process. The Precast/Prestressed Concrete Institute
Professor
monitoring, model verification, and of Civil Engineering
validation and
by test. is not responsible for statements made by authors of
Construction, director of the papers in PCI Journal. No payment is offered.
Garrett
Constructed
Pataky,Facilities
PhD, isLaboratory,
an assis-
tant
and director
professorofinthetheNational
Department Publishing details
of
Science
Mechanical
Foundation
Engineering
Center onat
Clemson
Integration University.
of CompositesHis research
into This paper appears in PCI Journal (ISSN 0887-9672)
interests include
Infrastructure experimental
at NCSU. V. 68, No. 2, March–April 2023, and can be found
solid mechanics to understand at https://doi.org/10.15554/pcij68.2-03. PCI Journal
fracture and PhD,
Paul Z. Zia, fatiguePE,mechanisms
FPCI, is a is published bimonthly by the Precast/Prestressed
of materials. University
Distinguished Concrete Institute, 8770 W. Bryn Mawr Ave., Suite
Professor Emeritus in the 1150, Chicago, IL 60631. Copyright © 2023, Precast/
Department of Civil, Construction, Prestressed Concrete Institute.
and Environmental Engineering at
Abstract NCSU. Reader comments
A design method for precast concrete buckling-re- Please address any reader comments to PCI Journal edi-
strained braced frames with
Gary J. traditional
Klein, PE, gusset plate
is executive tor-in-chief Tom Klemens at tklemens@pci.org or Precast/
connections has not yet
vicebeen codified.
president andIn addition,
senior principal Prestressed Concrete Institute, c/o PCI Journal, 8770 W.
experimental data forfor
this novel
Wiss, systemElstner
Janney, are lacking. Bryn Mawr Ave., Suite 1150, Chicago, IL 60631. J
The primary objectives of this research
Associates were to do the
Inc. in Northbrook, Ill.
following:

Investigation of repair techniques
for deteriorated end regions
of prestressed concrete bridge girders
R
eliable, cost-effective repair solutions for bridge
structures are essential for transportation agencies
managing aging infrastructure or bridges experienc-
ing premature deterioration. When damage is concentrated
within localized regions, repair options that restore the per-
formance of deteriorated components are attractive because
they are less disruptive and less expensive than the replace-
ment of entire bridge components. Furthermore, repair
solutions that minimize or eliminate road closures enhance
the safety of the traveling public.
■ Deterioration of the end regions of prestressed In areas with relatively harsh winter conditions, the end
concrete bridge girders is commonly observed in regions of prestressed concrete superstructure girders are
the field when girders are exposed to chloride-laden an example of localized regions of bridge components that
water that has leaked through failed expansion joints. often experience premature deterioration.1–8 If chloride-
laden water forms through the use of deicing salt and leaks
■ Reliable repair techniques can provide a means to ex- through failed expansion joints in the deck or between the
tend girder service life, avoiding the need for immedi- deck and approach slab, the end regions of the girders below
ate superstructure replacement. Three repair systems the joints are then exposed to the chloride-laden water, lead-
were evaluated, but only the externally bonded FRP ing to a corrosive environment in which concrete spalling
system successfully restored both the strength and and reinforcement section loss occur (Fig. 1).1,2,6–8 Cycles of
initial stiffness of the girder. freezing and thawing can exacerbate deterioration.1,6,7 Given
the frequency with which bridges with end region deteriora-
■ The tests also demonstrate that end region deterio- tion are observed and the fact that the deterioration is often
ration can cause significant reductions in strength, localized to the end region of the girders, a repair technique
underscoring the importance of addressing such is needed that can effectively extend the service life of these
deterioration observed in the field. bridges, avoiding the need for superstructure replacement.

Deterioration of bottom flange Deterioration within web
Figure 1. Prestressed concrete bridge girders with end region deterioration.
Previous research on repair methods on the side faces of the girder oriented 90 degrees relative to
the longitudinal axis of the component combined with longi-
Few studies have evaluated potential methods for the repair tudinal CFRP strips used to anchor the U wraps was the most
of deteriorated end regions of prestressed concrete girders to effective because of its “increases in shear, consistency, ease
restore the strength and stiffness of the components. The pri- of application, and simplicity of design.”5
mary techniques that researchers have explored can be divided
into two categories: applying fiber-reinforced-polymer (FRP) Ramseyer and Kang6 examined the effectiveness of externally
systems and fabricating a concrete end block that encases the bonded glass-fiber-reinforced polymer (GFRP) and CFRP for
damaged region. repairing AASHTO Type II prestressed concrete girders that
had been damaged in the laboratory by failing the girder ends
FRP systems in shear prior to repair. After rapid-set cement was added to re-
place concrete missing because of the shear failures, FRP was
FRP systems have a high strength-to-weight ratio, are natural- installed in a U-wrap configuration. Cracks were also injected
ly corrosion resistant, come in a variety of materials (carbon, with epoxy on select beam ends. The only repair system that
glass, and aramid), and have high installation flexibility.9–12 successfully restored the shear strength of the end region was
Two common applications of FRP for the repair and strength- GFRP with epoxy-injected cracks; however, the authors con-
ening of structural concrete components are externally bonded cluded that the CFRP recovered more stiffness than the GFRP.
and near-surface-mounted (NSM) systems. Externally bonded
FRP consists of fibers and resin combined to form a laminate Andrawes et al.1 also evaluated end region repairs using
that is applied to the surface of a concrete component with an FRP on AASHTO Type II prestressed concrete girders. The
adhesive. NSM reinforcement consists of FRP bars or strips specimens were damaged by mechanically removing the
installed in grooves cut into the surface of a concrete compo- concrete cover within the end regions of the components. For
nent.9 Although the use of FRP as a repair and strengthening a full-scale girder receiving an externally bonded FRP repair,
system has been widely studied, relatively few researchers rapid-set mortar was first used to restore the girder cross sec-
have explored the use of FRP systems specifically for repair- tion. Vertically oriented CFRP sheets were then applied to the
ing deteriorated end regions of bridge girders.1,5,6,13 sides of the girder, and longitudinal strips were used to anchor
these vertical sheets, a configuration similar to that studied by
Petty et al.5 studied the effectiveness of different externally Petty et al.5 The repair system resulted in the shear capacity
bonded FRP configurations on deteriorated end regions of and ductility of the specimen exceeding those of an undam-
American Association of State and Highway Transportation aged control girder; however, a specimen repaired with mortar
Officials (AASHTO) Type II prestressed concrete girders alone did not restore the strength or ductility to that of the
that were salvaged from two bridges. After evaluating five control specimen. Andrawes and colleagues1 also evaluated an
carbon-fiber-reinforced polymer (CFRP) repair configurations NSM FRP repair system in small-scale girders, but the system
through shear tests on repaired girders, the authors concluded was unsuccessful in restoring the shear capacity to that of an
that the configuration consisting of U-wrap sheets with fibers undamaged specimen.

Concrete end block for developing repair techniques that can be implemented in
the field.
The fabrication of a concrete end block within the damaged
end region is another technique that has been implemented Experimental program
to repair prestressed concrete bridge girders with end region
deterioration.2–4,8 End block repairs increase the size of the Specimen overview
original cross section of the girder and rely on supplemental
mild reinforcement or concrete anchors to transfer stresses For the experimental program, five decommissioned
from the original cross section into the concrete end block. AASHTO Type I precast, prestressed concrete bridge gird-
ers were salvaged from a bridge located on Interstate 469
Needham3,4 described the development and implementation of near Fort Wayne, Ind., that was constructed in 1988. Many
an end block repair procedure for the deteriorated end regions of the bridge girders showed signs of significant end region
of prestressed concrete I-beams of an in-service bridge. A deterioration (Fig. 1). Given the condition of the girders, the
latex-modified concrete was used to form the end block. bridge superstructure was replaced in 2018, at which time
Although several problems arose during the girder repairs that five girders were extracted from the bridge for testing in the
were performed in the field, the repair procedure was consid- laboratory.
ered a cost-effective method for extending the service life of
deteriorated girders.4 Figure 2 shows details of the test specimens. Each specimen
was 38.5 ft (11.7 m) long. The AASHTO Type I girders were
Shield and Bergson8 examined the performance of end blocks composite with the original reinforced concrete deck, which
formed using shotcrete that were added to in-service pre- had a nominal thickness of 8 in. (200 mm). A thin epoxy
stressed concrete I-shaped girders with significant end region overlay had been applied to the top surface of the deck. As
deterioration. Approximately 3.5 years after the end blocks specified in the bridge plans,14 the girders were prestressed
were added, the repaired girders were removed from the with eight 0.5 in. (12.7 mm) diameter seven-wire prestress-
bridge and load tested in shear. They failed at slightly greater ing strands with an ultimate tensile strength fpu of 270 ksi
loads than undamaged companion specimens extracted from (1860 MPa). Four of the eight prestressing strands were
the same bridge, demonstrating the success of the repair. straight and located 2 in. (50 mm) from the bottom surface of
the beam. The remaining four strands were harped with harp-
Floyd et al.2 investigated end region repair using ultra-high- ing points located at one-third of the girder length from each
performance concrete (UHPC), fiber-reinforced self-consoli- end. All prestressing strands were initially stressed to 189 ksi
dating concrete (SCC), and magnesium-alumino-liquid-phos- (1300 MPa), or 0.7fpu. Figure 2 shows the spacing used for the
phate concrete. Six specimens were loaded to failure in shear stirrups, which were fabricated from no. 4 (13M) deformed
and then repaired using one of the specialty concrete types to reinforcing bars. The specified 28-day concrete compressive
cast a thickened region at the ends of the damaged prestressed strength f c′ was 5000 psi (34.5 MPa).14
girders. Shear tests on the repaired end regions resulted in
greater strengths than the capacities achieved during the initial When extracting the girders from the bridge, longitudinal cuts
tests. Despite having a smaller thickness, the UHPC repair were made approximately 2 in. (50 mm) from the edge of the
provided the greatest increase in strength compared with the top flange (Fig. 2). The portion of the deck that remained on
other two materials. the girder was kept in place for the experimental program.
A transverse edge beam had been cast monolithically with
Research scope and significance the deck and extended across the width of the bridge through
the 6 in. (150 mm) notch in the elevation (Fig. 2). A portion
While the results of past research investigating potential end of this edge beam remained on all test girders except one.
region repair methods have been promising, studies on girders A gray patch material was applied to the deteriorated end of
with deterioration from decades of service in the field are
limited, and no study has directly compared FRP and end Table 1. Specimen details
block repair procedures. To compare different repair methods Girder End region condition Repair technique
and identify key design considerations for end region repair,
an experimental program was conducted on prestressed con- C Good Control
crete girders obtained from a decommissioned superstructure. Tested in damaged
More specifically, three repair techniques were investigated: D Deteriorated
condition
an externally bonded FRP system, an NSM FRP system, and
a concrete end block. Shear tests were performed to evaluate R-EXT Deteriorated Externally bonded FRP
the performance of each repair. The inclusion of three dif-
R-NSM Deteriorated NSM FRP
ferent techniques within the same study allowed the relative
effectiveness of each method to be established. The evaluation R-BLK Deteriorated End block
of each technique applied to 30-year-old, full-scale girders
Note: FRP = fiber-reinforced polymer; NSM = near-surface-mounted.
with severe end region deterioration offers valuable insights

~2 in.
(deck)
8 in.
Varies
(1/2 in. at midspan)
4 in.
12 in.
3 in. 0.5-in. diameter

strands (draped)
11 in.
6 in.
28 in.
No. 4 stirrup
0.5-in. diameter strands
5 in.
2 in.
5 in.
16 in. 0.75 in. chamfer
Cross
Cross sectionof
section of girder
girder specimens
specimens
[INDOT 1987]
Figure 2. Girder details. Note: Note: 1 in. = 25.4 mm; 1 ft = 0.305 m.
Elevation of girder specimens

(deck not shown)
Edge beam at ends of girders
Figure 2. Girder details. Source: drawings adapted from Indiana Department of Highways, 1987. Note: 1 in. = 25.4 mm; 1 ft =
0.305 m.
some bridge girders while in service (Fig. 1). This measure is performance of a deteriorated girder in its field condition. The
assumed to have been performed to mitigate corrosion. three remaining test specimens, girders R-EXT, R-NSM, and
R-BLK (where R is repaired, EXT is externally bonded FRP
Table 1 presents the test matrix for the five girder speci- system, NSM is NSM FRP system, and BLK is concrete end
mens. One of the five test girders, girder C (control), had block), were repaired using three techniques (Table 1). For
an end region in good condition and was used as a control the externally bonded FRP and NSM FRP repairs, the girder
specimen. The other four girders exhibited severe end region cross section within the end region was restored using mortar
deterioration. Girder D (damaged) was tested to evaluate the before the FRP was applied.

Test setup and procedure ometers were placed to measure displacement at each side of
both bearing pads to capture deflections at the supports.
Figure 3 shows the test setup for the experimental program.
Each girder was loaded in shear with a point load applied Each test specimen was loaded monotonically to failure.
45 in. (1140 mm) from the centerline of the support located Failure of the test specimens was defined by either a sudden
at the end of the girder. The relatively short shear span was loss in load-carrying capacity or when the applied shear had
selected based on the observation that direct compressive decreased by at least 20 kip (89 kN) from its maximum value.
stresses transferred from the load to the support would be crit-
ical for the end regions. The original elastomeric bearing pads Repair systems
acquired from the bridge were used to support the specimens.
Because the top surface of the deck was sloped, gypsum Externally bonded FRP system For the externally
cement was used to cast a thin wedge-shaped block at the load bonded FRP repair system applied to girder R-EXT, car-
point to provide a level loading surface. A steel loading plate bon-fiber fabric was selected because it offers both a high
was placed on top of the gypsum wedge. A 5 ft (1.5 m) over- ultimate tensile strength and a high elastic modulus and has
hang was provided at the right end of the girder (Fig. 3) to been demonstrated to be an effective material for the repair
ensure that this end of the girder remained unloaded, allowing and strengthening of concrete structures.1,5,9,15 Given the
the overhanging portion to be tested at a future date if needed. bond-critical nature of the externally bonded laminate within
the end region of the girder, special consideration was given
A hydraulic cylinder applied load to the specimens, and the to the proper anchorage of the FRP reinforcement. FRP spike
load was measured using a 300 kip (1330 kN) capacity load anchors (also known as FRP fan anchors) were used in the
cell that was installed in line with the hydraulic cylinder. At design of the repair system. A spike anchor consists of a bun-
the load point and at midspan, linear string potentiometers dle of FRP fibers that are inserted into an anchor hole in the
were used to measure deflections. Additional linear potenti- concrete substrate and bonded to the primary FRP reinforce-
45 in.
12 in. x 8 in. x 2 in.
Loading plate
~3 ft
14 in. x 7 in. x 2.5 in. 14 in. x 7 in. x 2.5 in.

Bearing pad 33 ft Bearing pad 5 ft
6 in.
38.5 ft
Figure 3. Loading configuration. Note: 1 in. = 25.4 mm; 1 ft = 0.305 m.
Figure 3. Loading configuration. Note: 1 in. = 25.4 mm; 1 ft = 0.305 m.
Anchor hole
Anchor prior to saturation with epoxy
Installed anchor
Figure 4. Fiber-reinforced-polymer spike anchor.

ment (Fig. 4). FRP spike anchors have been shown to provide a ing FRP spike anchors. The longitudinal strips applied to the
reliable means of anchoring externally bonded FRP laminates sloped surface of the bottom flange were two discrete strips
when it is not possible to wrap the FRP around the entire com- on either side of the girder. These strips were not anchored
ponent. When properly detailed, FRP anchors can develop the with spike anchors. For bridge girders that contain multiple
full strength of the primary FRP laminate, allowing the frac- layers of prestressing strands in the bottom flange, drilling
ture capacity of the laminate to be reached.15,16 anchor holes perpendicular to the sloped surface presents a
high risk of hitting prestressing stands.
The externally bonded CFRP fabric was combined with an
epoxy to form the laminate that was adhered to the concrete The second FRP layer with fibers oriented vertically on the
surface. The FRP spike anchors were cut from a premanufac- side surfaces of the girder aided with the anchorage of the
tured CFRP rope produced by the same manufacturer as the longitudinal strips of the first layer and added tensile strength
primary FRP reinforcement and were combined with the same in the vertical direction. The two sheets located farthest from
epoxy as the CFRP fabric. Table 2 provides the applicable the end of the girder were installed in a U-wrap configura-
design properties of the externally bonded FRP reinforcement tion; however, the sheet closest to the end of the girder was
(FRP strips and sheets) and anchors, both in the form of a a face-bonded sheet that did not wrap around the bottom of
cured laminate (except as noted for ultimate rupture strain εfu* the girder due to the support bearing that would be present
of the FRP anchor). during in-field installations. Following the recommendations
of the manufacturer and research conducted by Andrawes et
Figure 5 shows details of the externally bonded FRP repair sys- al.,1 a minimum 1 in. (25 mm) space was provided between
tem. The repair system was composed of three layers of FRP. all externally bonded FRP sheets. The 10 in. (250 mm) width
The first layer consisted of FRP sheets that were cut into strips of the U-wrap sheets was selected based on practical limits
and applied with the fibers running parallel to the longitudinal and the results of tests conducted by Pudleiner17 on specimens
axis of the girder. The second layer consisted of FRP sheets with 10 in. wide sheets anchored with two spike anchors.
with fibers oriented vertically on the side surfaces of the girder. Because the layout of the internal steel at the end of the girder
Spike anchors were used to anchor the longitudinal strips and controlled the locations of the spike anchors, the width of the
vertical sheets. The third layer of FRP consisted of externally face-bonded sheet was reduced to 5 in. (130 mm), and the
bonded FRP patches installed over the FRP anchors. space between the face-bonded sheet and the adjacent U-wrap
sheet was increased to 2.25 in. (57.2 mm). FRP spike anchors
Longitudinal FRP strips were installed as the first layer be- were provided near the ends of each sheet and at the reentrant
cause of the importance of restoring the tensile capacity of the corner between the bottom flange and the web.
girder caused by deterioration of the prestressing strands, es-
pecially along the bottom flange. This first layer consisted of The order with which the first two layers of FRP were
three strips (Fig. 5). A continuous strip that wrapped around applied in this experimental program was different from
the end of the girder was applied both to the vertical surface the order used for systems investigated by Andrawes et al.1
of the bottom flange and to the girder web. Continuous strips and Petty et al.5 in which longitudinal strips were used to
were used to improve the anchorage of the strips and provide anchor vertical sheets. The reason for deviating from these
confinement to the mortar used to restore the girder cross schemes was the importance placed in this study on provid-
section within the end region. Continuous strips are possible ing longitudinal tensile capacity in the bottom flange. Given
because space is typically available behind the girders in the this priority, placing longitudinal strips in the first layer was
field. These longitudinal strips were anchored at the ends us- deemed to be important.
Table 2. FRP repair system components and properties
FRP repair Constituent Nominal ply Cross-sectional

Component ffu*, ksi εfu* Ef, ksi
system materials thickness, in. area, in.2
Externally bonded
FRP fabric and
FRP strips and 0.04 n/a 160.9 0.0145 10,390
Externally bonded epoxy
sheets (laminate)
carbon FRP
FRP anchors FRP rope and
n/a 0.1† 304 0.016‡ 33,300
(laminate) epoxy
NSM carbon FRP NSM strip FRP tape n/a 0.049 325 0.0181 18,000
Note: Ef = tensile modulus of elasticity of FRP reinforcement; ffu* = ultimate tensile strength of FRP reinforcement; FRP = fiber-reinforced polymer;
n/a = not applicable; NSM = near-surface-mounted. εfu* = ultimate rupture strain of FRP reinforcement. 1 in. = 25.4 mm; 1 in.2 = 645 mm2; 1 ksi = 6.895 MPa.
†
Based on single rope segment; assumes 50% fiber content.
‡
Based on dry fibers, not the cured laminate.

