Parking Structures: Precast/Prestressed Concrete Institute

SEPTEMBER–OCTOBER 2021
VOLUME 66, NUMBER 5
Precast/Prestressed Concrete Institute
Parking Structures
23 39 54
Behavior of CFRP- Hollow-core wall Seismic design
reinforced precast, panel influence of precast concrete
prestressed concrete on steel framing buckling-
sandwich panels system cyclic restrained
behavior braced frames
Precast manufacturers
have an injury problem.
SAVE LIVES
PCI and Optimum Safety Management Average 57% reduction
have partnered to bring you the solution: in TRIR in first year!
The PCI Safety Culture

SAVE MONEY
Improvement Process™ Reduction in EMR by 42%
and $337,000 savings in
Get the knowledge to lead the charge, the less than two years!
action plan to follow, and the educational
support for your team to make it all stick.
DECREASE TURNOVER
Recruit and retain top talent
as “the firm that truly values
safety”
See how it works:

www.optimumsafetymanagement.com/SCIP
Scan with your
phone camera
Connections…
Always a Bet!
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
Parking Structures
Long-Term Behavior of Precast, Prestressed Concrete Sandwich Panels 23
Reinforced with Carbon-Fiber-Reinforced Polymer Shear Grid
Mohamed K. Nafadi, Gregory Lucier, Tugce Sevil Yaman, Harry Gleich, and Sami Rizkalla
23 Influence of Hollow-Core Wall Panels on the Cyclic Behavior

of Different Types of Steel Framing Systems
39
Parsa Monfaredi, Mehdi Nazarpour, and Abdoreza S. Moghadam
Seismic Design and Analysis of Precast Concrete 54

Buckling-Restrained Braced Frames
Shane Oh, Yahya C. Kurama, Jon Mohle, and Brandt W. Saxey
39
Index of advertisers
ALP Supply..........................................18–19 JVI ....................................................................1
alpsupply.com jvi-inc.com
Burgess Pigment .....................................16 Midwest Structure Engineering ........14
burgesspigment.com midweststructure.com
CEG.............................. Inside Back Cover Optimum Safety Management
cegengineers.com (SCIP) ................... Inside Front Cover
Endicott ....................................................... 6 optimumsafetymanagement.com/
endicott.com SCIP
Hamilton Form ......................Back Cover Prestress Supply Inc. .............................. 4
hamiltonform.com prestresssupply.com
High Concrete Accessories ................14 Tucker’s........................................................ 8
highconcreteaccessories.com tuckerbilt.com
PCI Journal | September–October 2021

2
SEPTEMBER–OCTOBER 2021 • VOLUME 66, NUMBER 5
September–OctOber 2021
VOlume 66, Number 5
Departments
On the cover
Precast/Prestressed Concrete Institute
Chairman’s Message 5
A five-level precast concrete
underground parking Get Involved Again
structure for the new Penn
Pavilion in Philadelphia, Pa., Parking Structures
President’s Message 7
supports structural loads PCI Foundation at 20: Making
from the steel-framed tower
Headway on our ‘Problem’
being constructed above.
23 39 54
Courtesy of the Shockey Behavior of CFRP-
reinforced precast,
prestressed concrete
Hollow-core wall Seismic design
panel influence of precast concrete
on steel framing buckling-
From PCI Headquarters 9

sandwich panels system cyclic restrained
Precast Group.
behavior braced frames
PCI Calendar 13
Our Members 15
In the News 17
Industry Calendar 17
Project Spotlight 21
PCI Directories 85
Board of Directors and
Technical Activities Council 85
PCI Staff Directory 86
Regional Offices 87
JOURNAL ADVISORY COMMITTEE Millard J. Barney Pinar Okumus Coming Ahead 87

Chair Richard Alan Miller Amir Fam Arturo Ernest Schultz
Vice Chair vacant Alexander G. Mihaylov Sri Sritharan Meet Mary Ann Griggas-Smith 88
Secretary Collin Moriarty Clay Naito
Staff Liaison Tom Klemens
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 © 2021, 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.
Rory Cleveland Copy 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
Elizabeth Nishiura Copy Editor international, and $80 per year and $200 for three years for electronic-only subscriptions anywhere in the world. A single or
Laura Vidale Copy Editor back issue is $15. International subscriptions are delivered by an international carrier; allow one to three weeks.
Laura Bedolla Technical Activities Program Manager Postmaster: Please send address changes to PCI Journal, 8770 W. Bryn Mawr Ave., Suite 1150, Chicago, IL 60631.
Periodicals postage rates paid at Chicago and additional mailing offices.
This paper is milled from a 3rd-party certified source
PCI Journal | September–October 2021

3
Chairman’s Message
Get involved
again
I hope everyone has had an enjoyable and productive summer. It seems as if I have mentioned
in every Chairman’s Message I’ve written that we can be hopeful that a federal transportation
bill is just around the corner. Three seasons have passed now, and still no bill, but I choose to
remain hopeful that our elected representatives can do what is needed to maintain and improve
our country’s aging infrastructure.
I encourage members to attend the PCI Committee Days in Rosemont, Ill., in September in
person. If the PCI Convention at The Precast Show was an indication, we are expecting atten-
dance to be high. Come to Chicago and get involved in the committees of interest to you and
help guide the institute and the industry forward. There is always room for new faces and fresh
ideas. You will find the folks sitting on the committees very welcoming and enthusiastic to share
the work they do.
Here is some more good news. The Productivity Tour is back on again! It is scheduled for
October 19 to 21 in Charlotte, N.C. After several postponements, expectations are great for a
highly attended and educational tour. The tour will kick off with a quality summit created by the
Quality Enhancement Committee. The Productivity Tour itself is planned and supported by the
Productivity Committee, which is near and dear to my heart because I was a chair for that com-
mittee and my son Andrew Fink is the current chair. It is gratifying to see these two committees
working together with the PCI staff to provide real value and learning opportunities for our
industry associates.
It remains a core value at PCI to foster strong relationships with our industry partners.
During the PCI summer Board of Directors meeting in Alexandria, Va., we met with the leaders
of the National Precast Concrete Association to discuss and plan The Precast Show and review
our Joint Concrete Products List. In addition, the PCI Executive Committee will attend the
ACI Fall Convention in October in Atlanta, Ga., and meet with ACI’s leadership to discuss our
common interests and strategies.
I am approximately at the midterm of my year as PCI chair. By the time this is printed, I will
have attended at least six chapter or affiliate meetings, with the intention of getting to at least
one meeting in all regions before my term expires. The hospitality and willingness to share best
practices and ideas have been overwhelming in a good way. Thank you!
I look forward to seeing you at Committee Days in September. J
Dennis R. Fink
2021 PCI Board Chair
President
Northeast Prestressed Products LLC
Cressona, Pa.
PCI Journal | September–October 2021 5

President’s Message
PCI Foundation at 20: Making

headway on our ‘problem’
“Y ou have a problem!” For those who were at the awards luncheon at the 2019 PCI
Convention in Louisville, Ky., we will never forget University of Southern California
(USC) professor Doug Noble’s passionate speech about how we, as an industry, have a problem
with architecture. The problem, he said, is that despite attending and then teaching at some of
the finest architecture schools, he never learned about precast concrete. That is, until the PCI
Foundation teamed up with California producers and PCI West to start a studio at USC.
In 2001, a group of precast concrete industry visionaries led by Jim Voss, president and gen-
eral manager of JVI Inc., recognized the problem and the need to step up the industry’s efforts to
educate the next generation of students about precast concrete and started the PCI Foundation.
Twenty years later, those efforts are paying off and continue to grow at a fantastic rate.
Early on, the PCI Foundation focused on raising money to give scholarships to students in
architecture and engineering. Then in 2007, PCI Foundation leadership learned about a program
at the University of Wisconsin–Milwaukee. PCI member Spancrete, whose headquarters was in
the Milwaukee suburbs, had partnered with architecture faculty to develop a hands-on, semes-
ter-long design studio centered on precast concrete and its design possibilities. With that, the
current era of the PCI Foundation’s support of architecture studios and immersive programs was
born.
Today there is PCI Foundation–supported curriculum in almost every PCI region, many of
them considering the continuation of the program at universities in their areas a key element of
their strategic plans. To date, with 1920 weeks of precast concrete instruction and more than
$3.4 million invested, the PCI Foundation has reached 35 universities, 6000 students, and 50
professors teaching precast concrete. The key to the success of these programs continues to be the
local PCI producers’ strong involvement and their presentations, planning curriculum with the
professor, plant tours, competition judging, and donations of time and materials.
And it’s working. On the PCI Foundation website, you’ll find dozens of testimonials and
examples of students armed with precast concrete knowledge looking for precast concrete solu-
tions for projects after graduation.
Several years ago, the PCI Board of Directors implemented a program where members can
donate to the Foundation as part of their annual membership renewal. I’m happy to report that
through this process, PCI members have pledged more than $73,000 for this year. If you donat-
ed, thank you. If you’d like to, visit the PCI Foundation website at PCI-Foundation.org or text
PRECAST to 41-444.
I hope to see many of you at the upcoming PCI Committee Days meetings. If you attend,
please consider joining the Concrete Chefs to continue supporting the work of the PCI
Foundation and to celebrate the foundation’s 20th anniversary. See you in Chicago! J
Bob Risser, PE
PCI President and CEO