First layer (longitudinal strips)
Second layer (vertical sheets)
Third layer (patches)
Figure 5. Externally bonded fiber-reinforced polymer (FRP) details for girder R-EXT. Note: 1” = 1 in. = 25.4 mm.<N

Nineteen FRP spike anchors were used on each side of the ers of patches were placed over the anchors. The fibers of the
girder to anchor the externally bonded strips and sheets first layer were orientated perpendicular to the fibers of the
(Fig. 5). The design of the anchors was based on recommen- externally bonded strip or sheet, and the fibers of the second
dations developed by previous researchers.16–19 The anchor layer were orientated parallel to the fibers of the externally
holes in the bottom flange of the girder were placed such bonded strip or sheet.
that the holes would be positioned between the first and sec-
ond rows of prestressing strands considering the typical 2 in. NSM FRP system The NSM FRP system applied to gird-
(50 mm) strand grid pattern. For these holes, an anchor hole er R-NSM was designed in an effort to reestablish tensile
depth of only 4 in. (102 mm) was selected to minimize the capacity along the bottom flange of the girder within the end
risk associated with drilling holes in the bottom flange near region that had been lost due to strand deterioration and to
prestressing strands. For anchors installed on opposite sides evaluate whether simply providing this tensile strength would
of the girder web, it was not feasible to drill separate anchor be sufficient to restore the performance of the component.
holes into each side of the 6 in. (152 mm) thick web of the Furthermore, using NSM FRP may be an attractive option
specimens. Therefore, anchor holes were drilled through because installing NSM reinforcement can be easier than
the entirety of the web and continuous anchors cut from the installing the strips, sheets, and anchors of the externally
FRP rope material were installed in the holes and fanned out bonded system. As with the externally bonded system, the
on both sides of the girder. A fan angle of 60 degrees was selected NSM strips were manufactured from carbon fibers.
selected for all spike anchors, and the splayed (fan) portion Table 2 provides design properties of the NSM strips. The
of each anchor was 6 in. (150 mm) long. The number of nominal cross-sectional dimensions of the strips were 0.079
anchors across the width of each strip and sheet, the re- × 0.63 in. (2.0 × 16 mm), and each strip had a nominal area of
quired anchor weight and area, and the anchor hole diameter 0.049 in.2 (31.67 mm2).
followed the recommendations and calculations outlined by
Pudleiner.17 Considering the area of a single FRP rope seg- Figure 6 shows details of the NSM FRP repair system used
ment (Table 2), each spike anchor had a cross-sectional area for the experimental program. The system consisted of eight
of approximately 0.31 in.2 (200 mm2) and was formed by NSM FRP strips installed in grooves that were cut within
combining FRP fibers from multiple rope segments. Rich et the bottom flange along the sides of the girder. Although
al.12 provides further information about the selected details ACI PRC-440.2R-179 suggests a groove depth of at least
of the spike anchors. 0.95 in. (24 mm) for the NSM strips, this study used a depth
of 0.875 in. (22.2 mm) to maintain a groove depth less than
The third FRP layer of the repair system consisted of exter- the clear cover of 1 in. (25 mm) that is typical for girders with
nally bonded FRP patches applied over the spike anchors confinement reinforcement enclosing pretensioned strands
(Fig. 5). Previous investigations have found that such patches in the bottom flange within the end region. The clear spacing
help transfer stresses between the FRP strips or sheets and the between the grooves and clear edge distance was also limited
spike anchors,16,17 and the patches used in this study consisted by the geometry of the girder and less than the recommenda-
of the same fabric and epoxy used for the primary strips and tions in ACI PRC-440.2R-17. These limitations, however, are
sheets. Based on the results of previous research,16,17 two lay- representative of actual field conditions.
Figure 6. Near-surface-mounted (NSM) fiber-reinforced polymer details for girder R-NSM. Note: 1” = 1 in. = 25.4 mm.

End block system The intention of the end block repair assumption that the ability to lift the girder from its original
for girder R-BLK was to provide an alternative load path for elevation would not be possible in the field and that geomet-
the most deteriorated portion of the girder, allowing load ric constraints would not allow reinforcement to pass under
to transfer to new bearings located away from the original the bottom surface of the girder.
support bearing where concrete was severely deteriorated.
Figure 7 presents details of the end block repair system. The To account for severe deterioration at the original bearing
repaired region extended 24 in. (610 mm) along the length of location, a bearing pad was not placed at this original loca-
the girder, which was the minimum length needed to repair tion during testing, simulating the complete loss of bearing
the portion of the girder that experienced significant section capacity at the original support. Instead, two bearing pads
loss. All reinforcement within the end block was Grade 60 with widths equal to half of the 14 in. (360 mm) width of the
(414 MPa) epoxy-coated bars conforming to ASTM A615.20 original bearing pad (measured transverse to the longitudinal
An epoxy anchoring gel was used to install four no. 3 (10M) axis of the girder) were placed under the end block (Fig. 7).
reinforcing bars into holes drilled through the web of the
girder, and these reinforcing bars acted as dowels to aid in the It was anticipated that properly vibrating concrete during
transfer of stresses from the web to the new concrete of the casting of the end block in the field would be difficult due to
end block. To attain the shape in Fig. 7, one end of each no. limited access at the girder ends. Therefore, SCC was used to
3 (10M) bar had to be bent after the bar was inserted through cast the end block. Table 3 presents the SCC mixture propor-
the web. The remainder of the reinforcing cage consisted tions.21,22 The SCC had a 28-day design strength of 6000 psi
of no. 4 (13M) bars. The end block was designed with the (41 MPa). A 6 in. (150 mm) clearance was left between the
Figure 7. End block details for girder R-BLK. Note: no. 3 = 10M; no. 4 = 13M; 1” = 1 in. = 25.4 mm.
Table 3. Self-consolidating concrete mixture proportions for end block
Material Details Design quantity
ASTM C15021 Type 1 cement 580 lb/yd3

Cementitious materials
ASTM C61822 Class F fly ash 145 lb/yd3
Coarse aggregate ⅜ in. pea gravel 1650 lb/yd3
Fine aggregate Natural sand 1379 lb/yd3
Water n/a 279.5 lb/yd3
High-range water-reducing admixture 72.5 oz/yd3

Admixtures
Viscosity modifier 29.0 oz/yd3
Water-cement ratio n/a 0.39
Note: n/a = not applicable. 1 in. = 25.4 mm; 1 oz/yd3 = 38.681 mL/m3; 1 lb/yd3 = 0.593 kg/m3.

top of the end block and the top surface of the precast concrete to reestablish the bearing area of the girder (Fig. 9) to permit
girder (Fig. 7) to provide sufficient space for the SCC to be testing. This was accomplished with the application of a pre-
pumped into the top of the forms during field implementation. packaged 6500 psi (45 MPa) fast-setting mortar mixture with
low-shrinkage characteristics. To minimize the extent of the
Repair procedures repair, the location of the bearing pad was shifted into the gird-
er span by 4 in. (100 mm) from its original location while the
Control specimen The control specimen (girder C) was in shear span of the girder remained at 45 in. (1140 mm) (Fig. 3).
good condition, with the only notable sign of destress being
cracking in the bottom flange near the end of the girder. No Externally bonded FRP repair During the repair pro-
repairs were made. The bearing pad was shifted 3 in. (76 mm) cedures for both girders R-EXT and R-NSM, efforts were
relative to its original location (Fig. 8) to avoid the cracks. The made to simulate conditions that would be present in the field.
shear span remained at a length of 45 in. (1140 mm) (Fig. 3). Because lifting the ends of the girders in the field would not be
possible, bearing pads were placed at their original locations
Damaged specimen Girder D was tested to evaluate the as the repairs were performed. Furthermore, a plywood board
performance of a deteriorated girder that was left unrepaired was placed approximately 2 in. (50 mm) from the end of the
after being extracted from the bridge; however, given the girders to represent a mud wall and provide another realistic
severe deterioration of girder D, minimal repairs were required constraint (Fig. 10).
Figure 10 presents the progression of the repair for gird-

er R-EXT. The repair began by using an electric chipping
hammer to remove delaminated and loose concrete from the
end region until sound concrete was reached. The regions
where concrete was removed or had previously fallen from
the specimen were then sandblasted to remove corrosion
product and mitigate microcracking caused by the chipping
hammer. Next, the original cross section of the girder was
restored by applying the same fast-setting mortar used to
reestablish the bearing area of girder D. After the mortar
cured, the surface of the concrete and mortar to which FRP
was to be applied was sandblasted to a concrete surface
profile of 3, as defined in the International Concrete Rein-
forcing Institute’s ICRI 310.2R-201323 and recommended by
ACI PRC-440.2R-17.9 Anchor holes were then drilled at the
locations in Fig. 5. In accordance with ACI PRC-440.2R-17,
the edges of the anchor holes and the corners of the girder
around which FRP strips and sheets were to be wrapped
Figure 8. Shifting bearing pad for girder C. Note: 1 in. = were rounded to a radius of 0.5 in. (13 mm) to reduce stress
25.4 mm.
concentrations in the FRP.
Original bearing
location
Prior to repair of bearing area After mortar repair of bearing area Girder D prior to testing
Figure 9. Reestablishing bearing area of girder D.

Deteriorated end region After mortar repair Completed FRP repair
Figure 10. Progression of externally bonded fiber-reinforced polymer (FRP) repair for girder R-EXT.
Before the FRP reinforcement was installed, the surface of the chor was inserted, the dowel was removed from the anchor
concrete was primed and sealed using the appropriate epoxy. hole and the end of the spike anchor extending from the
As recommended by the manufacturer of the FRP reinforce- hole was fanned out at a 60-degree angle. The anchor was
ment, a different epoxy resin was used to prepare the concrete checked to ensure that it was fully impregnated with epoxy.
than was used to saturate the FRP strips, sheets, and spike For the anchors extending through the girder web, both ends
anchors to improve the tack of the concrete surface during of the anchor were fanned out on the sides of the girder.
vertical and overhead applications. After the surface of the Last, additional epoxy was injected into the anchor holes
girder was prepared, the three layers of FRP were sequentially to eliminate possible air voids. Rich et al.12 provides the
installed (Fig. 5). Before the FRP strips and sheets were in- detailed procedure for installing the spike anchors. The FRP
stalled, they were saturated with epoxy using plastic laminat- patches were installed with the same procedure used for the
ing rollers. Once the FRP strips and sheets were in place on FRP strips and sheets. Figure 10 shows the completed exter-
the concrete surface, the rollers were used to fully impregnate nally bonded FRP repair system on girder R-EXT.
them with epoxy and eliminate air voids. Plastic squeegees
were then used to remove excess epoxy. NSM FRP repair Figure 11 shows the repair procedure for
girder R-NSM. First, unsound concrete was removed and the
Next, the spike anchors were installed. For each anchor, a surface was sandblasted. Then, the cross section of the girder
razor blade was used to separate the fibers of the FRP sheets was restored with the same mortar mixture used for girders D
installed on the girder to expose the anchor hole. A caulk and R-EXT. After the mortar cured, the NSM FRP strips were
dispensing gun was used to inject epoxy into the hole. The installed in accordance with the following procedure:
spike anchor was saturated with epoxy before installation.
Wooden dowels fastened to the anchor with zip ties were 1. Grooves with the dimensions in Fig. 6 were cut into the
used to help insert the anchors into the holes. After an an- concrete substrate using a tuckpointing grinder.
Deteriorated end region Inserting NSM strips Completed FRP repair
Figure 11. Progression of near-surface-mounted (NSM) fiber-reinforced polymer (FRP) repair for girder R-NSM.

2. Dust and debris were removed from the grooves using Fig. 7, and the end block after construction was completed.
compressed air.
Experimental results
3. The grooves were partially filled with epoxy grout.
Table 4 summarizes the test results for the five girder spec-
4. The FRP strips were inserted into the grooves using a imens and corresponding material strengths. These results
sawing motion until the strips were centered at approxi- include the following:
mately the middepth of the groove.
• the maximum shear force resisted by the specimen Vtest
5. The remainder of the grooves was filled with epoxy grout.
• the ratio of the maximum shear resisted by the specimen
6. Excess grout was removed, and the surface was leveled to the maximum shear resisted by girder C Vtest/Vcontrol
using squeegees.
• the ratio of the maximum shear resisted by the specimen
The transverse edge beam present in the bridge superstructure to the maximum shear resisted by girder D Vtest/Vdamaged
(Fig. 2) was unintentionally separated from girder R-NSM
during extraction or transportation of the component. A por- To measure the strength of the precast concrete of each girder,
tion of this edge beam remained on all of the other test girders at least three 4 × 6 in. (100 × 150 mm) concrete cores were
(compare Fig. 11 with Fig. 10). taken from the webs of each specimen following the girder
test. Table 4 provides the average compressive strength of the
End block repair Minimal surface preparation was per- concrete cores fc obtained from each girder in accordance with
formed on girder R-BLK before the end block was construct- ASTM C42.24 In addition, for specimens that received mortar
ed. Figure 12 shows the state of the specimen after loose repairs, Table 4 provides the compressive strength of the
concrete was removed. Figure 12 also shows the reinforcing mortar fm on the day of the girder tests, which was determined
cage, which was fabricated in accordance with the details in from 2 in. (50 mm) mortar cubes tested in accordance with
Deteriorated end region Completed reinforcing cage Completed end block repair
Figure 12. Progression of end block repair for girder R-BLK.
Table 4. Summary of test results
fc of cored fm of fc of end block fct of end block

Girder Vtest, kip Vtest/Vcontrol Vtest/Vdamaged
concrete, ksi mortar, ksi concrete, ksi concrete, ksi
C 7.27 n/a n/a n/a 141 1.00 1.76
D 9.24 9.13 n/a n/a 80 0.57 1.00
R-EXT 7.44 16.10 n/a n/a 189 1.34 2.36
R-NSM 9.07 12.17 n/a n/a 31 0.22 0.39
R-BLK 7.85 n/a 7.07 0.63 81 0.57 1.01
Note: fc = compressive strength of concrete; fct = splitting tensile strength of concrete; fm = compressive strength of mortar; n/a = not applicable; Vcontrol
= maximum shear resisted by girder C; Vdamaged = maximum shear resisted by girder D; Vtest = maximum shear resisted by girder. 1 kip = 4.448 kN; 1 ksi =
6.895 MPa.

Figure 13. Shear versus deflection at load point for girder specimens. Note: 1 in. = 25.4 mm; 1 kip = 4.448 kN.
ASTM C109.25 Finally, for girder R-BLK, Table 4 includes lack of development of the prestressing strands in the bottom
the compressive strength fc and splitting tensile strength fct of flange near the end of the girder. The load-carrying capacity
the concrete of the end block obtained on the day of the girder decreased as the diagonal cracks widened, and widening of
test from 4 × 8 in. (100 × 200 mm) cylinders in accordance the cracks corresponded with slippage of the prestressing
with ASTM C3926 and ASTM C496,27 respectively. Although strands in the bottom flange at the end of the girder. The ends
variations in the compressive strengths of the concrete and of the strands slipped approximately 1.25 in. (31.8 mm) into
mortar are evident among the girder specimens, these differ- the girder by the end of the test. The relatively sudden loss
ences are not considered to be significant to the overall perfor- in load-carrying capacity observed at a deflection of 1.24 in.
mance of the specimens based on the failure modes observed (31.5 mm) was likely due to strand slip.
during the tests.
Girder D
Figure 13 plots the shear within the test region (that is, the
45 in. [1140 mm] shear span) versus the deflection at the load For the damaged specimen, the primary failure crack (Fig. 14)
point for each of the specimens. The shear due to the self- initiated at the bottom of the component approximately 42 in.
weight of the girder is not reflected in the response curves or (1070 mm) from the end of the specimen at a shear of 61 kip
the values in Table 4, but it is estimated to be approximately (270 kN), propagated vertically through the web upon further
6.5 kip (29 kN) at the middle of the 45 in. (1140 mm) shear loading, and finally propagated diagonally through the top
span considering the original cross section of the girders. flange and deck toward the load point. Failure was defined by
Deflections caused by deformation of the bearing pads were an abrupt drop in the load-carrying capacity at a deflection of
small (for example, deflection at the load point due to bearing 2.43 in. (61.7 mm) (Fig. 13). By the end of the test, portions
pad deformation was 0.039 in. [0.99 mm] for girder C at Vtest) of the bottom flange on both sides of the web had detached
and are therefore neglected. from the specimen at the support (Fig. 14). This separation of
the bottom flange began early in the test due to the deteriorat-
Girder C ed state of the girder. As indicated in Table 4, the ratio of the
shear capacity of girder D to the capacity of girder C is 0.57.
Failure of the control specimen was characterized by the for-
mation of a diagonal strut within the test region, correspond- Girder R-EXT
ing to a crack angle of approximately 43 degrees measured
from horizontal (Fig. 14). Diagonal cracking developed in the Figure 13 shows that girder R-EXT exhibited an initial stiff-
web of the girder at a shear of 98 kip (440 kN). The prima- ness greater than that of girder C. As load was increased, the
ry diagonal crack did not extend to the support but instead behavior of the specimen became dominated by the opening
extended to a point at the bottom of the girder located inside of a flexural crack located at the termination of the longitudi-
the span from the bearing region. This occurred due to the nal FRP strips (approximately 49 in. [1240 mm] from the end

Control specimen (girder C) Damaged specimen (girder D)
Figure 14. Control and damaged specimens after failure.
Critical
flexural crack
Splitting
crack
Separated
bottom flange
Girder R-EXT Girder R-NSM
Side view End view
Girder R-BLK
Figure 15. Repaired girders after failure.

Figure 15. Repaired girders after failure.
of the girder). This primary flexural crack was first visually Figure 15 shows girder R-EXT after failure. The specimen
observed at a shear of 140 kip (620 kN) but may have initiated experienced a flexural failure characterized by complete
earlier; close visual examination of the specimen was not fracture of two prestressing strands in the bottom flange at
conducted at high loads because of safety concerns. During the wide critical flexural crack (Fig. 15), which was located
the test, relatively minor diagonal cracking also developed at the termination of the longitudinal FRP strips. Based on
near the load point in the region not covered by FRP rein- observations during the test, the sudden losses in load-car-
forcement; however, the portions of the diagonal cracks that rying capacity that occurred at a deflection of 1.39 in. (35.3
were visible did not widen significantly after their formation. mm) and a deflection of 2.35 in. (59.7 mm) correspond with
Propagation of the cracks toward the support is unknown the strand fractures (Fig. 13). The externally bonded FRP
because of the presence of the FRP. prevented failure from occurring within the repaired region.

Failure instead occurred outside the deteriorated region of Analysis of results
the girder.
A diagonal strut developed within the test region of the con-
Minimal FRP debonding from the concrete surface was trol girder (girder C) (Fig. 14); however, the damaged girder
observed after testing. It was confined to areas along the (girder D) exhibited behavior governed by the inability of
longitudinal strips that were not directly covered with patch- the corroded prestressing strands to develop sufficient tensile
es or vertical FRP sheets. Figure 15 outlines the debonded force along the bottom flange at the end of the component. A
areas on one side of the girder in red. No rupture of the diagonal strut could not form within the test region without
FRP occurred. adequate tensile capacity along the bottom flange to equil-
ibrate the horizontal component of the strut. The primary
Girder R-NSM failure crack of the damaged specimen was nearly vertical.
The behaviors of the control and damaged specimens clearly
At the beginning of the test on girder R-NSM, the girder’s indicate the importance of the tensile capacity along the bot-
initial stiffness was less than that of the other specimens tom flange in developing the full strength of the girder.
(Fig. 13). Relatively early in the test, when a shear of 31
kip (140 kN) was reached, the portion of the web located The test results present a distinction between the satisfac-
above the support bearing experienced a splitting crack that tory performance achieved by the specimen with externally
effectively caused the end of the specimen to separate from bonded FRP and issues for the specimens with the NSM
the rest of the girder (Fig. 15). The splitting crack appeared FRP and end block repair systems that prevented the girders
suddenly over the depth of the girder, intersecting with the from achieving restoration of the strength and stiffness of
reentrant corner at the notch located along the top flange of the damaged end regions. For girder R-NSM, the maximum
the girder, and resulted in a sudden loss in load-carrying ca- shear Vtest resisted by the specimen was only 22% of the
pacity. Once the end of the specimen that separated from the shear resisted by girder C and 39% of the shear resisted by
rest of the girder was no longer effective in transferring load girder D. Although the lack of a portion of the transverse
to the bearing, the reaction force was primarily transferred edge beam on the specimen may have negatively affect-
through a relatively small portion of the bottom flange in ed specimen performance, the splitting above the support
contact with the bearing pad. After reaching a second peak bearing demonstrates the importance of providing ade-
in the response curve, load-carrying capacity was again lost quate confinement within the region of the girder repaired
when the outer portions of the bottom flange separated from with mortar. Because the NSM strips were not effectively
the girder at a shear of 27 kip (120 kN) (Fig. 15). Because engaged, the behavior of the specimen essentially represents
of the premature failure, the NSM strips were not effective- the performance of a girder repaired only with mortar and
ly engaged. highlights the need to provide strengthening measures be-
yond simply restoring the cross section of the girder using a
As previously discussed, girder R-NSM lacked a portion of fast-setting mortar.
the transverse edge beam of the bridge superstructure that was
present on the other specimens (Fig. 2). The absence of a por- Girder R-BLK achieved a maximum shear Vtest that was only
tion of this edge beam may have contributed to the splitting 57% of the strength of girder C and essentially equal to the
observed in the vicinity of the notch during the test. strength of girder D. Although the initial stiffness of gird-
er R-BLK was similar to that of girder C, its postcracking
Girder R-BLK behavior more closely resembled the behavior of girder D.
Cracking and loss of stiffness occurred at a relatively low
The response curve of girder R-BLK presents an initial shear, causing the linear portion of the response curve to
stiffness similar to that of girder C (Fig. 13). As loading end at a shear of 44 kip (200 kN), less than half the shear
increased, the end face of the girder R-BLK end block began achieved by girder C before diagonal crack formation (98 kip
to experience minor cracking at a shear of 8.9 kip (40 kN). At [440 kN]). The formation of cracks in girder R-BLK at a low
a shear of 44 kip (200 kN), cracking had propagated vertically shear was caused by the elimination of the original center
through the full depth of the end block and extended longitu- bearing, which forced the load to be transferred through the
dinally along the bottom of the block. This cracking led to the end block to the two new bearing areas. Furthermore, the
block eventually splitting into two pieces (Fig. 15). Outside of absence of continuous reinforcement near the bottom (that
the end block, a diagonal crack initiated at a shear of approxi- is, along the tension face) of the end block transverse to the
mately 53 kip (240 kN). Diagonal cracking extended from the longitudinal axis of the girder caused the splitting of the end
bottom surface of the girder near the edge of the end block to- block to quickly increase in severity upon further loading
ward the load point, indicating the development of a diagonal due to the absence of a tension tie. Continuous reinforcement
strut. The load test was terminated after the interface between with proper development is essential for a successful end
the end block and the original girder failed (that is, the end block repair to properly transfer stresses to the bearing pads
block separated for the concrete of the original girder), and supporting the end block and to control any cracks that de-
the end block experienced significant rotation relative to the velop. The end block essentially behaved as an unreinforced
girder (Fig. 15). deep transfer beam.

The externally bonded FRP repair system of girder R-EXT longitudinal FRP strips, is needed to prevent the prema-
provided a shear capacity that exceeded the capacity of girder ture splitting failure mode observed during the test on
C by 34%. Furthermore, the linear portion of the response the specimen with NSM FRP reinforcement. Providing
curve ends at a shear that is 17% greater than the shear when confinement around the repaired region also mitigates
diagonal cracking developed in girder C. Because of the high some concerns about the condition of the concrete at the
stiffness of the FRP laminate material, the repair system was interface between the original girder concrete and repair
able to restore the stiffness lost due to the deterioration of material, as well as concerns about the resulting bond
the end region, providing a greater initial stiffness than that between the two materials.
of the control specimen. Considering that a flexural failure
outside of the repaired region was achieved, the externally • The externally bonded FRP repair system proved to be a
bonded FRP repair system was able to restore shear capacity viable technique for restoring the strength and stiffness
within the deteriorated end region, reestablish the tie force of the prestressed concrete girder with severe end region
in the bottom flange that was lost due to deterioration that deterioration. The repaired specimen achieved a great-
caused the failure behavior exhibited by girder D, and provide er shear capacity and a greater initial stiffness than the
sufficient confinement to the region repaired with mortar. control specimen, leading to a flexural failure outside of
The confinement provided by the FRP prevented any splitting the repaired region. The chosen spike anchor details were
above the support bearing as observed for girder R-NSM, successful in preventing FRP failure due to debonding.
and it precluded separation of portions of the bottom flange
of the component as experienced by girders D and R-NSM. • The NSM FRP repair system did not provide adequate
Moreover, the confinement allowed two strands to reach their confinement to the region repaired with mortar; there-
ultimate rupture strengths at the end of the repaired region and fore, the strength and stiffness of the prestressed con-
prevented the significant slip of the strands observed during crete bridge girder were not restored. If NSM strips are
the test on girder C. installed along the vertical and sloped surfaces of the
bottom flange as a strengthening measure, additional
Considering that the NSM strips in girder R-NSM were not consideration should be given to confining the portion of
effectively engaged due to premature splitting of the repaired the girder that is restored using a repair material.
region, the addition of externally bonded FRP to confine the
end region and prevent such splitting may allow NSM strips • The end block system also did not reestablish the strength
to provide tensile capacity along the bottom flange. Such a of the deteriorated end region, though the system restored
hybrid repair system that incorporates both NSM strips in the initial stiffness of the component. Elimination of the
the bottom flange and externally bonded FRP reinforcement original bearing under the girder web and the absence of
could be a viable technique for restoring the strength and stiff- continuous transverse reinforcement along the bottom of
ness of a deteriorated end region and deserves further study. the end block to aid with transferring stresses to the new
support areas resulted in premature failure. Therefore, the
Conclusion failure prevented evaluation of the potential benefit of this
approach.
The primary conclusions resulting from the experimental
investigation conducted to evaluate repair techniques for With modifications, the NSM FRP and end block repair
prestressed concrete bridge girders with severe end region systems could potentially be viable techniques to restore the
deterioration are as follows: behavior of prestressed concrete girders with end region dete-
rioration. Further research, however, is needed to evaluate the
• Deterioration of the end regions of prestressed concrete success of modified versions of these repair systems.
girders can result in significant reductions in strength
(43% or greater shear strength reduction considering Acknowledgments
results of the control and damaged specimens).
This work was supported by the Joint Transportation Research
• When designing end region repair systems for prestressed Program administered by the Indiana Department of Trans-
concrete girders, it is critical to restore the tensile capacity portation and Purdue University. The authors would like to
lost due to corroded and ineffective prestressing strands in thank the members of the Study Advisory Committee for pro-
the bottom flange of the deteriorated end regions. Without viding guidance and direction for the research. Appreciation
adequate tensile capacity in the bottom flange, a diago- is also extended to Sika Corp., Pilgrim Permacoat, and Owens
nal strut cannot form within the end region, resulting in Corning for their assistance with material acquisition for the
premature failure and decreased capacity. experimental program. The contents of this paper reflect the
views of the authors, who are responsible for the facts and
• It is also critical to ensure that confinement of the repair accuracy of the data presented herein and do not necessarily
material (for example, mortar) used to restore the cross reflect the official views or policies of the sponsoring organi-
section of the girder is adequate. End confinement, zation. These contents do not constitute a standard, specifica-
such as the confinement provided by externally bonded tion, or regulation.

References Bonded FRP Systems for Strengthening Concrete Struc-
tures. ACI PRC-440.2R-17. Farmington Hills, MI: ACI.
1. Andrawes, B., I. D. Shaw, and H. Zhao. 2018. Repair
and Strengthening of Distressed/Damaged Ends of Pre- 10. Cai, C. S, and M. Xia. 2015. Repairing/Strengthening of
stressed Beams with FRP Composites. Report FHWA- Bridges with Post-Tensioned FRP Materials and Per-
ICT-18-001. Urbana, IL: Illinois Center for Transporta- formance Evaluation. Report FHWA/LA.11/488. Baton
tion, University of Illinois at Urbana-Champaign. https:// Rouge, LA: Louisiana Transportation Research Center,
doi.org/10.36501/0197-9191/18-001. Louisiana Department of Transportation and Develop-
ment. https://rosap.ntl.bts.gov/view/dot/29399.
2. Floyd, R. W., J. S. Volz, T. Looney, M. Mesigh, M. Ah-
madi, S. Roswurm, P. Huynh, and M. Manwarren. 2021. 11. Pevey, J. M., W. B. Rich, C. S. Williams, and R. J. Frosch.
Evaluation of Ultra-High Performance Concrete, Fiber 2021. Repair and Strengthening of Bridges in Indiana Us-
Reinforced Self-Consolidating Concrete, and MALP ing Fiber Reinforced Polymer Systems: Volume 1–Review
Concrete for Prestressed Girder Repair. Report FHWA- of Current FRP Repair Systems and Application Method-
OK-21-03. Oklahoma City, OK: Oklahoma Department ologies. Report FHWA/IN/JTRP-2021/09. West Lafay-
of Transportation. https://shareok.org/handle/11244 ette, IN: Joint Transportation Research Program, Purdue
/330981. University. https://doi.org/10.5703/1288284317309.
3. Needham, D. E. 1999. Prestressed Concrete Beam End 12. Rich, W. B., R. R. Jacobs, C. S. Williams, and R. J.
Repair (Interim Report). Report R-1373. Lansing: Mich- Frosch. 2021. Repair and Strengthening of Bridges
igan Department of Transportation. https://mdotjboss. in Indiana Using Fiber Reinforced Polymer Systems:
state.mi.us/SpecProv/getDocumentById.htm?docGuid Volume 2–FRP Flexural Strengthening and End Region
=bcf5caef-cd66-4f6c-91c6-5bc1e86c1ab8. Repair Experimental Programs. Report FHWA/IN/
JTRP-2021/10. West Lafayette, IN: Joint Transportation
4. Needham, D. E. 2000. Prestressed Concrete Beam End Research Program, Purdue University. https://doi.org/10
Repair (Final Report). Report R-1380. Lansing, MI: .5703/1288284317310.
Michigan Department of Transportation. https://mdotj-
boss.state.mi.us/SpecProv/getDocumentById.htm 13. Pei, J. S., R. D. Martin, C. J. Sandburg, and T. H.-K. Kang.
?docGuid=b47813a4-cafe-4d32-b4fd-51cee10d7b9e. 2008. Rating Precast Prestressed Concrete Bridges for
Shear. Report FHWA-OK-08-08. Oklahoma City, OK:
5. Petty, D. A., P. J. Barr, P. G. Osborn, M. W. Halling, and Oklahoma Department of Transportation. https://www.odot
T. R. Brackus. 2011. “Carbon Fiber Shear Retrofit of For- .org/hqdiv/p-r-div/spr-rip/library/reports/fhwa-ok0808.pdf.
ty-Two-Year-Old AASHTO I-Shaped Girders.” Journal of
Composites for Construction 15 (5): 773–781. https://doi 14. Indiana Department of Highways. 1987. Bridge Plans
.org/10.1061/(ASCE)CC.1943-5614.0000208. for Spans over 20 Feet on State Road No. 24 (Project No.
MAF-170-1). Indianapolis, IN: Indiana Department of
6. Ramseyer, C., and T. H.-K. Kang. 2012. “Post-Damage Highways.
Repair of Prestressed Concrete Girders.” Internation-
al Journal of Concrete Structures and Materials 6 (3): 15. Ghannoum, W. M., N. K. Alotaibi, J. Garcia, C. H.
199–207. https://doi.org/10.1007/s40069-012-0019-7. Kim, Y. Kim, D. Pudleiner, K. Quinn, N. Satrom, W.
Shekarchi, W. Sun, H. Wang, and J. Jirsa. 2018. “Shear
7. Shafei, B., B. Phares, and S. Weizhuo. 2020. Beam Strengthening of Concrete Bridges Using CFRP Strips
End Repair for Prestressed Concrete Beams. Report and Anchors and Quality Control Procedures.” ACI Sym-
IHRB Project TR-715. Ames, IA: Bridge Engineering posium Publication 327: 43.1–43.20. https://doi.org/10
Center, Iowa State University. http://publications.iowa. .14359/51713364.
gov/33803/1/TR-715_Final%20Report_Beam%20
End%20Repair%20for%20Prestressed%20Concrete%20 16. Shekarchi, W. A., D. K. Pudleiner, N. K. Alotaibi, W.
Beams.pdf. M. Ghannoum, and J. O. Jirsa. 2020. “Carbon Fiber-Re-
inforced Polymer Spike Anchor Design Recommenda-
8. Shield, C., and P. Bergson. 2018. BR27568—Experimen- tions.” ACI Structural Journal 117 (6): 171–182. https://
tal Shear Capacity Comparison between Repaired and doi.org/10.14359/51728065.
Unrepaired Girder Ends. Report MN/RC 2018-07. Min-
neapolis, MN: Minnesota Department of Transportation. 17. Pudleiner, D. K. 2016. “Design Considerations Based on
http://www.dot.state.mn.us/research/reports/2018 Size Effects of Anchored Carbon Fiber Reinforced Poly-
/201807.pdf. mer (CFRP) Systems.” MS thesis, University of Texas at
Austin. http://hdl.handle.net/2152/39031.
9. American Concrete Institute (ACI) Committee 440. 2017.
Guide for the Design and Construction of Externally 18. Kim, Y. G. 2011. “Shear Behavior of Reinforced Con-

crete T-Beams Strengthened with Carbon Fiber Rein- Notation
forced Polymer (CFRP) Sheets and CFRP Anchors.” PhD
dissertation, University of Texas at Austin. http://hdl Ef = tensile modulus of elasticity of fiber-reinforced
.handle.net/2152/ETD-UT-2011-12-4398. polymer reinforcement reported by manufacturer
19. Orton, S. L. 2007. “Development of a CFRP System fc = compressive strength of concrete

to Provide Continuity in Existing Reinforced Concrete
Buildings Vulnerable to Progressive Collapse.” PhD dis- f c′ = specified 28-day concrete compressive strength
sertation, University of Texas at Austin. http://hdl.handle
.net/2152/3241. fct = splitting tensile strength of concrete
20. ASTM International. 2018. Standard Specification for ffu* = ultimate tensile strength of fiber-reinforced polymer
Deformed and Plain Carbon-Steel Bars for Concrete reinforcement reported by manufacturer
Reinforcement. ASTM A615/A615M-18. West Consho-
hocken, PA: ASTM International. https://doi.org/10.1520 fm = compressive strength of mortar
/A0615_A0615M-18.
fpu = ultimate tensile strength of prestressing strand
21. ASTM International. 2019. Standard Specification for
Portland Cement. ASTM C150/C150M-19a. West Con- Vcontrol = maximum shear force resisted by girder C
shohocken, PA: ASTM International. https://doi.org/10
.1520/C0150_C0150M-19A. Vdamaged = maximum shear force resisted by girder D
22. ASTM International. 2019. Standard Specification for Vtest = maximum shear force resisted by girder
Coal Fly Ash and Raw or Calcined Natural Pozzolan for
Use in Concrete. ASTM C618-19. West Conshohocken, εfu* = ultimate rupture strain of fiber-reinforced polymer
PA: ASTM International. https://doi.org/10.1520 reinforcement reported by manufacturer
/C0618-19.
23. ICRI (International Concrete Repair Institute) Committee

310. 2013. Selecting and Specifying Concrete Surface
Preparation for Sealers, Coatings, Polymer Overlays,
and Concrete Repair. ICRI Guideline 310.2R-2013.
Rosemont, IL: ICRI.
24. ASTM International. 2019. Standard Test Method for

Obtaining and Testing Drilled Cores and Sawed Beams of
Concrete. ASTM C42/C42M-18a. West Conshohocken,
PA: ASTM International. https://doi.org/10.1520/C0042
_C0042M-18A.

Compressive Strength of Hydraulic Cement Mortars
(Using 2-in. or [50-mm] Cube Specimens). ASTM C109/
C109M-16a. West Conshohocken, PA: ASTM Interna-
tional. https://doi.org/10.1520/C0109_C0109M-16A.

Compressive Strength of Cylindrical Concrete Specimens.
ASTM C39/C39M-18. West Conshohocken, PA: ASTM
International. https://doi.org/10.1520/C0039
_C0039M-18.

Splitting Tensile Strength of Cylindrical Concrete Speci-
mens. ASTM C496/C496M-17. West Conshohocken, PA:
ASTM International. https://doi.org/10.1520/C0496
_C0496M-17.

systems were evaluated: an externally bonded fiber-
reinforced-polymer (FRP) system, a near-surface-
William
<Body>Mohamed
B. Rich, MS, K. Nafadi,
is a structur- Body textFRP system, and a concrete end block. Only
mounted
al
PhD,
engineer
is an assistant
at Martin/Martin
professorinof the externally bonded FRP system successfully re-
Denver,
structuralColo.
engineering
He received
at Assiut
his Keywords
stored both the strength and initial stiffness of the gird-
bachelor’s
University degree
in Assiut,
in civil
Egypt.engi-
He is er. Although the other two methods were not success-
neering
a formerfrom
graduate
the University
research of Body
ful, thetext
tests on the repaired girders highlight important
Wyoming
assistant ininthe
Laramie
Department
and his of factors that must be considered when designing repairs
master’s
degree fromand Purdue
Environ- Review
or conductingpolicy
further research. The tests also demon-
University
mental Engineering
in West Lafayette,
at North Ind. strate that end region deterioration can cause signifi-
Carolina State University (NCSU) Body(>text
cant 40%) reductions in strength, underscoring the
in Raleigh. Christopher S. Williams, PhD, is importance of addressing such deterioration observed
an assistant professor of civil Reader
in the field. comments
Omar M. Khalafalla,
engineering at PurdueisUniversity.
a graduate
research
He received
andhisteaching
bachelor’s
assistant
degree Keywords
Body text
andcivil
in PhDengineering
candidate in from
the Southern
Depart-
ment ofUniversity
Illinois Civil, Construction,
Carbondale and
and Corrosion, bridge, end region repair, externally bonded
Environmental
his master’s andEngineering
PhD degreesatfrom reinforcement, fiber-reinforced polymer, FRP, girder,
NCSU.
the University of Texas at Austin. near-surface-mounted reinforcement, NSM, prestressed
concrete bridge girders, shear strengthening.
Gregory
Robert J.W. Lucier,PhD,
Frosch, PhD,PE, is aFACI,
research assistant professor
FASCE, is a professor of civil in the Review policy
Department of Civil, Construction,
engineering and vice provost for
and Environmental
academic facilities Engineering
at Purdue This paper was reviewed in accordance with the
and
University. Aof
manager the Constructed
fellow of the Precast/Prestressed Concrete Institute’s peer-review
Facilities
AmericanLaboratory at NCSU.
Concrete Institute and process. The Precast/Prestressed Concrete Institute
American Society of Civil is not responsible for statements made by authors of
Sami H. Rizkalla,
Engineers, PhD,
he is the edi-FPCI, papers in PCI Journal. No payment is offered.
FACI, FASCE, FIIFC,
tor-in-chief of the ACI Structural Journal and FEIC,serves
on the ACI 318 Structural Concrete Building Profes-
FCSCE, is Distinguished Code Publishing details
Committee, for which sorheofchairs
Civil Engineering and
ACI 318D, Structural
Members. His research,
Construction,
which focuses
director
on of
thethe
design This paper appears in PCI Journal (ISSN 0887-9672)
and behavior of structural
Constructed
concrete,
Facilities
has resulted
Laboratory,
in V. 68, No. 2, March–April 2023, and can be found
changes to the ACI building
and director
codeofand
thethe
National
American at https://doi.org/10.15554/pcij68.2-02. PCI Journal
Association of StateScience
Highway Foundation
and Transportation
Center on is published bimonthly by the Precast/Prestressed
Officials’ design specifications.
Integration of Composites into Concrete Institute, 8770 W. Bryn Mawr Ave., Suite
Infrastructure at NCSU. 1150, Chicago, IL 60631. Copyright © 2023, Precast/
Abstract Prestressed Concrete Institute.
Paul Z. Zia, PhD, PE, FPCI, is a
Deterioration of the end regions of University
Distinguished prestressedProfes-
concrete Reader comments
bridge girders is commonly observed
sor Emeritus in Department
in the the field whenof
girders are exposed to chloride-laden
Civil, water
Construction, andthat has
Environ- Please address any reader comments to PCI Journal
leaked through failedmental
expansion joints. Because
Engineering at NCSU.the editor-in-chief Tom Klemens at tklemens@pci.org or
deterioration is often localized to the end regions of the Precast/Prestressed Concrete Institute, c/o PCI Journal,
girders, reliable repair techniques can provide a means 8770 W. Bryn Mawr Ave., Suite 1150, Chicago, IL
to extend girder service
Gary life, avoiding
J. Klein, PE,the
is need for
executive 60631. J
immediate superstructure replacement.
vice president and To evaluate
senior dif-
principal
ferent repair methodsforand identify
Wiss, keyElstner
Janney, design Associ-
consider-
ations for end regionates
repair,
Inc.shear tests to failure
in Northbrook, Ill. were
conducted on prestressed concrete girders extracted
from a decommissioned superstructure. Three repair

Experimental background behind
new AASHTO LRFD specifications
for partially debonded strands
Mathew W. Bolduc, Bahram M. Shahrooz, Kent A. Harries, Richard A. Miller,

Henry G. Russell, and William A. Potter
I
n prestressed concrete bridge girder fabrication, the fol-
lowing four methods are commonly used to meet limits
on the extreme fiber concrete tensile stress at prestress
strand release:
• partial debonding of several strands near the girder end
• harping some strands
• adding top strands
• a combination of these methods
Strands that can be harped are limited to those aligned

within the member webs and may be further limited by the
casting bed. For some girder shapes, such as boxes, harping
■ To develop a unified approach for the design of is not practical. The preference for partial debonding or
partially debonded strands in prestressed concrete harping to relieve extreme-fiber tensile stress varies among
highway bridge girders, a coordinated analytical and states.1 Partial debonding may also be used to satisfy the
experimental investigation was conducted. American Association of State Highway and Transportation
Officials’ AASHTO LRFD Bridge Design Specifications2
■ The results from the testing of full-scale I- and article 5.9.4.4.1 requirements for splitting resistance at the
U-shaped girders indicate that partially debonding ends of prestressed concrete girders, which require trans-
strands does not result in deleterious performance verse confining reinforcement sufficient to resist 4% of the
if adequate reinforcement is provided to resist the total prestressing force at transfer to be located within h/4
longitudinal tension due to bending and shear. of the girder end, where h is the component depth. Partial
debonding reduces the prestressing force, which can po-
■ The requirements for debonded strands were re- tentially cause splitting at the end of the girder. In addition,
vised significantly in the ninth edition of the AAS- partial debonding can be used to relieve the compressive
HTO LRFD specifications based on the presented stress (AASHTO LRFD Specifications article 5.9.2.3.1a) at
research. the end region when required.