From PCI Headquarters
New PCI white paper Fifth edition of PCI quality

scrutinizes need for sprinklers control manual released
in open parking structures
T he fifth edition of the Manual for Quality Control for
Plants and Production of Structural Precast Concrete
A new PCI white paper lays out the institute’s view

on requiring sprinklers in open parking structures.
Although such requirements were recently added to the
Products, MNL 116-21, was released in July and is now avail-
able in the PCI Bookstore in either PDF or printed format.
A PDF copy of the new manual was provided to
2021 International Fire Code (IFC) and International the authorized audit contacts at all plants in July.
Building Code, this review, authored by the PCI Fire Implementation of the new requirements is scheduled to be
Committee, finds that the underlying documentation for effective January 1, 2022. In advance of that implementation
the addition is inadequate. date, information describing these changes will be provided
This white paper provides the historical background on to certified plants and other stakeholders through various
fire experience in open parking structures, a summary of the publications and newsletters. A webinar is planned for early
documentation used to require sprinkler protection for open October. Watch for details in My PCI Update and the
parking structures in the 2021 IFC, a review of the potential monthly Education and Certification newsletters.
fire hazards in open parking structures, and an explanation of
why this added expense for the construction of open parking
structures is not justified. PCI Foundation-sponsored
Download the white paper Why Sprinklers Should Not Be
Required in Open Parking Structures free of charge at https:// barbecue competition returns
doi.org/10.15554/wp-21-01.
Brown added as new

A fter a hiatus in 2020 due to the pan-
demic, the PCI Foundation BBQ
Competition has returned.
Any company, association or individual
transportation systems in the precast concrete industry can join
the fun by holding a barbecue between
program manager August 1 and December 31, 2021. By
using peer-to-peer fundraising (similar to
P CI welcomes Trina Brown, as the new

PCI transportation systems program
manager. She comes to PCI from LTCI
a walkathon), each team will have its own
fundraising webpage where donations can be made in honor
of the team’s effort.
Partners LLC, where she was director of This year, the first $50,000 raised by this competi-
business development. tion will fund student participation at the 2022 PCI
Brown works with William Nickas, Convention at The Precast Show in Kansas City, Mo.,
transportation systems managing director, Trina Brown including a student welcoming party, special student educa-
to ensure that PCI is efficiently utilizing tional sessions, precast concrete presentations, and student
resources. She provides periodic transporta- and professor travel grants.
tion account and activity updates to the PCI Transportation The grand prize, a top-quality industrial grill or smoker val-
Activities Council (TrAC) leadership. She also works as a ued at $2500, goes to the team that raises the most money for
staff liaison and assists volunteer members with committee the PCI Foundation. Other prizes may be given in categories
and subcommittee activities. In addition, she manages Aspire such as most unusual menu item, family and friends award,
and activities with the National Concrete Bridge Council. student attendees award, most meat grilled, most creative
Brown graduated from Florida State University in fundraiser, or best event photo. Participants can register at
Tallahassee with a degree in international finance. http://pci-foundation.org.