Excessive debonding, however, can reduce the flexural and percentage based on “successful past practice” may be
shear capacity near the girder ends as the tensile resistance considered. Some states relax the 25% limit for certain
provided by prestressing reinforcement in the debonded conditions and girder shapes; for example, Texas permits
region (Apsfps, where Aps is the area of prestressing steel and up to 75% debonding. In lieu of the AASHTO debonding
fps is average stress in prestressing steel at nominal flexural requirements, the following rules were followed to detail
resistance) is reduced. The concrete component of shear the test specimens.
strength is also reduced because the prestressing force in
regions of partial debonding is smaller. Accordingly, the • single-web flanged girders (AASHTO BT-54, AASHTO
AASHTO LRFD specifications places limits on the amount Type III, and Nebraska NU-1100)
of partial debonding.
— Do not debond more than 50% of the bottom-row
As part of National Cooperative Highway Research Program strands.
(NCHRP) project 12-91, a comprehensive study was
conducted to develop a unified approach for the design of — Keep the outermost strands in all rows located
prestressed concrete bridge girders with partially debonded within the full-width section of the flange bonded.
strands.1 Based on the results and recommendations of this
study and previous studies,3–8 the requirements for debonded — With the exception of the outermost strands, debond
strands were revised in the ninth edition of the AASHTO strands further from the vertical centerline of the
LRFD specifications.2 This paper summarizes the experi- section preferentially to those nearer the centerline.
mental component of NCHRP project 12-91 and the results
that support some of these new requirements. — Use a strut-and-tie (STM) model to check capacity
of transverse bulb splitting reinforcing steel based
Experimental program on strand geometry. The STM model has been de-
scribed by Harries et al.10 and Shahrooz et al.1
Both ends of six full-scale prestressed concrete bridge
girders (12 tests) with different debonding ratios (area of • double-web sections with bottom flanges (AASHTO
debonded strands/total area of strands) were tested to failure. BI-36 and Texas U-40)
One end of all but one of the girders (designated as end B)
had a debonding ratio of less than 25%; the remaining girder — Do not debond more than 50% of the bottom-row
had a debonding ratio of 27%. The debonding ratio in the strands.
other end (designated as end A) was greater than 25% in
all test girders. Except for the level of debonding ratio, the — Debond strands from the centerline of the section
detailing and loading of end A and end B were identical. The outward.
performances of end A and end B were compared. The test
variables were girder shape (single-web girder, box girder, — For bearings placed below webs not connected by an
or U girder), debonding ratio, concrete compressive strength, end diaphragm, engage a width equal to the exten-
and strand diameter. sion of all webs at supports.
Test specimen design • all girders
The test specimens were designed according to the eighth — In accordance with article 5.9.4.3.3 of the
edition of the AASHTO LRFD specifications.9 All subsequent AASHTO LRFD specifications, debonding termi-
references to the AASHTO LRFD specifications in this paper nations (that is, initiation of the strand embedment)
will be to the eighth edition unless otherwise noted. All appli- were staggered such that no more than 40% of
cable requirements (in particular, providing sufficient split- debonded strands and four strands were terminated
ting resistance and confinement reinforcement and ensuring at any section. Terminations were staggered 36 in.
adequate longitudinal tensile resistance) were satisfied with (910 mm) along the girder length. Debonded strands
the exception of the following: were distributed symmetrically about the vertical
centerline of the component cross section.
• “The number of partially debonded strands should not
exceed 25 percent of the total number of strands.” The provision for greater amounts of partial debonding
required additional nonprestressed longitudinal reinforce-
• “The number of debonded strands in any horizontal row ment to ensure that shear capacity near the girder ends would
shall not exceed 40 percent of the strands in that row.” comply with article 5.7.3.5 of the AASHTO LRFD specifica-
tions.1,11 These additional longitudinal bars were provided in
The wording for the first requirement in the AASHTO the form of hairpins (two longitudinal bars per unit) or stan-
LRFD specifications was nonmandatory (“should,” not dard 180-degree hooks so that they could be fully developed
“shall”), and the commentary indicated that a greater close to the face of the girder support.

Specimen details forcement. The strands were 270 ksi (1860 MPa) low-relax-
ation strands.
Tables 1 and 2 summarize details of the specimens, includ-
ing the maximum debonding ratio dr for each section, the In each girder, end B satisfied all extant requirements and
debonding ratios for each layer of reinforcement and each ter- guidelines of the eighth edition of the AASHTO LRFD specifi-
mination section, and measured concrete strengths at release cations9 with respect to strand partial debonding. End A exceed-
and at time of test. Figure 1 shows the strand patterns and ed the recommended partial debonding limits. Nevertheless,
placement of nonprestressed reinforcement. Except for the additional nonprestressed reinforcement was required to satisfy
AASHTO BI-36 girder, all the girders had a 6 in. (150 mm) article 5.7.3.5 in all but one case (AASHTO BI-36 end B).
thick slab over the entire width of the top flange. The deck
slab reinforcement was designed according to the empirical Test setup
design procedure described in article 9.7.2.5 of the AASHTO
LRFD specifications.9 The AASHTO BI-36 specimen was The girders were supported on 12 in. (300 mm) long neoprene
tested without a slab because it is common practice to use this pads like those typically used in bridge construction. For
section in adjacent box-girder bridges with only an asphalt single-web flanged sections, full-flange-width pads, having a
deck. In addition, a 2.5 ft (0.76 m) thick end diaphragm was thickness of 1.375 in. (34.93 mm), were provided for all the
provided for the AASHTO BI-36 girder to replicate common girders except for the AASHTO BT-54 girder, which had only
practice used in box girders. 22 in. (560 mm) of the 24.5 in. (622 mm) bottom flange width
supported due to pad availability. For the AASHTO BI-36
Table 3 presents the material properties for the reinforcing specimen, two 9 in. (230 mm) wide by 3 in. (76 mm) thick
steel, which was ASTM A61512 Grade 60 (414 MPa) rein- neoprene pads were placed under each web. Two 3 in. thick
End A End B End A End B
AASHTO BI-36 (w = 36 in.) AASHTO BT-54 (w = 26 in.)
AASHTO III-a (w = 22 in.) AASHTO III-b (w = 22 in.)
NU-1100 (w = 38 83 in.) U-40 (w = 55 in.)
Debond strand 3 ft from end Debond strand 6 ft from end

Debond strand 9 ft from end Debond strand 12 ft from end
Nonprestressed reinforcement to satisfy AASHTO Article 5.7.3.5
Bearing pad (typical)
Figure 1. Debonding pattern and locations of prestressed and nonprestressed reinforcement. Note: w = width of bottom flange.
1 in. = 25.4 mm; 1 ft = 0.305 m.

Table 1. Concrete and prestressing details of the test specimens
Number Debonding ratio dr

fci , fc at time db,
Girder End Distribution of 0 to 3 to 6 to 9 to
ksi of test, ksi in.
strands 3 ft 6 ft 9 ft 12 ft
Section 22 0.50* 0.36 0.18 0.09
A 12.6 Row 1 15 0.40 0.27 0.13 0
Row 2 7 0.71 0.57 0.29 0.29

AASHTO BI-36 7.4 0.5
Section 22 0.18* 0.09 0 0
B 12.2 Row 1 15 0.13 0 0 0
Row 2 7 0.29 0.29 0 0
Section 20 0.60* 0.40 0.20 0
A 17.4 Row 1 10 0.40 0.40 0 0
Row 2 10 0.80 0.40 0.40 0

AASHTO BT-54 10.2 0.6
Section 20 0.10* 0 0 0
B 15.2 Row 1 10 0.20 0 0 0
Row 2 10 0.00 0 0 0
Section 16 0.50* 0.25 0.13 0
A 12.6 Row 1 8 0.25 0 0 0
Row 2 8 0.75 0.50 0.25 0

AASHTO Type III-a 6.9 0.5
Section 16 0.25* 0.13 0 0
B 12.2 Row 1 8 0.25 0 0 0
Row 2 8 0.25 0.25 0 0
Section 18 0.56* 0.33 0.11 0
A 13.8 Row 1 8 0.25 0 0 0
Row 2 10 0.80 0.60 0.20 0

AASHTO Type III-b 8.3 0.5
Section 18 0.22* 0.11 0 0
B 13.2 Row 1 8 0.25 0 0 0
Row 2 10 0.20 0.20 0 0
Section 22 0.45* 0.27 0.18 0.09
A 14.0 Row 1 18 0.44 0.33 0.22 0.11
Row 2 4 0.50 0 0 0
Nebraska NU-1100 8.4 0.7
Section 22 0.27* 0.18 0.18 0.18
B 13.2 Row 1 18 0.33 0.22 0.22 0.22
Row 2 4 0.00 0 0 0
Section 26 0.50* 0.35 0.19 0
A 12.8 Row 1 19 0.42 0.21 0 0
Row 2 7 0.71 0.71 0.71 0

Texas U-40 6.9 0.6
Section 26 0.23* 0.15 0.08 0
B 12.0 Row 1 19 0.21 0.11 0 0
Row 2 7 0.29 0.29 0.29 0

Note: db = nominal strand diameter; dr = debonding ratio; fc′ = compressive strength of concrete; fci′ = compressive strength of concrete at time of initial
loading or prestressing. 1 in. = 25.4 mm; 1ft = 0.305 m; 1 ksi = 6.895 MPa.
*The maximum debonding ratio dr for each section. Other values represent debonding ration dr for each layer of reinforcement and each termination
section.

Table 2. Nonprestressed ASTM A615 Grade 60 reinforcement details of the test specimens
Nonprestressed ASTM A615 Grade 60 reinforcement
Longitudinal Transverse (no. 4)
Girder End Distribution

Web
Cutoff
Reinforcement Bottom flange
point, ft
U shaped
Section
A Row 1 Six no. 6 8.5

Three at 3 in., eight at
Row 2 At 12 in. (outside of end
AASHTO BI-36 6 in., and at 12 in. to
Section diaphragm)
midspan
B Row 1 None n/a
Row 2
Section Two no. 6 6.5
A Row 1
Eight no. 6 13.5 Nine at 3 in., eleven at
Row 2 Five at 3 in. and at 18 in.
AASHTO BT-54 6 in., and at 18 in. to
Section to midspan
midspan
B Row 1 Six no. 6 6.5
Row 2
A Row 1
Six no. 5 10.5 Four at 3 in., eleven at
Row 2 Four at 3 in. and at 18
AASHTO Type III-a 6 in., and at 18 in. to
Section Two no. 5 6.5 in. to midspan
midspan
B Row 1
Four no. 5 9.5
Row 2
A Row 1
Six no. 5 10.5 Four at 3 in., eleven at
Row 2 Four at 3 in. and at 18
AASHTO Type III-b 6 in., and at 18 in. to
Section Four no. 5 5.5 in. to midspan
midspan
B Row 1
Four no. 5 9.5
Row 2
Section
A Row 1 Six no. 6 5.5

Four at 3 in., eleven at
Row 2 Four at 3 in. and at
Nebraska NU-1100 6 in., and at 12 in. to
Section 12 in. to midspan
midspan
B Row 1 Four no. 6 5.5
Row 2
Section
Twenty-two
A Row 1 14.5
no. 6
Four at 3 in., Four at 3 in.,
Row 2
Texas U-40 twenty-three at 4 in., twenty-three at 4 in.,
Section and at 6 in. to midspan and at 6 in. to midspan
B Row 1 Sixteen no. 6 13.5
Row 2
Note: no. 5 = 16M; no. 6 = 19M; 1 in. = 25.4 mm; 1 ft = 0305 m; 1 ksi = 6.895 MPa; Grade 60 = 414 MPa.

neoprene pads were also placed under each web of the Texas Table 3. Measured material properties of nonpre-
U-40 girder. These pads engaged the outer 7 in. (180 mm) stressed reinforcement
width of the bottom flange. Displacement transducers were
Girder Bar size fy, ksi fu, ksi εu
attached to the girder to measure the compression of pads
during testing, and reported girder deflections were corrected No. 3 82.1 120 0.126
for this movement. AASHTO
No. 4 72.7 112 0.128
BI-36
The girders were tested in three-point bending using either No. 6 65.4 102 0.186
a single 1200 kip (5300 kN) hydraulic ram or dual 300 kip
(1300 kN) hydraulic rams in parallel. The location of the AASHTO No. 4 69.7 107 0.127
load was selected such that the shear span–to–depth ratio a/dv BT-54 No. 6 65.9 106 0.132
would be greater than 2.0 to prevent direct transfer of the load
to the support through arching action. All reported deflections AASHTO No. 4 63.6 100 0.191
were measured at the point of application of load. Each girder Type III-a No. 5 75.6 113 0.159
end was tested separately, with end B tested first. After testing
end B, the girder was repositioned to test end A, which had AASHTO No. 4 63.6 100 0.191
a greater debonding ratio. The Texas U-40 girder was tested Type III-b No. 5 75.6 113 0.159
as a simply supported span; in all other cases, testing of each
end consisted of a simple span L with a propped cantilever No. 3 75.1 101 0.238
overhang (Fig. 2 and Table 4), isolating the other test end
Nebraska No. 4 79.0 106 0.254
of the girder. To prevent cracking due to the self-weight of
the cantilevered portion, an air jack was used to support the NU-1100 No. 5 70.1 103 0.128
cantilevered end of the girder. The air pressure was calibrated
No. 6 69.2 109 0.120
such that the force in the jack actively compensated for the
self-weight of the cantilevered portion throughout the duration No. 4 70.5 110 0.157
of the test; thus, the girder was effectively tested as a simply Texas
No. 5 67.1 105 0.093
supported span. Regions 1 and 2 in Fig. 2 identify locations U-40
where shear deformations were measured. The Texas U-40 No. 6 67.6 110 0.145
girder was tested as a simply supported span because testing
Note: fu = ultimate strength; fy = yield strength; εu = ultimate strain. No. 3
similar to that used for the other girders would have required
= 10M; no. 4 = 13M; no. 5 = 16M; no. 6 = 19M; 1 ksi = 6.895 MPa.
a longer girder, which would have exceeded the laboratory’s
main crane capacity.
Figure 2. Test setup for test specimen AASHTO BT-54. Note: a = shear span; L = span length; LG = overall girder length; P = load.

Test results and discussion Table 4. Test specimen span and location of
applied load
Except for the Nebraska NU-1100 girder and end B of the
Girder LG, ft L, ft a, ft a/dv
Texas U-40 girder, all specimens were loaded to failure.
Failure at end B of the Texas U-40 specimen would have com- AASHTO
40 29 5 2.73
promised, if not prevented, testing of end A; therefore, end B BI-36
of this girder was loaded to only the capacity predicted by the
AASHTO LRFD specifications. It was deemed unsafe to load AASHTO
55 34 10 2.34
the Nebraska NU-1100 girder, which had twenty-two 0.7 in. BT-54
(18 mm) diameter strands, to failure given the amount of AASHTO
energy that would have been released in the event of a sudden 55 39 7.75 2.15
Type III-a
failure. Therefore, this girder was loaded to only slightly
above its predicted capacity. AASHTO
55 39 7.75 2.16
Type III-b
Capacity, stiffness, and failure mode Nebraska
55 39 7.75 2.20
NU-1100
Figure 3 shows the failure patterns of the girders in which both
ends were loaded to their ultimate capacities. Based on these Texas U-40 32 31 7.75 2.39
photographs, the failure modes were characterized as indicated in Note: a = shear span; dv = effective shear depth; L = span length;
the figures. In general, comparing end A and end B of the same LG = overall girder length. 1 ft = 0.305 m.
girder demonstrated that the failure modes were not influenced
by the amount of debonding. End A of the Texas U-40 girder
failed due to a combination of shear tension and bearing (Fig. 3). comparable. The normalized load-deflection for the AASHTO
BI-36 girder (Fig. 4) clearly illustrates that end A achieved
The measured material properties were used to calculate the its peak capacity at a larger deflection than end B. Figure 5
expected capacity of each specimen per the AASHTO LRFD indicates that the strands in end B of the AASHTO BI-36
specifications (referred to herein as the AASHTO-predicted girder experienced a sudden dramatic slip after achieving the
capacity). For capacity calculations, the experimental loading peak normalized load of 1.60. End B of this specimen did
conditions were used (single-point loading) and all failure not have any nonprestressed reinforcement to compensate
modes were considered. In all but one case, shear was found for the reduction of prestressing force due to slip. As a result,
to be the controlling mode; however, end B of the AASHTO load-carrying capacity at this end began to drop once the
BI-36 specimen was controlled by the available tensile capac- strands slipped (Fig. 5). The two ends of the AASHTO BT-54
ity at the inside face of the support. In the calculations, resis- girder experienced different failure modes (shear compression
tance factors were taken as unity because the test girders were and sliding shear), which may explain why the normalized
cast under controlled conditions, the loading was well defined deflections at peak load differed between the ends of these
and known from theoretical deduction, and the purpose of the two girders. The two ends of the Nebraska NU-1100 and
calculation was to determine a predicted capacity and not a Texas U-40 specimens could not be compared because in both
design load. The measured applied loads (and shears) were girders, end B was not tested to failure.
normalized with respect to the calculated capacities of each
girder. The measured deflections at the point of loading were For all girders, the slopes of the normalized load-deflection
normalized with respect to the deflection measured at the relationships at end A and end B are essentially the same up
predicted capacity. Figure 4 shows the resulting normalized to the predicted capacities (that is, when the value of normal-
load-deflection responses. In each case, end B met the extant ized load is equal to 1). The larger amount of debonding at
limits on the amount of debonding from the AASHTO LRFD end A did not have a noticeable effect on the overall stiffness
specifications,9 whereas end A exceeded these limits. of the girders. This observation should be expected because
the relatively small area of prestressing reinforcement does
All specimens successfully developed their predicted capaci- not affect the stiffness, and debonding, which is localized near
ties. The failure loads were at least 50% greater than the the girder ends, has little effect on deflection.
AASHTO-predicted capacities (without reduction factors) cal-
culated using measured material properties. Moreover, at peak Shear deformation
load, which corresponded to failure if the specimen was loaded
to its ultimate capacity, the deflection was at least 2.3 times that Using diagonal displacement transducers mounted to the
when the predicted capacity was developed. The large amounts specimen webs, the average shear deformations in two adja-
of debonding present in the end A tests did not affect the ex- cent regions were obtained (Fig. 2). Region 1 is approximate-
pected load-carrying or displacement capacity (ductility). ly one-half the shear span closer to the support, and region 2
is the other half of the shear span closer to the applied load.
For the AASHTO Type III-a and Type III-b girders, the Table 5 compares the average shear strains at AASHTO-
normalized deflections at peak load for end A and end B were predicted capacity in these regions.

Failure mode: shear compression Failure mode: shear compression
End A (dr = 0.50) End B (dr = 0.18)
AASHTO BI-36
Failure mode: shear compression Failure mode: sliding shear at web-flange interface
End A (dr = 0.60) End B (dr = 0.10)
AASHTO BT-54

End A (dr = 0.50) End B (dr = 0.25)
AASHTO Type III-a

End A (dr = 0.56) End B (dr = 0.22)
AASHTO Type III-b
Failure mode: Shear tension and bearing

End B (dr = 0.50)
Figure 3. Crack patterns and modes of failure (compound image from multiple photos). Note: dr = debonding ratio.

1.8 1.8
1.6 1.6
1.4 1.4
1.2 1.2
Normalized load
Normalized load
1.0 1.0
0.8 0.8
0.6 0.6
End A End A
0.4 End B 0.4 End B
0.2 0.2
0 0
0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7
Normalized deflection Normalized deflection
AASHTO BI-36 AASHTO BT-54
1.8 1.8
1.6 1.6
1.4 1.4
1.2 1.2
Normalized load
Normalized load
1.0 1.0
0.8 0.8
0.6 0.6
End A End A
0.4 End B 0.4 End B
0.2 0.2
0 0
0 1 2 3 4 5 6 7 8 9 10 11 0 1 2 3 4 5 6 7 8 9 10 11
AASHTO Type III-a AASHTO Type III-b
1.6 1.6
1.4 1.4
1.2 1.2
Normalized load
Noramlized load
1.0 1.0
0.8 0.8
0.6 0.6
End A
0.4 End A 0.4
End B
End B
0.2 0.2
0 0
0 0.4 0.8 1.2 1.6 2.0 0 1 2 3 4 5
Nebraska NU-1100 Texas U-40
Figure 4. Normalized load-deflection responses.
In general, the shear strain at end A was larger than that at end B; patterns. The photos generally do not suggest any discern-
on average, the strain at end A was 17% and 5% larger for region able differences between the crack patterns at the two ends
1 and 5% larger for region 2. It is theorized that the smaller of a given girder. However, for some girders (such as the
amount of prestressing force (resulting from the larger debonding AASHTO Type III-a and Type III-b specimens), end A expe-
ratio) at end A could not restrain the growth and widening of the rienced more cracking and exhibited more of a flexure-shear
cracks to the same extent as the larger amount of prestressing behavior than end B, where behavior was predominantly
force at end B, and, hence, the shear strains were larger at end A. controlled by web shear. These observations are consistent
with the smaller amount of prestressing force (due to greater
Crack angles and widths debonding) at end A.
Photos (Fig. 3) of each girder at failure were used to deter- For a single girder, the average crack angles θcr (measured
mine the angles of diagonal cracks and to compare crack from horizontal) were essentially the same for the two ends

Apparent slip, in.
1.8 0 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 0.55
1.6
1.4
1.2
Normalized load
1.0
0.8
Normalized deflection
0.6
Slip: bonded
0.4 Slip: debonded 3 ft
Slip: debonded 6 ft
0.2
0
0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0
Normalized deflection
Figure 5. Normalized load versus normalized deflection and apparent strand slip for end B in test specimen AASHTO BI-36.
Note: 1 in. = 25.4 mm; 1 ft = 0.305 m.
with different debonding ratios (Table 6). The maximum Table 5. Average shear strain at capacity predicted
crack widths wmax at end A, which had a larger debonding by the American Association of State Highway and
ratio than end B, were generally wider than those at end B. Transportation Officials’ AASHTO LRFD Bridge
However, the maximum measured crack widths corresponding Design Specifications
to the AASHTO-predicted capacities were less than 0.03 in.
End A End B
(0.76 mm) in all cases. The larger debonding ratio did not Girder
have a deleterious effect on the observed crack angles or crack Region 1 Region 2 Region 1 Region 2
widths. Table 5 also presents the load at which the first crack
occurred P@1stcrack normalized with respect to the AASHTO- AASHTO
0.00042 0.00068 0.00016 0.00048
predicted capacities PAASHTO for each end of the six girders. BI-36
End A cracked at a load that was, on average, 6% lower than AASHTO
that for end B. This observation is consistent with smaller 0.00090 0.00052 n/a n/a
BT-54
amount of prestressing force at end A, which had a higher
level of debonding than end B. AASHTO
0.00113 0.00181 0.00068 0.00067
Type III-a
Shear resistance from transverse AASHTO
reinforcement Type III-b
0.00104 0.00065 0.00106 0.00100
Stress-strain relationships from materials testing were used Nebraska

0.00142 0.00044 0.00143 0.00089
to infer stresses in the transverse reinforcement from trans- NU-1100
verse reinforcing bar strains measured in the girder tests.1,11 In Texas U-40 0.00085 0.00125 0.00069 0.00119
this manner, the shear resistance provided by the transverse
reinforcement Vs can be determined using Eq. (1), which is Note: n/a = not applicable (because instruments had been removed).
Eq. (5.7.3.3-4) of the AASHTO LRFD specifications for a

case in which transverse reinforcement is perpendicular to where
the longitudinal axis and the yield strength of the transverse
reinforcement fy is substituted with the inferred stress fv: Av = area of shear reinforcement
Vs =
(
Av f v dv cot θ ) (1) fv = stress in shear reinforcement
s

Table 6. Average measured angles and widths of cracks at capacity predicted by the American Association of
State Highway and Transportation Officials’ AASHTO LRFD Bridge Design Specifications
End A End B
Girder
θcr, deg wmax, in. P@1stcrack /PAASHTO θcr, deg wmax, in. P@1stcrack /PAASHTO
AASHTO BI-36 29 ≤0.01 0.61 30 ≤0.01 0.67
AASHTO BT-54 34 0.025 0.77 32 0.022 0.78
AASHTO Type III-a 35 0.028 0.80 35 0.014 0.89
AASHTO Type III-b 33 0.015 0.79 34 0.025 0.73
Nebraska NU-1100 32 ≤0.01 0.62 32 0.015 0.70
Texas U-40 32 0.014 0.55 34 ≤0.01 0.63
Note: P@1stcrack = load at which the first crack occurred; PAASHTO = load at capacity predicted by the AASHTO LRFD specifications; wmax = maximum crack
width; θcr = average crack angle measured from horizontal. 1 in. = 25.4 mm.
dv = effective shear depth

0
2
θ = angle of inclination of diagonal compressive R = 0.68
Change in normalized Vc
stresses −0.01 Test
AASHTO
−0.02
s = average spacing of shear reinforcement
−0.03
The values of θcr in Table 5 were substituted for θ in the Vs 2
−0.04 R = 0.67
calculations. This calculation was performed at six locations
(every 1 ft [0.3 m] up to 6 ft [1.8 m] from the ends of the −0.05
girder) where the transverse reinforcement had been instru- 0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 0.55
mented. Because there was no harped prestressing strand Change in debonding ratio
component, the difference between the applied shear and the

average of Vs from these six locations was the experimentally Figure 6. Normalized Vc at end A minus Vc at end B versus
inferred concrete contribution to shear Vc. The value of Vc differences in the debonding ratios. Note: R2 = coefficient
was also determined based on Eq. (5.7.3.3-3) of the of determination; Vc = nominal shear resistance provided by
tensile stresses in the concrete.
AASHTO LRFD specifications. Both the experimental and
calculated values of Vc were normalized by f c′bv dv .
Figure 7 compares the measured slip data at AASHTO-
Figure 6 plots the difference between the normalized values predicted capacities for end A and end B. The slip of bonded
of Vc at end A and end B versus the difference between the strands was 0.004 in. (0.1 mm) or less for all cases except for
level of debonding ratio at the two ends. As expected, the end A of the Texas U-40 girder, which experienced a slip of
value of Vc at end A, which had a larger debonding ratio, was 0.011 in. (0.28 mm). The measured slip of debonded strands
smaller than its counterpart at end B, which had a smaller rarely exceeded 0.043 in. (1.1 mm), except for the following
debonding ratio. Thus, the beneficial effect of precompres- cases in the Texas U-40 girder:
sion on the concrete contribution to shear capacity is evident.
AASHTO Eq. (5.7.3.3-3) indicates a similar rate of reduction • strands debonded 3 ft (0.9 m) at end B with slip of
compared with the test data (Fig. 6). 0.060 in. (1.5 mm)
Apparent strand slip • end A with a debonding length of 6 ft (1.8 m) with slip of
0.073 in (1.9 mm)
For each specimen, several bonded strands and all debonded
strands were instrumented with displacement transducers to • both end A and end B in which the strands were debond-
measure the movements of the strands relative to the end face ed for 9 ft (2.7 m) with slips of 0.060 and 0.099 in. (1.5
of the girder. For fully bonded strands, this movement is the and 2.5 mm), respectively
actual slip. In the case of debonded strands, the measured slip
is affected by the length of partial debonding but in a manner The partially debonded strands exhibited greater slip at end
that cannot be corrected. A than at end B, except for the debonding lengths of 12 ft

0.12
0.02
1.8
magnification
up to 0.02 in. 1.6
0.10
1.4
End B normalized shear

End B slip at AASHTO capacity, in.
NU-1100
0.08 1.2
0 1.0
0.06 0 0.02
0.8
U-40
0.6
0.04 Bonded
NU-1100
0.4 Debonded (3 ft)
Bonded Debonded (6 ft)
0.02 Debonded (3 ft) 0.2 Debonded (9 ft)
Debonded (6 ft)
Debonded (12 ft)
Debonded (9 ft)
Debonded (12 ft)
0
0 0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8
0 0.02 0.04 0.06 0.08 0.10 0.12
End A slip at AASHTO capacity, in. End A normalized shear
Figure 7. Comparison of slips at the capacity predicted by the Figure 8. Comparison of normalized shear corresponding to
American Association of State Highway and Transportation 0.01 in. slip. Note: 1 in. = 25.4 mm; 1 ft = 0.305 m.
Officials’ AASHTO LRFD Bridge Design Specifications. Note:
1 in. = 25.4 mm; 1 ft = 0.305 m.
ϕf = resistance factor for moment resistance
(3.7 m) in the Nebraska NU-1100 girder and 3 ft (0.9 m) in
the Texas U-40 girder. Nu = applied factored axial force at the section under
consideration, taken as positive if tensile
The amount of debonding affected the onset of slip of bonded
and debonded strands. To illustrate this dependency, Fig. 8 com- ϕc = resistance factor for axial resistance
pares the normalized shears corresponding to 0.01 in. (0.25 mm)
slip at end A and end B. This value, which is one-tenth of the slip Vu = applied factored shear force at section
used for strand evaluation according to ASTM A1081,13 was ar-
bitrarily selected, and similar observations can be made for other ϕv = resistance factor for shear resistance
values of slip displacement. The normalized shears required to
develop 0.01 in. slip for the bonded and debonded strands at end Vp = component in the direction of the applied shear of
B were higher than their counterparts in end A. That is, slippage the effective prestressing force
of strands in the cases with greater levels of debonding occurred
sooner than in those cases with less debonding. θ = angle of inclination of diagonal compressive stresses
Contribution of longitudinal ⎛V ⎞
Aps f ps + As f y ≥ ⎜ u − 0.5Vs − V p ⎟ cot θ (AASHTO 5.7.3.5-2)
reinforcement φ
⎝ v ⎠
As part of design of the test specimens, the amounts of nonpre-
stressed longitudinal reinforcement required at the critical section Based on a similar procedure used to calculate shear resis-
and at the interior face of the support were determined according tance from transverse reinforcement, stresses in the nonpre-
to AASHTO Eq. (5.7.3.5-1) and (5.7.3.5-2), respectively: stressed longitudinal reinforcement were inferred from the ex-
perimentally determined stress-strain relationships.1,11 Table 7
Mu N u ⎛ Vu ⎞ presents the resulting nonprestressed reinforcement stresses
Aps f ps + As f y ≥ + 0.5 +⎜ − V p − 0.5Vs ⎟ cot θ
dvφ f φc ⎝ φv ⎠ fs normalized with respect to their measured yield strengths
for the AASHTO-predicted capacity and at the maximum
(AASHTO 5.7.3.5-1)
experimentally applied load. If available, stresses at three sec-
where tions are given at the critical section near the support (XS2), a
distance dv from the interior face of the support (XS3), and at
As = area of nonprestressed tension reinforcement the point where the load was applied (XS4).
Mu = applied factored bending moment at the section At the AASHTO-predicted capacity, the stress in the
under consideration nonprestressed reinforcing steel is equal to or less than

Table 7. Normalized stress in nonprestressed longitudinal reinforcement
At AASHTO-predicted capacity At maximum load
End A End B End A End B

Girder
fs /fy fs /fy fs /fy fs /fy
XS2 XS3 XS4 XS2 XS3 XS4 XS2 XS3 XS4 XS2 XS3 XS4
AASHTO BI-36 0.024 0.094 0.173 n/a n/a n/a 1.02 1.11 0.89 n/a n/a n/a
AASHTO BT-54 0.037 0.065 0.124 0.12 0.14 n/a 1.01 1.02 0.45 1.02 0.76 n/a
AASHTO Type III-a 0.104 0.127 0.148 0.04 0.11 0.16 0.77 1.08 0.97 0.67 0.99 0.96
AASHTO Type III-b 0.051 0.130 0.163 0.18 0.07 0.03 1.03 1.04 0.94 0.86 0.90 0.71
Nebraska NU-1100 0.261 0.074 n/a 0.24 0.08 n/a 0.79 0.17 n/a 0.31 0.09 n/a
Texas U-40 0.350 0.281 n/a 0.29 0.15 n/a 0.99 0.93 n/a 0.30 0.19 n/a
Note: dv = effective shear depth; fs = nonprestressed reinforcement stresses; fy = yield strength of reinforcing bar; n/a = not applicable; XS2 = critical
section near support; XS3 = section a distance dv from the interior face of support; XS4 = point of load application.
0.35fy; however, the longitudinal nonprestressed reinforce- • The experimentally determined girder capacities were in
ment is assumed to have yielded according to AASHTO excess of those computed based on the AASHTO LRFD
Eq. (5.7.3.5-1) and (5.7.3.5-2). A plausible explanation for specifications using measured material properties and
this difference is that AASHTO LRFD specifications do not prestress losses with no strength reduction factor. The
account for the tensile strength of the precompressed con- large amounts of debonding at end A were not detrimen-
crete. Hence, the available capacity is greater than the predict- tal to the load-carrying capacity of the girders.
ed Apsfps + Asfy term.
• Regardless of the debonding ratios, the measured
The results in Table 7 indicate that normalized nonpre- deflection at the peak load was several times larger than
stressed longitudinal reinforcement stresses at end A were the measured deflection at the capacity calculated using
generally larger than those at end B. This observation is the AASHTO LRFD specifications, which indicates ad-
consistent with the differences in the amount of prestressing equate performance in terms of displacement capacity
force at the two ends resulting from partial debonding. The (ductility).
smaller prestressing force at end A resulted in more crack-
ing and, hence, an earlier onset and higher redistribution • Up to the AASHTO-predicted capacity, the slopes of the
of the tensile force from the precompressed concrete to the normalized load-deflection curves were nearly identical
reinforcement. For most of the girders that were loaded to for end A and end B. The larger debonding ratio at end
failure, the nonprestressed longitudinal bars began to yield at A did not have a noticeable impact on normalized girder
either section XS2 (the critical section near the support) or stiffness.
section XS3 (the section at the distance dv from the interior
face of the support), or at both sections. All nonprestressed • In general, the average measured concrete shear strain for a
reinforcement was fully developed at these locations using given value of applied shear was larger at end A than at end
hairpins or standard hooks. B. This observation should be expected because the smaller
amount of prestressing force (resulting from the larger
Conclusion debonding ratio) at end A would not restrain the growth and
widening of the cracks to the same extent as in end B.
Experimental testing of six full-sized girders was conducted
to examine the effects of debonding on girder performance. • The crack widths at end A (larger debonding ratio)
The main test variables were girder shape, debonding ratio, were in general slightly wider than those at end B, the
concrete strength, and strand diameter. Each girder end was maximum measured crack widths corresponding to the
tested separately, with end A having a greater debonding ratio AASHTO-predicted capacities were small regardless of
than end B. Detailing and loading of the two girder ends were the debonding ratio, and the largest recorded crack width
identical with the exception of the debonding ratios. Girders was less than 0.03 in. (0.76 mm). Moreover, for a single
were designed to meet all requirements set forth in the eighth girder, the average crack angles were essentially the same
edition of the AASHTO LRFD specifications except for the for the two ends with different debonding ratios. The
25% limit on debonding ratios. Based on the presented experi- larger debonding ratio did not have a deleterious effect on
mental results, the following conclusions are made: observed crack angles or crack widths.

• The smaller prestressing force at end A resulted in more specimens. Up to 80% of strands per row were debonded
cracking, and, hence, the contribution of the concrete to the in the test girders.
shear resistance was reduced. However, the performance
was not adversely affected because shear equations in the • The AASHTO limit of 45% debonded strands per row
AASHTO LRFD specifications capture the smaller con- also means that the total number of debonded strands
crete shear strength as the level of debonding increases. is limited to 45%. In the test girders, overall debonding
ranged from 10% to 60% of the total strands in the cross
• In general, there were no discernable differences section.
between the crack patterns at the ends of a given girder.
However, for some girders (such as AASHTO Type III-a • The AASHTO LRFD specifications require that the
and Type III-b), end A experienced more cracking and debonding termination locations be at least 60db,
exhibited more of a flexure-shear behavior than end B, where db is the nominal strand diameter. A longitudinal
whose behavior was predominantly controlled by web spacing of debonding termination locations of 36 in.
shear. These observations are consistent with the smaller (910 mm) was used in the test specimens. For 0.5 and
amount of prestressing force (due to greater debonding) 0.6 in. (13 and 15 mm) diameter strands, this spacing
at end A. corresponds to 72db and 60db, respectively; thus, this
requirement was met by all test specimens except the
• At AASHTO-predicted girder capacities, measured slip Nebraska NU-1100 girder. For the NU-1100 girder,
did not exceed 0.03 in. (0.76 mm), except in the debond- which had 0.7 in. (18 mm) diameter strands, the spacing
ed strands in the Texas U-40 specimen, which slipped corresponded to 51.4db.
between 0.036 and 0.104 in. (0.91 and 2.64 mm). The
measured slip of fully bonded strands was negligible in • The AASHTO LRFD specifications require that the loca-
all cases. All girders exhibited acceptable amounts of tions of bonded and debonded strands be alternated both
strand slip up to the capacity required in the AASHTO horizontally and vertically. This requirement was exceed-
LRFD specifications. ed in the reported research. Many of the test specimens
had adjacent strands debonded horizontally, vertically, or
• The nonprestressed reinforcement used to compensate in both directions.
for larger debonding ratios was found to be adequate
in terms of capacity. Although this reinforcement could • The AASHTO LRFD specifications limit the debond-
not replicate the effects of prestressing force in bonded ing length, measured from the girder end, for sim-
strands, investigators found only small differences in ple-span girders to the smaller of 20% of the span
the overall stiffness, crack widths, and crack patterns length or one-half the span length minus the develop-
and angles of cracks of the two girder ends with dif- ment length. This requirement was not satisfied in the
ferent magnitudes of debonding ratios. The results are test program. To accommodate practical requirements
consistent with the hypothesis that bonded strand and for laboratory testing, all of the girders tested had
nonprestressed tension reinforcement work together to relatively short spans.
resist longitudinal forces induced by shear (that is, those
calculated using Eq. [5.7.3.5-1] and [5.7.3.5-2] of the • For single-web flanged sections such as I-beams and bulb
AASHTO LRFD specifications). tees, the AASHTO LRFD specifications require the fol-
lowing. The single-web flange test specimens (Nebraska
AASHTO and the American Society of Civil Engineers have NU-1100, AASHTO Type III, and AASHTO BT-54) met
addition resources that cover related topics.14–18 the last requirement but did not meet the first two require-
ments. Instead, an STM model1,10 was used to check the
Ninth edition of the AASHTO LRFD capacity of the confining reinforcement.
specifications updates
— All strands within the horizontal limits of the web
The reported research was conducted as part of NCHRP must remain bonded if the total number of debonded
project 12-91, “Strand Debonding for Pretensioned Girders.” strands exceeds 25%.
The results of this research program1 along with previous
studies3–8 were used to update the requirements for debonded — All strands within all rows that are within the pro-
strands in the ninth edition of the AASHTO LRFD spec- jected width of the flange must be bonded.
ifications. Following are differences between the revised
AASHTO article 5.9.4.3.3 and the experimental program: — The strands that are the farthest from the vertical
centerline must be debonded.
• The ninth edition of the AASHTO LRFD specifications
limits the number of debonded strands per row to 45% • For box beams, U girders, or voided slabs, the AASHTO
of the strands in that row unless approved by the owner. LRFD specifications require the following. The multi-
This revised requirement was exceeded at end A of all web flange test specimens (AASHTO BI-36 and Texas

U-40) met the last requirement but did not meet the first Concrete Beams Containing Debonded Strands.” PCI
two requirements. Journal 39 (5): 60–77.
— Debonded strands must be distributed uniformly 6. Russell, B. W., and N. H. Burns. 1994. “Fatigue Tests
between the webs. on Prestressed Concrete Beams Made with Debonded
Strands.” PCI Journal 39 (6): 70–88.
— Strands must be bonded within 1.0 times the web-
width projection. 7. Langefeld, D. P., and O. Bayrak. 2012. Anchorage-
Controlled Shear Capacity of Prestressed Concrete Bridge
— The outermost strands must be bonded. Girders. Technical report. Austin, TX: University of Texas.
Although the test specimens did not meet most of the revised 8. Hamilton, H. R., G. R. Consolazio, and B. E. Ross.
requirements in the ninth edition of the AASHTO LRFD spec- 2013. End Region Detailing of Pretensioned Concrete
ifications, the performance of the test girders was found to be Bridge Girders: Final Report. Tallahassee, FL: Florida
adequate. In conclusion, the experimental results demonstrate Department of Transportation.
that the changes to the AASHTO LRFD specifications are
supported and continue to be conservative. 9. AASHTO. 2017. AASHTO LRFD Bridge Design
Specifications, 8th ed. Washington, DC: AASHTO.
Acknowledgments
10. Harries, K. A., B. M. Shahrooz, B. E. Ross, P. Ball,
The research presented in this paper was a part of NCHRP and H. R. Hamilton. 2019. “Modeling and Detailing
project 12-91, “Strand Debonding for Pretensioned Girders.” Pretensioned Concrete Bridge Girder End Regions Using
The authors wish to thank the NCHRP Project Panel and the the Strut and Tie Approach.” ASCE Journal of Bridge
senior program officer, Waseem Dekelbab, for their project Engineering 24 (3): 04018123. https://doi.org/10.1061
oversight and valuable insight and feedback throughout /(ASCE)BE.1943-5592.0001354.
the project. Malory Gooding, a student at the University of
Cincinnati, and James Thompson, a student at Cincinnati 11. Bolduc, M. W. 2020. “Full-Scale Testing of Pretensioned
State Technical and Community College working through Concrete Girders with Partially Debonded Strands.” PhD
a cooperative education program with the University of dissertation. University of Cincinnati, Cincinnati, OH.
Cincinnati, were instrumental in the experimental phase, and
their contributions are gratefully acknowledged. 12. ASTM International. 2022. Standard Specification for
Deformed and Plain Carbon-Steel Bars for Concrete
References Reinforcement. ASTM A615/A615M-22. West
Conshohocken, PA: ASTM International.
1. Shahrooz, B. M., R. A. Miller, K. A. Harries, Q.
Yu, and H. G. Russell. 2017. Strand Debonding for 13. ASTM International. 2015. Standard Test Method for
Pretensioned Girders. NCHRP (National Cooperative Evaluating Bond of Seven-Wire Steel Prestressing Strand.
Highway Research Program) report 849. Washington, ASTM A1081/A1081M-15. West Conshohocken, PA:
DC: Transportation Research Board, National Research ASTM International.
Council.
14. AASHTO. 2012. Standard Specification for Steel Strand,
2. AASHTO (American Association of State Highway and Uncoated Seven-Wire for Concrete Reinforcement.
Transportation Officials). 2020. AASHTO LRFD Bridge AASHTO M 203M. Washington, DC: AASHTO.
Design Specifications, 9th ed. Washington, DC: AASHTO.
15. AASHTO. 2014. Standard Method of Test for
3. Shahawy, M., and B. de V Batchelor. 1991. Bond and Compressive Strength of Cylindrical Concrete Specimens.
Shear Behavior of Prestressed AASHTO Type II Beams. AASHTO T 22. Washington, DC: AASHTO.
Progress report. Tallahassee, FL: Structural Research
Center, Florida Department of Transportation. 16. AASHTO. 2014.Standard Method of Test for Mechanical
Testing of Steel Products. AASHTO T 244. Washington,
4. Shahawy, M., B. Robinson, and B. de V. Batchelor. DC: AASHTO.
1993. An Investigation of Shear Strength of Prestressed
Concrete AASHTO Type II Girders. Research report. 17. AASHTO. 2016. AASHTO LRFD Bridge Design
Tallahassee, FL: Structures Research Center, Florida Specifications, 7th ed., with 2015 and 2016 interim revi-
Department of Transportation. sions. Washington, DC: AASHTO.
5. Russell, B. W., N. H. Burns, and L. G. ZumBrunnen. 18. Bolduc, M. W., A. Gaur, B. M. Shahrooz, K. A. Harries,
1994. “Predicting the Bond Behavior of Prestressed R. A. Miller, and H. G. Russell, H.G. 2020. “Evaluation

of Pretensioned Girders with Partial Strand Debonding.” Vp = component in the direction of the applied shear of
ASCE Journal of Bridge Engineering 25 (8). https://doi. the effective prestressing force
org/10.1061/(ASCE)BE.1943-5592.0001585.
Vs = shear resistance provided by shear reinforcement
Notation
Vu = applied factored shear force at section
a = shear span
w = width of bottom flange
Aps = area of prestressing steel
wmax = maximum crack width
As = area of nonprestressed tension reinforcement
εu = ultimate strain of nonprestressed reinforcement
Av = area of shear reinforcement
θ = angle of inclination of diagonal compressive stresses
dr = debonding ratio
θcr = average crack angle measured from horizontal
db = nominal strand diameter
ϕc = resistance factor for axial resistance
dv = effective shear depth
ϕf = resistance factor for moment resistance
f c′ = compressive strength of concrete
ϕv = resistance factor for shear resistance
f ci′ = compressive strength of concrete at time of initial
loading or prestressing
fps = average stress in prestressing steel at nominal flex-

ural resistance
fs = nonprestressed reinforcement stresses
fv = stress in shear reinforcement
fy = yield strength of nonprestressed reinforcement
h = overall depth of a component
L = span length
LG = overall girder length
Mu = applied factored bending moment at section
Nu = applied factored axial force at section, taken as

positive if tensile
P = load
P@1stcrack = load at which the first crack occurred
PAASHTO = load at capacity predicted by the AASHTO LRFD

specifications
R2 = coefficient of determination
s = average spacing between mild shear reinforcement
Vc = nominal shear resistance provided by tensile stress-

es in the concrete

About the authors distribution of partially debonded strands within the
cross section, debonded lengths, locations and stagger-
Mathew W. Bolduc, PhD, PE, is a ing of termination of debonded strands, confinement
former graduate student at the of debonded regions and their termination points, and
University of Cincinnati in the impact of adding nonprestressed reinforcement
Cincinnati, Ohio. in the region of partial debonding. The results from
testing full-scale I- and U-shaped girders indicate that
partially debonding strands does not result in deleteri-
ous performance if adequate reinforcement is provided
Bahram M. Shahrooz, PhD, FACI, to resist the longitudinal tension due to bending and
FASCE, FSEI, PE, is a professor shear. Crack patterns and angles were not noticeably
of structural engineering at the affected by the amount of debonding. Regardless of
University of Cincinnati. the debonding ratio, the maximum measured crack
widths at the capacities predicted by the eighth edition
of the American Association of State Highway and
Transportation Officials’ AASHTO LRFD Bridge
Kent A. Harries, PhD, PEng, Design Specifications remained small. The require-
FASCE, FACI, FIIFC, is a ments for debonded strands were revised significantly
professor of structural engineering in the ninth edition of the AASTHO LRFD specifica-
and mechanics at the University of tions based on the presented research.
Pittsburgh in Pittsburgh, Pa.
Keywords
Richard A. Miller, PhD, FPCI, PE, Bridge, multiweb flange section, nonprestressed
is a professor of civil engineering longitudinal reinforcement, partially debonded strand,
and department head at the prestressed concrete girder, single-web flanged section,
University of Cincinnati. strand debonding.
Review policy
Henry G. Russell, PhD, PE, is an This paper was reviewed in accordance with the
engineering consultant who has Precast/Prestressed Concrete Institute’s peer-review
been involved with applications of process. The Precast/Prestressed Concrete Institute
concrete for bridges for about 50 is not responsible for statements made by authors of
years. He is a former managing papers in PCI Journal. No payment is offered.
technical editor of Aspire, the
concrete bridge magazine. Publishing details
William A. Potter, PE, is the This paper appears in PCI Journal (ISSN 0887-9672)
Florida state structures design V. 68, No. 2, March–April 2023, and can be found
engineer. at https://doi.org/10.15554/pcij68.2-01. PCI Journal
is published bimonthly by the Precast/Prestressed
Concrete Institute, 8770 W. Bryn Mawr Ave., Suite
1150, Chicago, IL 60631. Copyright © 2023, Precast/
Prestressed Concrete Institute.
Abstract
Reader comments
To develop a unified approach for the design of
partially debonded strands in prestressed concrete Please address any reader comments to PCI Journal edi-
highway bridge girders, a coordinated analytical tor-in-chief Tom Klemens at tklemens@pci.org or Precast/
and experimental investigation was conducted. This Prestressed Concrete Institute, c/o PCI Journal, 8770 W.
investigation examined amounts of partial debonding, Bryn Mawr Ave., Suite 1150, Chicago, IL 60631. J

Influence of structural form on
hydration-heat-induced temperature
rise of precast concrete lining segments
for a metro transit station
C
oncrete is a composite material composed of
cement, water, aggregate, and admixtures.1 During
the concrete casting process, the exothermic
chemical reaction between cement and water (hydration
of cement) releases a large amount of heat.2,3 The large
amount of hydration heat generated in a short time leads to
a rapid rise in the concrete temperature at its early age. The
high temperatures (greater than 70°C [158°F]) can give rise
to delayed ettringite formation, which could cause the con-
crete to expand and crack and, in turn, reduce its strength
and durability.4–6 Furthermore, in the interior of concrete,
heat is trapped because of the poor thermal conductivity
of concrete, whereas on the surface of concrete, heat is
released into the atmosphere.3,7 The trapped heat causes
the temperature of the interior to be relatively higher than
the surface temperature. The temperature gradient from
■ A metro transit station with multiring segmental the core to the surface leads to an internal restraint, which
lining was designed in Changchun, China. Because of results in compressive stresses in the interior and tensile
the size of the precast concrete lining segment, heat stresses in the surface.8,9 Consequently, thermal cracks
from hydration of cement on the segments could not could occur as the restraint-induced tensile stress in con-
be neglected. crete exceeds its tensile strength.
■ This paper describes a heat of hydration analysis of The American Concrete Institute (ACI)10 defines mass
a segment that was conducted using finite element concrete as “any volume of concrete with dimensions large
analysis to examine the ability of the closed-cavity enough to require that measures be taken to cope with gen-
and solid forms to control temperature increases. eration of heat from hydration of the cement and attendant
volume change to minimize cracking.” In mass concrete
■ The segment was numerically modeled using a structures, the nonuniform distribution of temperatures is
summer ambient temperature of 25°C (77°F) and a particularly apparent. To control or avoid the risk of thermal
winter ambient temperature of 10°C (50°F). cracking in mass concrete, its early-age temperatures must

be reduced.7,11 In recent decades, engineers and scientists in concrete and found that this technique was useful for reduc-
have made great efforts to reduce the internal temperature ing the maximum concrete temperature.
of concrete. Research has so far mainly focused on optimiz-
ing the raw materials in concrete (for example, developing Postcooling methods aim to lower the temperature of con-
low-heat cement12–14) and choosing casting methods (such as crete while it is hydrating. Postcooling is mainly performed
embedding cooling pipes7,15–18) to reduce the temperature in by circulating cold water in pipes embedded in concrete.7,15–18
mass concrete. In contrast, scant attention has been paid to Tasri and Susilawati7 concluded that cooling water tempera-
improving the structural form of mass concrete to avert ex- ture, space between cooling pipes, and coefficient of convec-
cessively high temperatures. This paper describes research on tions from cooling pipes to cooling water could be optimized
the effect of structural form on hydration-heat-related changes to obtain a temperature distribution that resulted in thermal
in precast concrete lining segments for a metro transit station stresses being less than the tensile strength of concrete. In
in Changchun, China, a location where conventional cast-in- addition, other methods, such as layered casting31–33 and
place construction for mass concrete structures is interrupted strengthened curing,34,35 have been adopted to reduce the tem-
by months of cold winter weather. perature of concrete and avoid the risk of thermal cracking.
Reducing concrete temperature Metro transit station model
There are two main types of methods of reducing the con- In China, urban rail transit has emerged as a growing form
crete temperature: precooling and postcooling methods.19,20 of transportation in recent decades as the nation has rapidly
Methods to precool concrete often involve the improvement urbanized. For example, the total length of the rail transit
of materials, the reduction of placing temperatures, or both network in Changchun, China, was 106.9 km (66.4 mi) in
material improvements and reduced placing temperatures. 2019 and is expected to expand to 235 km (146 mi) with
Material improvements involve adjusting the mixture propor- 174 stations by 2024.36 Notably, because Changchun is in
tions of ingredients in concrete or developing new materials to the middle temperate zone, it has been necessary to take a
reduce or inhibit the release of hydration heat. For example, four- to six-month winter break when conventional cast-in-
the content of cement may be reduced,21–23 low-heat cement place concrete techniques have been used for construction of
may be used,12–14 or hydration heat inhibitor may be added.24–26 metro transit stations. This winter break exerts considerable
Jia et al.26 incorporated a microcapsule sustained-release-type pressure on construction schedules. To shorten the construc-
hydration heat inhibitor into concrete and succeeded in tion time line for Metro Line 2 in Changchun, an innovative
decreasing its peak temperature by delaying the hydration of underground metro transit station design with fully precast
low-heat portland cement. The primary objective of reducing concrete components was applied.37,38 The station, which
placing temperature is to decelerate the hydration of cement was constructed with a cut-and-cover method, is mainly
by cooling concrete-forming components (cement, water, and composed of multiple rings of precast concrete lining seg-
aggregate) before they are mixed to form concrete.21,23,27–29 ments (Fig. 1). The dimensions and mass of these segments
Aniskin et al.30 used ice as a substitute for some of the water are detailed in Table 1. Given the large dimensions, these
Multiple-ring design Segments in one ring
Figure 1. Lining segments for the precast concrete metro station. Note: 1 m = 3.281 ft.

Table 1. Information for the precast concrete lining segments
Segment Length, m Width, m Thickness, m Cavity volume, m3 Mass, tonne
A 10.50 0.80 to 1.26 2.00 4.60 40.0
B 5.00 3.20 2.00 6.00 40.0
C 8.95 0.70 to 1.02 2.00 1.74 31.0
D 10.20 5.30 2.00 4.71 48.5
E 10.30 5.30 2.00 4.77 54.5
Note: 1 m = 3.281 ft; 1 m3 = 1.308 yd3; 1 tonne = 1.102 tons.
segments were considered mass concrete structures; there- T = temperature of the concrete
fore, the effect of hydration heat during their casting could
not be neglected.39 To avert excessively high temperatures, t = time
the design included closed cavities in the interior of the seg-
ments. The closed cavities were located at the neutral zone, λ = thermal conductivity of concrete
and the total volume proportion ranged from 15% to 22%.
Reinforcement design was sufficient to ensure that each c = specific heat of the concrete
segment met strength requirements.37
ρ = density of the concrete
The finite element method (FEM) can be an economical and
powerful tool for accurately predicting the development of θ(t) = adiabatic temperature rise due to the heat from
temperature and stress in mass concrete structures.3,29,40,41 hydration of cement
Using FEM, it is feasible to put forward effective measures
for controlling the temperature in mass concrete structures at The relation between θ(t) and t can be expressed by the fol-
the design stage. In the study described in this paper, FEM lowing formula:43
was used to numerically simulate the development of tem-
perature and stress in a precast concrete lining segment during θ(t) = θ0(1 – e-mt)
its production. The FEM model and selected parameters
were proved to be reasonable by comparing the temperature where
predictions with experimental temperature observations.
To investigate the effect of the closed cavity on controlling θ0 = the final value of the adiabatic temperature rise
the temperature rise, the segment was simulated in both
closed-cavity and solid forms for comparison. The results m = rate coefficient of the adiabatic temperature rise
from the finite element analysis were used to confirm whether
the closed-cavity form could effectively control the tempera- Initial condition
ture rise and the distributions of normal and shear stresses in
the precast concrete segment. Initial temperature, which is the average temperature of water,
cement, and aggregates at the time when concrete is placed, is
Theoretical foundations the initial condition for thermal analysis. The initial tem-
of thermal analysis perature is continuously distributed in the concrete, and the
concrete is assumed to be isothermal at the initial moment.44
This section introduces a differential equation of heat conduc-
tion, as well as initial and boundary conditions, to calculate Thermal boundary condition
the distribution of temperature in mass concrete.
To solve the differential equation governing heat conduction
Heat conduction equation within a body, it is necessary to apply thermal conditions
at the boundaries of the body. The three common boundary
Heat conduction is the transfer of heat within a body due to conditions are convection, specified temperature, and heat flux
temperature gradient. Given that concrete is isotropic and at the surface.43 Convection is a form of heat transfer whereby
homogeneous, the heat conduction within it can be governed heat is transmitted between a fluid (for example, water or air)
by the following equation:42 and the surface of a solid through a fluid’s relative molecular
motion. In this study, convection was applied to consider the
∂T λ 2 ∂θ (t) heat transmitted between the concrete surface and the atmo-
− ∇T=
∂t cρ ∂t sphere.9,42,45 The heat transfer due to convection is governed
where by the following:

qconv = h(T – Ta) three-dimensional FEMs in both solid and closed-cavity forms
are presented in Fig. 2. The reinforcing bars were ignored in
where the models because their interaction with the concrete was
beyond the scope of this work. The concrete mixture pro-
qconv = convective heat flux portions of the lining segments were designed with a 28-day
compressive strength of 50 MPa (7.25 ksi) (Grade C50) and
h = convection coefficient the mixture components were determined according to the
Chinese standard.46 Table 2 lists the main components of the
Ta = ambient temperature concrete mixture. This study used common portland cement
(Table 3).47 Table 4 presents the main model parameters for
Finite element method model the heat of hydration analysis. The thermal properties of the
concrete (thermal conductivity λ, specific heat c, convection
The heat of hydration analysis in the study was carried out coefficient h, final value of adiabatic temperature rise θ0, rate
with a commercial FEM simulator (Midas/FEA). Segment B coefficient of adiabatic temperature rise m, and coefficient of
in Fig. 1 was selected for the hydration heat analysis and its thermal expansion) were set to reasonable ranges from the
Table 2. Concrete mixture proportions Table 4. Input parameters for the heat of hydration
analysis
Component Dosage, kg/m3
Parameter Value
Cement 400
Fly ash 70 Thermal conductivity λ, kJ•m-1•hr-1•°C-1 10.0
Ultra-fine sand 326 Specific heat c, kJ•kg-1•°°C-1 0.95
Medium-coarse sand 490 Density r, kg/m3 2400
Gravel (5 to 10 mm) 379 Final value of adiabatic temperature rise θ0, °C 56
Gravel (10 to 25 mm) 598 Rate coefficient of adiabatic temperature rise m 0.45
Water 150 Convection coefficient h, kJ•m-2•hr-1•°C-1 50.4
Note: 1 mm = 0.0394 in.; 1 kg/m3 = 1.6875 lb/yd3. Initial placing temperature T0, °C 20
Ambient temperature in summer Tas, °C 25

Table 3. Portland cement mineral contents
Ambient temperature in winter Taw, °C 10
Ca3SiO5 Ca2SiO4 Ca3Al2O4 Ca4AlnFe2-nO7
Poisson’s ratio 0.2
Proportion, % 40.62 29.58 10.42 9.15
Coefficient of thermal expansion, °C-1 1e-5
Note: Al = aluminum; Ca = calcium; Fe = iron; O = oxygen; Si = silicon.
Note: 1 kg/m3 = 1.6875 lb/yd3; 1 kJ/kg = 0.4299 Btu/lb; 1 kJ/m = 0.2889;
1 kJ/m2 = 0.0881 Btu/ft2; °C = (°F – 32)1.8.
Solid form Closed-cavity form
Figure 2. Finite element method models of segment B.

literature on similar materials.48–52 Density ρ and Poisson’s specific structural form, the temperature of the segment cast
ratio were measured using the Chinese standard.46 The initial in the summer was higher than the temperature of the segment
temperature for placing concrete (the initial placing tempera- cast in the winter. In each of the ambient temperatures, the
ture T0) was approximately maintained at 20°C (68°F). The temperature distribution in the closed-cavity segment was
ambient temperature Ta was specified or fixed during the man- more uniform than that in the solid one. In both structural
ufacturing of the segments as either 25°C (77°F) for concrete forms and at both ambient temperatures, the maximum tem-
being cast in the summer or 10°C (50°F) for winter casting. peratures of the segment occurred at the core.
Model validation To observe the evolution and differences of temperature in

a segment, two points in each segment were designated to
To validate the FEMs, the numerical predictions of temperature measure the evolution of temperature (Fig. 4). Points P1 and P2
evolution in the closed-cavity segment B were compared with were located at the core and surface of the segments, respective-
corresponding experimental observations. Two points around ly, and were set at the same locations in each segment. Figure 5
the surface and core of the closed-cavity segment B were presents temperature history at the measurement points. Because
selected to measure the evolution of temperature after concrete of the heat from the hydration of the cement, the temperatures
was placed. The corresponding measurement points in the of all segments rapidly increased up to a peak, and then segment
numerical model and experimental observations were set at the temperatures decreased to the value of the ambient temperature
same locations. Figure 3 compares the evolution of tempera- as hydration came to a halt. For the summer and winter ambient
ture with time at the selected points in the numerical models temperatures at both measurement points, the maximum tem-
and experimental measurements. The numerical predictions perature of the closed-cavity segment was much lower than that
were approximately consistent with the experimental obser- of the solid one and the peak temperature was reached earlier in
vations, although the temperature in the numerical model was the closed-cavity segment. These findings can be attributed to
slightly higher. This good consistency between the numerical the lesser volume of concrete in the closed-cavity segment and
and experimental results indicated that the FEM and selected hence the lesser amount of hydration heat.
parameters were reasonable.
When the summer ambient temperature was 25°C (77°F), the
Results maximum temperatures at point P1 were 89.48°C (193.06°F)
in the solid segment and 58.97°C (138.15°F) in the closed-cav-
Figure 4 presents the temperature distributions in the ity segment. Also at an ambient temperature of 25°C, the
closed-cavity and solid forms of segment B as the temperature maximum temperatures at point P2 were 80.18°C (176.32°F) in
rise reached its maximum (termed maximum temperature dis- the solid segment and 59.90°C (139.82°F) in the closed-cavity
tributions herein) during 28 days of curing. Notably, for each segment. In other words, compared with the maximum tem-
Surface of closed-cavity segment B Core of closed-cavity segment B
Figure 3. Comparison between the experimental measurements and numerical predictions of time-temperature history. Note: °C
= (°F – 32)1.8.

Closed-cavity segment B at summer Solid segment B at summer
ambient temperature of 25°C ambient temperature of 25°C
Closed-cavity segment B at winter Solid segment B at winter ambient

ambient temperature of 10°C temperature of 10°C
Figure 4. Maximum concrete temperature distributions of segment B forms. Note: P1 = temperature measurement point 1; P2 =
temperature measurement point 2; T = temperature of concrete. °C = (°F – 32)1.8.
Summer ambient temperature of 25°C Winter ambient temperature of 10°C
Figure 5. Time-temperature history at measurement points (P1, P2) of segment B forms. Note: °C = (°F – 32)1.8.

perature in the solid segment, the maximum temperature in the segment, the maximum temperature difference between P1
closed-cavity segment was approximately 31.4% lower at point and P2 was about 5°C (41°F) when the ambient tempera-
P1 and 25.3% lower at point P2 in the summer scenario. ture was 10°C and almost negligible when the ambient
temperature was 25°C (77°F). The lower temperatures and
At the winter ambient temperature of 10°C (50°F) , the smaller temperature differences in the closed-cavity segment
maximum temperatures at point P1 were 72.12°C (161.82°F) indicated that the closed-cavity form was more effective in
in the solid segment and 37.49°C (99.48°F) in the closed-cav- suppressing the temperature rise, which lowered the risk of
ity segment. The maximum temperatures at point P2 were thermal cracking. ACI’s Guide to Mass Concrete10 says that
62.81°C (145.06°F) in the solid segment and 32.49°C to avoid the cracks caused by delayed ettringite formation and
(90.48°F) in the close-cavity segment. Thus, in the winter thermal stress, the maximum temperature and the temperature
scenario, the maximum temperatures of the closed-cavity difference inside the concrete must not exceed 70°C and 20°C
segment were approximately 48.0% lower at point P1 and (158°F and 68°F), respectively. For this study, the tempera-
48.3% lower at point P2 than maximum temperatures at the tures during casting of the closed-cavity segment completely
same points in the solid segment. satisfied the ACI requirements and the maximum temperature
of the solid segment exceeded the 70°C threshold for avoiding
Furthermore, at each of the ambient temperatures, the the risk of delayed ettringite formation.
maximum temperature differences between point P1 and point
P2 were smaller when the closed-cavity form was employed. Figures 6 and 7 present the distributions of normal and shear
In both the summer and winter scenarios, the maximum tem- stresses in the segments as the stresses reach a maximum.
perature difference between the two points was around 10°C (These stresses are termed maximum normal stress distribution
(50°F) in the solid segment. In contrast, in the closed-cavity and maximum shear stress distribution herein, respectively.)
Closed-cavity segment B at summer Solid segment B at summer ambient


Figure 6. Maximum normal stress distributions of segment B forms. Note: 1 MPa = 0.145 ksi; °C = (°F – 32)1.8.

Closed-cavity segment B at summer Solid segment B at summer ambient tem-
ambient temperature of 25°C perature of 25°C

Figure 7. Maximum shear stress distributions of segment B forms. Note: 1 MPa = 0.145 ksi; °C = (°F – 32)1.8.
For each specific structural form, the normal (shear) stress of The smaller stresses in the closed-cavity segment indicated
the segment cast in summer temperatures was larger than that greater effectiveness of the closed-cavity form to mitigate
in winter temperatures; also, at each ambient temperature, the the thermal stress and ensure the quality of the manufactured
stresses of the closed-cavity segment were smaller than those in segments.
the solid one.
Conclusion
At the summer ambient temperature of 25°C (77° F), the
maximum normal stress was 40.4% lower in the closed-cavity An innovative, fully prefabricated metro transit station with a
segment than in the solid segment (10.2 MPa [1.5 ksi] com- multiring segmental lining was designed to expedite construc-
pared with 17.1 MPa [2.5 ksi]). The maximum shear stress tion in Changchun, China, where cold winter weather stops
in the closed-cavity segment was 22.7 MPa (3.3 ksi), which conventional cast-in-place construction for several months. The
was 30.2% lower than the maximum shear stress of 32.5 MPa precast concrete lining segments with large dimensions were
(4.7 ksi) in the solid segment. mass concrete structures, so the effect of heat from hydration
of cement on the segments could not be neglected. To avert
At the winter ambient temperature of 10°C (50°F), the excessively high concrete temperatures due to hydration heat,
maximum normal stress was 8.1 MPa (1.2 ksi) in the closed-cav- which can lead to delayed ettringite formation– and thermal
ity segment and 15.7 MPa (2.3 ksi) in the solid segment (that is, stress–related cracking, a closed-cavity form was designed for
the maximum normal stress was 48.7% lower in the closed-cav- the segments. The segment was numerically modeled in both
ity segment). The maximum shear stress was 15.7 MPa in the closed-cavity and solid forms for comparison. Ambient tem-
closed-cavity segment and 37.5% lower than the maximum peratures of 25°C and 10°C (77°F and 50°F) were selected to
shear stress of 25.1 MPa (3.6 ksi) in the solid segment. simulate summer and winter production conditions, respectively.

The FEM and selected parameters were proved reasonable Furnace Slag.” CivilEng 2 (1): 254–271. https://doi.org
by the consistency between the numerical predictions and ex- /10.3390/civileng2010014.
perimental observations from the closed-cavity segment. The
results show that for both the closed-cavity and solid struc- 4. Myuran, K., N. S. A. Wanigaratne, and M. T. R.
tural forms, the temperature of the segment cast in summer Jayasinghe. 2015. “Strategies for Prevention of Delayed
conditions was higher than that in winter conditions. In both Ettringite Formation in Large Concrete Sections.”
of the ambient temperatures, the maximum temperature was Engineer: Journal of the Institution of Engineers, Sri
much lower in the closed-cavity segment than in the solid one Lanka 48 (2): 1–13. https://doi.org/10.4038/engineer
and the temperature distribution in the closed-cavity segment .v48i2.6829.
was more uniform. The lower temperature and more uniform
temperature distribution in the closed-cavity segment resulted 5. Lubej, S., I. Anžel, P. Jeluši , L. Kose , and A. Ivani , .
in less thermal stress in the closed-cavity segment, thereby 2016. “The Effect of Delayed Ettringite Formation on
lowering the probability of thermal cracking. Fine Grained Aerated Concrete Mechanical Properties.”
Science and Engineering of Composite Materials 23 (3):
During casting of the closed-cavity segment, the maximum 325–334. https://doi.org/10.1515/secm-2012-0107.
temperature and the difference in maximum temperatures
inside the concrete were kept below 70°C and 20°C (158°F 6. Kawabata ,Y., N. Ueda, T. Miura, and S. Multon. 2021.
and 68°F), respectively, fully satisfying the ACI requirements10 “The Influence of Restraint on the Expansion of Concrete
for avoiding cracks caused by delayed ettringite formation and due to Delayed Ettringite Formation.” Cement and
thermal stress. In contrast, the maximum temperature of the Concrete Composites, no. 121, 104062. https://doi.org
solid segment exceeded 70°C. Thus, compared with the solid /10.1016/j.cemconcomp.2021.104062.
segment, the closed-cavity segment can more effectively resist
cracking. With the closed-cavity segments and novel prefabri- 7. Tasri, A., and A. Susilawati. 2019. “Effect of Cooling
cation techniques, on which Yang and Lin53 have elaborated, six Water Temperature and Space between Cooling Pipes of
fully prefabricated metro transit stations have been completed Post-Cooling System on Temperature and Thermal Stress
in Changchun, China. The prefabrication of these segments in Mass Concrete.” Journal of Building Engineering, no.
in cold ambient temperatures has been proved to be effective 24, 100731. https://doi.org/10.1016/j.jobe.2019.100731.
in accelerating the construction of metro transit stations in the
cold regions. Moreover, this prefabricated construction could be 8. Jeon, S.-J. 2008. “Advanced Assessment of Cracking due
a promising approach to improve time-efficiency of other large- to Heat of Hydration and Internal Restraint.” Materials
scale underground construction projects, such as underground Journal 105 (4): 325–333. https://doi.org/10.14359
parking structures. /19893.
Acknowledgments 9. Soo Geun, K. 2010. “Effect of Heat Generation from

Cement Hydration on Mass Concrete Placement.” Master
The work discussed in this paper is supported by the Science of science thesis, Iowa State University, Ames. https://
and Technology Project on Developing the Xiong’an New doi.org/10.31274/etd-180810-763.
Area (grants 2021-07 and 2021-03).
10. ACI (American Concrete Institute) Committee 207. 2005.
References Guide to Mass Concrete. ACI 207.1R-05. Farmington
Hills, MI: ACI.
1. Gagg, C. R. 2014. “Cement and Concrete as an
Engineering Material: An Historic Appraisal and Case 11. Mizobuchi, T. 1998. “Discussion on the Experimental
Study Analysis.” Engineering Failure Analysis, no. 40, Evaluation of Reducing Effect of Thermal Stress of
114–140. https://doi.org/10.1016/j.engfailanal.2014 Expansive Additive Based on Uniaxial Restraint Testing
.02.004. Device.” JCI Conference 20 (2):1051–1056.
2. Ng, P. L., J. S. Du, X. F. Luo, and F. T. K. Au. 2015. 12. Huang, Y., G. Liu G, S. Huang, R. Rao, and C. Hu. 2018.
“Hydration Temperature Rise and Thermal Stresses “Experimental and Finite Element Investigations on
Induced in Segment-on-Pier of Prestressed Concrete the Temperature Field of a Massive Bridge Pier Caused
Box Girder Bridge.” In Proceedings of the Second by the Hydration Heat of Concrete.” Construction and
International Conference on Performance-based and Building Materials, no. 192, 240–252. https://doi.org/10
Life-cycle Structural Engineering, 679–688. Brisbane, .1016/j.conbuildmat.2018.10.128.
Australia: School of Civil Engineering, University of
Queensland. https://doi.org/10.14264/uql.2016.1185. 13. Xin, J., G. Zhang, Y. Liu, Z. Wang, N. Yang, Y. Wang, R.
Mou, Y. Qiao, J. Wang, and Z. Wu. 2020. “Environmental
3. Leon, G., and H.-L. Chen. 2021. “Thermal Analysis of Impact and Thermal Cracking Resistance of Low
Mass Concrete Containing Ground Granulated Blast Heat Cement (LHC) and Moderate Heat Cement

(MHC) Concrete at Early Ages.” Journal of Building “Implementing Ternary Supplementary Cementing
Engineering, no. 32, 101668. https://doi.org/10.1016/j Binder for Reduction of the Heat of Hydration of
.jobe.2020.101668. Concrete.” Journal of Cleaner Production, no. 112,
845–852. https://doi.org/10.1016/j.jclepro.2015.06.022.
14. Lee, M. H., B. S. Khil, and H. D. Yun. 2014. “Influence
of Cement Type on Heat of Hydration and Temperature 25. Makul, N., and G. Sua-iam. 2018. “Effect of Granular
Rise of the Mass Concrete.” Indian Journal of Urea on the Properties of Self-Consolidating Concrete
Engineering and Materials Sciences 21 (5): 536–542. Incorporating Untreated Rice Husk Ash: Flowability,
Compressive Strength and Temperature Rise.
15. Tasri, A., and A. Susilawati. 2019. “Effect of Material of Construction and Building Materials, no. 162, 489–502.
Post-Cooling Pipes on Temperature and Thermal Stress https://doi.org/10.1016/j.conbuildmat.2017.12.023.
in Mass Concrete. Structures, no. 20, 204–212. https://doi
.org/10.1016/j.istruc.2019.03.015. 26. Jia, F., Y. Yao, and J. Wang. 2021. “Influence and
Mechanism Research of Hydration Heat Inhibitor on
16. Kim, J. K., K. H. Kim, and J. K. Yang. 2001. “Thermal Low-Heat Portland Cement.” Frontiers Materials, no. 8,
Analysis of Hydration Heat in Concrete Structures with 697380. https://doi.org/10.3389/fmats.2021.697380.
Pipe-Cooling System.” Computers and Structures 79 (2):
163–171. https://doi.org/10.1016/S0045-7949(00)00128-0. 27. Chen, Y.L., C. J. Wang, S. Y. Li, and L. J. Chen. 2003.
“The Effect of Construction Designs on Temperature
17. Zuo, Z., Y. Hu, Q. Li, and L. Zhang. 2014. “Data Field of a Roller Compacted Concrete Dam—A
Mining of the Thermal Performance of Cool-Pipes in Simulation Analysis by a Finite Element Method.”
Massive Concrete via In Situ Monitoring.” Mathematical Canadian Journal of Civil Engineering 30 (6): 1153–
Problems in Engineering 2014 (5): 1–15. https://doi 1156. https://doi.org/10.1139/l03-076.
.org/10.1155/2014/985659 and errata https://doi.org
/10.1155/2014/608179. 28. Zhou, H., Y. Zhou, C. Zhao, and Z. Liang. 2016.
“Optimization of the Temperature Control Scheme for
18. Liu, K., Z. Wang, C. Jin, F. Wang, and X. Lu. 2015. “An Roller Compacted Concrete Dams Based on Finite
Experimental Study on Thermal Conductivity of Iron Element and Sensitivity Analysis Methods.” Civil
Ore Sand Cement Mortar.” Construction and Building Engineering Journal 25 (3): 14. https://doi.org/10.14311
Materials 101 (1): 932–941. https://doi.org/10.1016/j /CEJ.2016.03.0014.
.conbuildmat.2015.10.108.
29. Sargam, Y., M. Faytarouni, K. Riding, K. Wang,
19. ACI Committee 207. 2005. Cooling and Insulating C. Jahren, and J. Shen. 2019. “Predicting Thermal
Systems for Mass Concrete. ACI 207.4R-05 (Reapproved Performance of a Mass Concrete Foundation—A Field
2012). Farmington Hills, MI: ACI. Monitoring Case Study.” Case Studies in Construction
Materials, no. 11, e00289. https://doi.org/10.1016/j
20. KCI (Korea Concrete Institute). 2009. Standard .cscm.2019.e00289.
Specification for Concrete. [In Korean.] Seoul, South
Korea: KCI. 30. Aniskin, N., T. Chuc Nguyen, and A. Kiet Bui. 2021.
“The Use of Ice to Cool the Concrete Mix in the
21. Lee, M. H., B. S. Khil, and H. D. Yun. 2013. “Thermal Construction of Massive Structures.” E3S Web of
Analysis of Hydration Heat in Mass Concrete with Conferences, no. 264, 02047. https://doi.org/10.1051
Different Cement Binder Proportions.” Applied /e3sconf/202126402047.
Mechanics and Materials, no. 372, 199–202. https://doi
.org/10.4028/www.scientific.net/AMM.372.199. 31. Chen, J., Z. Quan, and X. Peng. 2016. “Simulation
Analysis on Mass Concrete Temperature Field of Lock
22. Hu, J., Z. Ge, and K. Wang. 2014. “Influence of Cement Floor Layered Pouring.” Journal of Environmental
Fineness and Water-to-Cement Ratio on Mortar Early- Science and Engineering 5 (9): 476–483. https://doi.org/
Age Heat of Hydration and Set Times.” Construction and 10.17265/2162-5298/2016.09.006.
Building Materials, no. 50, 657–663. https://doi.org/10
.1016/j.conbuildmat.2013.10.011. 32. Han, S. 2020. “Assessment of Curing Schemes for
Effectively Controlling Thermal Behavior of Mass
23. Lagundžija, S., and M. Thiam. 2017. “Temperature Concrete Foundation at Early Ages.” Construction and
Reduction during Concrete Hydration in Massive Building Materials, no. 230, 117004. https://doi.org/10
Structures.” Master’s thesis, KTH Royal Institute of .1016/j.conbuildmat.2019.117004.
Technology, Stockholm, Sweden.
33. Le, H.-H., C.-C. Vu, N.-K. Ho, and V.-T. Luu. 2020. “A
24. Yang, K.-H., G.-D. Moon, and Y.-S. Jeon. 2016. Method of Controlling Thermal Crack for Mass Concrete

Structures: Modelling and Experimental Study.” IOP “Effect of Thermal Parameters on Hydration Heat
Conference Series: Materials Science and Engineering Temperature and Thermal Stress of Mass Concrete.”
869 (7): 072054. https://doi.org/10.1088/1757-899X Advances in Materials Science and Engineering, vol.
/869/7/072054. 2021, 1–16. https://doi.org/10.1155/2021/5541181.
34. Wang, X., Q. Chen, J. Tao, R. Han, X. Ding, F. Xing, 43. Zhu, B. 2014. Thermal Stresses and Temperature Control
and N. Han. 2019. “Concrete Thermal Stress Analysis of Mass Concrete. Boston, MA: Elsevier/Butterworth-
during Tunnel Construction.” Advances in Mechanical Heinemann.
Engineering 11 (6): 168781401985223. https://doi.org/10
.1177/1687814019852232 44. Zeng, H., C. Lu, L. Zhang, T. Yang, M. Jin, Y. Ma, and
J. Liu. 2022. “Prediction of Temperature Distribution in
35. Jin, W., L. Jiang, L. Han, L. Chen, X. Yan, and C. Concrete under Variable Environmental Factors through
Chen. 2021. “Influence of Curing Temperature on the a Three-Dimensional Heat Transfer Model.” Materials
Mechanical Properties and Microstructure of Limestone 15 (4): 1510. https://doi.org/10.3390/ma15041510.
Powder Mass Concrete.” Structural Concrete 22 (S1):
E745–E755. https://doi.org/10.1002/suco.201900549 45. Lee, Y., M.-S. Choi, S.-T. Yi, and J.-K. Kim. 2009.
“Experimental Study on the Convective Heat Transfer
36. China’s National Development and Reform Coefficient of Early-Age Concrete.” Cement and
Commission. Accessed January 17, 2023. “Reply to The Concrete Composites 31 (1): 60–71. https://doi.org/10
Third Phase Construction Planning of Changchun Urban .1016/j.cemconcomp.2008.09.009.
Rail Transit (2019–2024).” [In Chinese.] https://www
.ndrc.gov.cn/fggz/zcssfz/zcgh/201812/t20181228 46. China Academy of Building Research. 2011.
_1145781.html. Specification for Mix Proportion Design of Ordinary
Concrete. [In Chinese.] Beijing, China: Ministry of
37. Yang, X., and Y. Han. 2017. “Closed Cavity Thin-Wall Housing and Urban-Rural Construction of the People’s
Components Design for Prefabricated Underground Republic of China.
Subway Structures.” In Geo-Risk 2017, 194–205.
Denver, CO: American Society of Civil Engineers. 47. Xiong, Z., P. Wang, and Y. Wang. 2015. “Hydration
https://doi.org/10.1061/9780784480700.019. Behaviors of Portland Cement with Different Lithologic
Stone Powders.” International Journal of Concrete
38. Tao, L., P. Ding, C. Shi, X. Wu, S. Wu, and S. Li. Structures and Materials 9 (1): 55–60. https://doi.org
2019. “Shaking Table Test on Seismic Response /10.1007/s40069-014-0086-z.
Characteristics of Prefabricated Subway Station
Structure.” Tunneling and Underground Space 48. Wang, H. L., and X. D. Zhang. 2015. “Study on the
Technology, no. 91, 102994. https://doi.org/10.1016/j Effect of Fly Ash Mix Amount in Large Volume
.tust.2019.102994. Pile Cap Concrete on Heat of Hydration.” Applied
Mechanics and Materials, vol. 744–746: 832–836.
39. Central Research Institute of Building and Construction https://doi.org/10.4028/www.scientific.net/AMM.744
Co. 2018. Standard for Construction of Mass Concrete. -746.832.
[In Chinese.] Beijing, China: Ministry of Housing and
Urban-Rural Construction of the People’s Republic of 49. Aniskin, N., C. Nguyen Trong, and L. Hoang Quoc.
China. 2018. “Influence of Size and Construction Schedule
of Massive Concrete Structures on Its Temperature
40. Lawrence, A.M., M. Tia, C. C. Ferraro, and M. Bergin. Regime.” MATEC Web of Conferences, no. 251, 02014.
2012. “Effect of Early Age Strength on Cracking in https://doi.org/10.1051/matecconf/201825102014.
Mass Concrete Containing Different Supplementary
Cementitious Materials: Experimental and Finite- 50. Qiao, G., G. Xue, and G. Zhu. 2022. “A Precise
Element Investigation.” Journal of Materials in Civil Internal-Stress Calculation Algorithm for the Evolution
Engineering 24 (4): 362–372. https://doi.org/10.1061 of Cracking Risk of the Steam-Curing Box Girder.”
/(ASCE)MT.1943-5533.0000389. Construction and Building Materials, no. 346, 128463.
https://doi.org/10.1016/j.conbuildmat.2022.128463.
41. Yikici, T. A., and H.-L. Chen. 2015. “Numerical
Prediction Model for Temperature Development in Mass 51. Liu, L., S. Yu, W. Xu, and Z. Wang. 2021. “Study
Concrete Structures.” Transportation Research Record: on Arrangement of Cooling Water Pipe to Control
Journal of the Transportation Research Board 2508 (1): Hydration Heat of Concrete in Beam of Cable-Stayed
102–110. https://doi.org/10.3141/2508-13. Bridge.” Stavební Obzor—Civil Engineering Journal 30
(2): 525–534. https://doi.org/10.14311/CEJ.2021
42. Zhao, Y., G. Li, C. Fan, W. Pang, and Y. Wang. 2021. .02.0039.

52. Xie, Y., C. Qian, Y. Xu, M. Wei, W. Du. 2022. “Effect
of Fine Aggregate Type on Early-Age Performance,
Cracking Analysis and Engineering Applications of C50
Concrete.” Construction and Building Materials, no. 323,
126633. https://doi.org/10.1016/j.conbuildmat.2022
.126633.
53. Yang, X., and F. Lin. 2021. “Prefabrication Technology

for Underground Metro Station Structure.” Tunnelling
and Underground Space Technology, no. 108, 103717.
https://doi.org/10.1016/j.tust.2020.103717.
Notation
c = specific heat of the concrete
h = convection coefficient
m = rate coefficient of adiabatic temperature rise
qconv = convective heat flux
t = time
T = temperature of the concrete
Ta = ambient temperature
Tas = ambient temperature in summer
Taw = ambient temperature in winter
T0 = initial placing temperature
θ0 = final value of adiabatic temperature rise
θ(t) = adiabatic temperature rise due to the heat from

hydration of cement
λ = thermal conductivity of concrete
ρ = density of the concrete

ty and durability, and the ability to prefabricate these
segments in cold ambient temperatures effectively
Yuzhen
<Body>Mohamed
Han, PhD, K.is aNafadi,
structural Body text construction on this project.
accelerated
PhD, is an assistant
engineering managerprofessor
from Beijing
of
structural
Urban Construction
engineeringDesign
at Assiut
and Keywords
University in Assiut,
Development Group Co.
Egypt.
Ltd.He
inis
a formerChina.
Beijing, graduate research FEM, text
Body finite element method, heat of hydration, lining
assistant in the Department of segment, prefabricated metro transit station, structural
Civil, Construction, and form, transit.policy
Review
Environmental
Lei Zhang, PhD, Engineering
is a geotechnical
at
North Carolina
engineer from Beijing
State University
Urban Body text policy
Review
(NCSU) in Raleigh. Construction Group Co. Ltd.
Reader comments
This paper was reviewed in accordance with the
Omar M. Khalafalla, is a graduate Precast/Prestressed Concrete Institute’s peer-review
research and teaching assistant Body textThe Precast/Prestressed Concrete Institute
process.
and PhD candidate in the is not responsible for statements made by authors of
Department
Jizhong He isofaCivil,
geotechnical
Construction, papers in PCI Journal. No payment is offered.
and Environmental
engineer from Beijing
Engineering
Urban at
NCSU.
Construction Design and Publishing details
Development Group Co. Ltd.
Gregory W. Lucier, PhD, is a This paper appears in PCI Journal (ISSN 0887-9672)
research assistant professor in the V. 68, No. 2, March–April 2023, and can be found
Department of Civil, Construction, at https://doi.org/10.15554/pcij68.2-04. PCI Journal
and Environmental Engineering is published bimonthly by the Precast/Prestressed
and manager of the Constructed Concrete Institute, 8770 W. Bryn Mawr Ave., Suite
Abstract Facilities Laboratory at NCSU. 1150, Chicago, IL 60631. Copyright © 2023, Precast/
Prestressed Concrete Institute.
This paper describes Sami H. Rizkalla,
research PhD,ofFPCI,
on the effect structural
FACI, FASCE, FIIFC, FEIC,
form on hydration-heat-related changes in large-scale Reader comments
precast concrete liningFCSCE, is Distinguished
segments. These segments of
mass concrete were used in theofconstruction
Professor Civil Engineering and
of an un- Please address any reader comments to PCI Journal
derground metro transitConstruction,
station in Changchun,
director of theChina, editor-in-chief Tom Klemens at tklemens@pci.org or
a location where conventional
Constructed cast-in-place
Facilities Laboratory,
construc- Precast/Prestressed Concrete Institute, c/o PCI Journal,
tion is interrupted forand
months
director
by cold
of thewinter
National
weather. 8770 W. Bryn Mawr Ave., Suite 1150, Chicago, IL
Heat of hydration analysis
Science ofFoundation
closed-cavity Center
and solid
on 60631. J
lining segments was Integration
conducted using the finite element
of Composites into
method (FEM). Two Infrastructure
ambient temperatures
at NCSU.of 25°C
and 10°C (77°F and 50°F) were selected to simulate
segment manufacturing Paulconditions in summer
Z. Zia, PhD, PE, FPCI,andis a
winter, respectively. The FEM and selected
Distinguished Universityparameters
were found to be reasonable
Professor because thein
Emeritus numerical
the
predictions and experimental
Department observations from the
of Civil, Construction,
closed-cavity segment andwere consistent. At
Environmental both of theat
Engineering
ambient temperatures, the maximum internal concrete
NCSU.
temperature of the closed-cavity segment was much
lower than that in the solid one and the temperature
distribution in the closed-cavity
Gary J. Klein,segment
PE, iswas more
executive
uniform. As a result, vice
thermal stress and
president wassenior
lower principal
in the
closed-cavity segment forthan in Janney,
Wiss, the solidElstner
segment, which
reduced the risk for cracking.
Associates The closed-cavity
Inc. in Northbrook,formIll.
of the lining segments enhanced the structural integri-

Directories
Board of Directors
Matt Ballain, Chair, Coreslab Troy Jenkins, Institute Program Patty Peterson, Institute Program
Structures (INDIANAPOLIS) Director, Transportation Activities, Director, Business Performance,
Jim Fabinski, Vice Chair, EnCon Northeast Prestressed Products Tindall Corp.
Colorado Lloyd Kennedy, Institute Program Richard Potts, Producer Member
Carlos Cerna, Secretary-Treasurer, Director, Educational Activities, Director, Georgia/Carolinas PCI,
Manco Structures Ltd. Finfrock Industries LLC Standard Concrete Products Inc.
J. Matt DeVoss, Immediate Past Brent Koch, Producer Member Jim Renda, Associate Member
Chair, Jackson Precast Inc. Director, PCI West, Con-Fab Director, Supplier, Cresset Chemical
Bob Risser, President and CEO, PCI California LLC Co.
Cheryl Lang, At-Large Member, Tindall Lenny Salvo, Producer Member
Dusty Andrews, Producer Member Corp. Director, Florida Chapter, Coreslab
Director, PCI Washington/Oregon, Matt Mahonski, Producer Member Structures (ORLANDO) Inc.
Knife River Corp.–Northwest Director, PCI Central, High Concrete Peter Simoneau, Producer Member
Dennis Cilley, Associate Member Group Director, PCI Northeast, Dailey
Director, Erector, American Steel & David Malaer, Producer Member Precast
Precast Erectors Director, Texas, Oklahoma, New Ben Spruill, Producer Member
Ned Cleland, Professional Member Mexico, Valley Prestress Products Inc. Director, PCI Gulf South, Gulf Coast
Director, Blue Ridge Design Inc. Jane Martin, Institute Program Director, Prestress Partners
Todd Culp, Producer Member Marketing, Gate Precast Co. Kimberly Wacker, At-Large Member,
Director, PCI Midwest, Coreslab Brian Miller, Associate Member Wells
Structures (OMAHA) Inc. Director, Supplier, GCP Applied Lee Wegner, Producer Member
Brandon Farley, Producer Member Technologies Inc. Director, PCI Mountain States,
Director, PCI Midwest, EnCon Field Richard Miller, Institute Program Forterra Structural Precast
Services LLC Director, Technical Activities, Mike Wolff, Institute Program
Evan Fink, Producer Member University of Cincinnati Director, Quality Activities,
Director, PCI Mid-Atlantic, Chris Mosley, Professional Member Mid-States Concrete Industries LLC
Northeast Prestressed Products Director, The Consulting Engineers Diep Tu, Regional Council
LLC Group Inc. Representative, Nonvoting, Florida
Matt Graf, Producer Member Director, Andrew Osborn, Institute Prestressed Concrete Association
PCI Illinois/Wisconsin, International Program Director, Research and Colin Van Kampen, Membership
Concrete Products Inc. Development, Wiss, Janney, Elstner Council Representative, Nonvoting,
Associates Inc. Structures Inc.
Technical Activities Council

Chair Harry Gleich Dusty Andrews Alex Mihaylov
Vice Chair Rich Miller Suzanne Aultman Barry McKinley
Secretary Jared Brewe Ned M. Cleland Christopher Mosley
Ex-officio, fib Representative Mary Ann Griggas-Smith Pinar Okumus
Larbi Sennour David Jablonsky Andrew Osborn
Ex-officio, Code Representatives Wayne Kassian Timothy Salmons
S. K. Ghosh and Stephen V. Skalko Yahya Kurama Stephen J. Seguirant
John Lawler Venkatesh Seshappa

PCI staff
Tom Bagsarian (312) 428-4945 tbagsarian@pci.org Editorial content manager
Laura Bedolla (312) 360-3218 lbedolla@pci.org Technical activities program manager
Lauren Bell (312) 583-6775 lbell@pci.org Education and publications coordinator
Jared Brewe (312) 360-3213 jbrewe@pci.org Technical services vice president
Trina Brown (312) 360-3590 tbrown@pci.org Transportation systems program manager
K. Michelle Burgess (312) 282-8160 mburgess@pci.org PCI Journal managing editor
Nikole Clow (312) 360-3202 nclow@pci.org Marketing coordinator
Royce Covington (312) 428-4946 rcovington@pci.org Member services manager
Timothy Cullen (312) 360-3206 tcullen@pci.org Technical activities director
Cher Doherty (312) 583-6781 cdoherty@pci.org Events director
Walter Furie (312) 583-6772 wfurie@pci.org Production senior specialist
Christopher Hurst (312) 360-3203 churst@pci.org Membership director
Cody Kauhl (312) 583-6778 ckauhl@pci.org Web developer
Michael Kesselmayer (312) 583-6770 mkesselmayer@pci.org Quality programs managing director

Market development and education managing
Becky King (312) 360-3201 bking@pci.org
director
Tom Klemens (312) 583-6773 tklemens@pci.org Publications director
Ken Kwilinski (312) 428-4944 kkwilinski@pci.org Quality systems manager
Carolina Lopez (312) 583-6774 clopez@pci.org Certification programs coordinator

Membership and administrative services
Philip McConnell (312) 583-6783 pmcconnell@pci.org
coordinator
John McConvill (312) 360-3208 jmcconvill@pci.org Controller
Noel Morales (312) 786-0300 nmorales@pci.org Certification coordinator
William Nickas (312) 583-6776 wnickas@pci.org Transportation systems managing director
Bob Risser (312) 360-3204 brisser@pci.org President and chief executive officer
Lisa Scacco (312) 583-6782 lscacco@pci.org Publications manager
Neal Sherman (312) 786-0300 nsherman@pci.org Senior staff accountant
Edith Smith (312) 360-3219 esmith@pci.org Codes and standards managing director
Mike Smith (312) 786-0300 msmith@pci.org Information technology manager
Beth Taylor (312) 583-6780 btaylor@pci.org Chief financial and administrative officer
Trice Turner (312) 583-6784 tturner@pci.org Business development manager

Information technology and events
Cindi Ward (312) 360-3214 cward@pci.org
coordinator
Randy Wilson (312) 428-4940 rwilson@pci.org Architectural precast systems director

Regional offices
Florida Prestressed Concrete PCI of Illinois & Wisconsin PCI Northeast
Association Joe Lombard Rita Seraderian
Diep Tu Phone: (312) 505-1858 Phone: (617) 484-0506
Phone: (407) 758-9966 Email: joe@pci-iw.org Email: contact@pcine.org
Email: diep@myfpca.org PCI-IW.org PCINE.org
MyFPCA.org
PCI Mid-Atlantic PCI West
Georgia/Carolinas PCI Dawn Decker Ruth Lehmann
Ray Clark Phone: (717) 682-1215 Phone: (949) 420-3638
Phone: (678) 402-7727 Email: dawn@pci-ma.org Email: info@pciwest.org
Email: ray.clark@gcpci.org PCI-MA.org PCIWest.org
GCPCI.org
PCI Midwest Precast Concrete Manufacturers’
PCI Central Region Mike Johnsrud Association
Phil Wiedemann Phone: (952) 806-9997 Chris Lechner
Phone: (937) 833-3900 Email: mike@pcimidwest.org Phone: (866) 944-7262
Email: phil@pci-central.org PCIMidwest.org Email: chris@precastcma.org
PCI-Central.org PrecastCMA.org
PCI Mountain States
PCI Gulf South Jim Schneider
Dan Eckenrode Phone: (303) 562-8685
Phone: (228) 239-3409 Email: jschneider@pcims.org
Email: pcigulfsouth1@att.net PCIMS.org
PCIGulfSouth.org
Coming ahead
Bridges
• Evaluating fatigue, relaxation, and creep rupture of
CFRP strands for highway bridge construction
• Analytical investigation of UHPC in deck bulb tee
girder connections
Also
• Numerical modeling of inverted U-shaped connectors
to enhance the performance of composite beams
• A novel reinforced-concrete buckling-restrained brace
for precast concrete lateral-load-resisting frames

Meet Nancy Peterson
Concrete harmony
Sarah Fister Gale
A s a child, Nancy Peterson loved two

things: math and playing the violin.
And she excelled at both. So when it came
Stadium in Tampa Bay, Fla., and the FTX Arena in Miami,
Fla. She also designed several prisons, including a supermax
penitentiary in Florence, Colo., and a landmark IKEA store
time for college, she faced a conundrum. for which she won a Women of Influence in Commercial Real
She wanted to pursue a career in both fields Estate award.
and even declined an offer from the School Throughout her career, Peterson has relied on PCI to pro-
of Mines in Colorado because they had no vide guidance and camaraderie as she navigated the industry.
music program. She joined PCI and attended her first convention in 1996.
Eventually, she landed at Colorado University in Boulder as “It was a little bit overwhelming,” she says. But then her boss,
an engineering major. “I realized I couldn’t major in music and Craig Barrett at Leap, introduced her to Jim Voss and she felt
do engineering on the side,” she says. She played in the Boulder immediately at home. “I remember thinking that I needed to get
Philharmonic Orchestra while in school and is currently a mem- to know these guys because they were so much fun,” she says.
ber of the Jefferson Symphony Orchestra. In 2000, she joined the Concrete Chefs, PCI members
Peterson completed her degree at Metropolitan State who host and cook for an annual fundraising dinner for the
University of Denver in 1988 and quickly landed a job at PCI Foundation, where she contributes her signature jamba-
Rockwell International. The following year, she and her hus- laya. She later joined and eventually chaired the Soundwall
band of three weeks relocated to the San Francisco Bay area of Committee, where she was the principal author of the “Guide
California where they both saw better job prospects. Specification for Reflective Precast Concrete Soundwalls.”
Within four days of arriving, the city was hit by the Later she became a member of PCI’s board of directors, and
magnitude 6.9 Loma Prieta earthquake during the World she is now on the PCI Foundation Board of Trustees.
Series. “It was terrifying,” she says, but it also led to her next In her current role, she is focused on getting the next genera-
job. Peterson spent the next two years working for Brewster tion of engineers, architects, and project managers excited about
Consulting Engineers retrofitting wood frame homes that had the precast concrete industry and what the material can do.
fallen off their foundations during the earthquake. It was inter- “This younger generation with their innovations can see things
esting work, but eventually the cost of living in San Francisco with new eyes,” she says. “It’s exciting to watch them work.”
got to be too much to bear, so they moved back to Denver. It’s also necessary to keep the precast concrete industry grow-
That’s when she met Jim Sirko, of Sirko Associates, who ing and thriving. Peterson says that engineering and architecture
introduced her to precast concrete. Before that, Peterson had students are still not being exposed to precast concrete in their
known virtually nothing about the material. She recalls one college studies, and she wants the foundation to change that.
professor dedicating only half of one lecture to prestressed con- “We are giving grants to schools to develop curriculum for pre-
crete. Once Sirko explained the material and how it worked to cast programs, and we are reaping benefits from it,” she says.
Peterson, though, she was hooked. A big part of this process is educating students about the
“You have to be very detail oriented, which is one of the safety, durability, and energy efficiency that precast concrete
things I love about it,” she says. “I love the precision and the can bring to a project, especially as climate crises continue to
tight tolerances. You’re striving for perfection all the time.” expand. “Whether it’s hurricanes or forest fires, we can design
Peterson spent the rest of her career working in precast things with precast that will save a lot of lives,” she says.
concrete. After two years with Sirko Associates, she spent a She says that any PCI member who ever had to educate an
decade with Leap Associates, then held positions with FDG architect about precast concrete or search for an engineering
Precast Engineering and Rocky Mountain Prestress before applicant with real knowledge of the industry should support
landing at Encon United. the foundation’s work. “We want to foster relationships with
Over the years, Peterson helped to design and expand mul- young architects, engineers, and construction managers to keep
tiple stadiums across the country, including the Jack Murphy the industry growing,” she says. “It’s a great way to give back to
Stadium in San Diego, Calif., and both the Raymond James the industry that has given us so much.”

28dbb993 4

Uploaded by

Copyright:

Available Formats

28dbb993 4

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

28dbb993 4

Uploaded by

Copyright:

Available Formats

MARCH–APRIL 2023

VOLUME 68, NUMBER 2

Precast/Prestressed Concrete Institute

Your Connection Connection

Investigation of Repair Techniques for Deteriorated End Regions 52

Experimental Background behind New AASHTO LRFD Specifications 71

71 Influence of Structural Form on Hydration-Heat-Induced Temperature Rise 88

New ACI 440.11 Code Adopted for Design of Concrete Reinforced 22

PCI Journal | March–April 2023

the downtown area. Building

on the high thermal mass 01_Cover_MA23.indd 1 2/14/23 12:47 PM

of concrete, the structure’s PCI Calendar 11

PCI Directories 101

PCI Staff Directory 102

Regional Offices 103

EDITORIAL DESIGN & PRODUCTION

This paper is milled from a 3rd-party certified source

PCI Journal | March–April 2023

The PCI Design Awards program showcases winning

VISIT PCI.ORG/DESIGNAWARDS FOR MORE INFORMATION

PCI Journal | March–April 2023 5

The Precast/Prestressed Concrete Institute (PCI) certifications are the industry’s

CERT21x40057_Certification_Halfpage_Ad_2022.indd 1 4/19/22 2:35 PM

It’s An Art Installation,

And, It Required Intricate Forms.

Project: Gov. Mario M. Cuomo Bridge

Our Role: Hamilton Form created the

Hamilton Form Company For more than 55 years,

The next class

PCI Journal | March–April 2023 7

PCI cancels ICC-ES Dixon, Fink added to

I CC-ES evaluation report ESR-3010 for “precast concrete

2023 T. Henry Clark Award call for nominations

8 PCI Journal | March–April 2023

T he National Science Foundation

Welcome the students

PCI Journal | March–April 2023 9

P CI West and the PCI Foundation hosted a ribbon cutting

The Carapace Pavilion, designed by University of Southern California

2023 Sidney Freedman Craftsmanship Award call for entries

Sarah Fister Gale

10 PCI Journal | March–April 2023

2023 Norman L. Scott Professional

PCI Journal | March–April 2023 11

PCI personnel training and certification schools

Compiled by K. Michelle Burgess (mburgess@pci.org) J

Guidelines for the Use of Ultra-High-Performance

• batching and placement considerations for production

and routine quality assurance.

PCI Con cret e M

12 PCI Journal | March–April 2023

S pencer Kubat has retired from his

ACI Center of Excellence

N EU: An ACI Center of Excellence

PCI Journal | March–April 2023 13

Charles E. “Budd” Hilgeman Jr.

14 PCI Journal | March–April 2023

Sylaj named CPCI president Wells promotes Kloos