Welcoming students to PCI
As we head back to school Student feedback is so beneficial, and they
this fall, the PCI Foundation is have told us that the lessons they learn in the
celebrating its 20th anniversary. precast studios go beyond the hard and fast rules
With four new schools coming on of designing a beam or creating a color palette.
board, we now have 35 schools By including industry partners and mentors in
in the PCI Foundation family and the programs, the PCI Foundation studios help
have worked alongside thousands students learn soft skills, collaboration, leader-
of students in our 14 years of of- Chris Pastorius ship, strategic thinking, and human relations. The
fering grants. PCI Foundation precast studio helps students understand deci-
We measure the success of the Chair sions and constraints in design, not just because
past 20 years in several ways, but of the material, but because of the partnerships
for me, it always goes back to the students. How we forge.
many are we reaching? How is our work affecting This year we proudly welcome four new schools
their educational outcome? The news could not be to our family. For the first time we are including
better. More than 6000 students have taken part in concrete industry management programs in our
a PCI Foundation studio in one form or another. As studio work, reaching out to a new type of student.
the years go on, paths of PCI members and studio In addition, new industry partners are emerging
participants are starting to cross more frequently. and will assist in helping these studios succeed.
Universities Locations Programs Professors
Chico State University

and Kansas State Uni- • Engineering
versity (joint program, Chico, Calif., and • Construction Mohammed T. Albahttiti
sponsored in partnership Manhattan, Kans. management Kimberly W. Kramer
with the National Precast • Concrete industry Hayder A. Rasheed
Concrete Association management
Foundation)
University of Delaware Newark, Del. • Engineering Jovan Tatar
• Engineering George Morcous

University of Nebraska Lincoln, Neb. • Construction Marc Maguire
management
• Engineering
New Jersey Institute of Newark, N.J. • Architecture Mohamed Mahgoub
Technology • Concrete industry
management
The transition of certified architectural plants to the new

New Architectural categories began with the publication of the Architectural
Certification Program Supplemental Requirements in
Certification Program December 2019 and certified architectural plants submitting
applications for their desired new categories during the first
categories effective October 1 quarter of 2020. Since that time, each plant has been required
to construct a set of three mock-up panels unique to their cat-
T he transition to the new architectural certification cate-

gories will be completed on October 1, 2021. All certified
plants that have met the requirements for the new categories
egory, which have been reviewed by the plant auditors during
on-site audits.
COVID-19 pandemic concerns affected the ability of the
will receive a new certificate reflecting their new category with auditors to visit many plants, and virtual audits were per-
an effective date of October 1, 2021. formed in those cases. The evaluation of the panels for confor-
10 PCI Journal | September–October 2021

mance to the acceptability of appearance standards necessitates requirements matrix, FAQ documents, and a brochure, which
an on-site review; therefore, many plants were unable to com- are all available on PCI’s website at https://www.pci.org
plete the panel evaluation. In addition, COVID-19 affected /archcert. Educational webinars are also being presented,
the workforce at many plants and prevented or delayed the including a series of four monthly webinars from June through
construction of the mock-up panels. September 2021, which will be available in the education sec-
Due to the pandemic’s impact on the plants, the tion of PCI.org. Revisions to the PCI and MasterSpec guide
Architectural Certification Committee made the decision in specs that will incorporate the new categories are also under-
November 2020 to delay the implementation of the new cate- way and are planned to be released in September 2021.
gories from the original July 1 date to October 1. This extension The official transition to the new categories will be com-
allowed plants that were experiencing restricted access an addi- pleted with the assignment of the categories on October 1.
tional opportunity for an on-site plant audit. The audit schedule Full transition will take much longer. All projects bidding on
for the third quarter of 2021 was structured to focus especially or after October 1, 2021, will need to reference the new cate-
on those plants where on-site access had been restricted. gories. As such, a full transition to the new categories will take
In support of the new program, PCI has developed var- several years as category A1 projects are completed and plants
ious informational resources, including the Architectural begin to produce projects with the new categories.
Certification Program Supplemental Requirements, a category This program is designed to advance the precast concrete
industry by promoting collaboration between our certified
2021 Professors Seminar attendees tour a parking structure at plants and customers early in the design process and help
Sacramento State University. Courtesy of PCI Foundation. establish realistic quality and appearance expectations.
Professors Seminar attracts

record number of professors
T he 2021 PCI Foundation Professors Seminar was

held June 2 through 4 in Sacramento, Calif., with
the assistance of PCI West, Sacramento State, and Clark
Pacific. Originally scheduled to be held in June 2020 on the
Sacramento State campus, the program was postponed by a
year due to the pandemic.
Eric Matsumoto helped develop a robust multiday agenda,
which included local project tours, a Clark Pacific plant tour,
and multiple presentations by precast concrete experts.
This was the sixth Professors Seminar and attracted the
largest number of professors to date. There were 20 profes-
sors attending in person, and one third of them were profes-
sors who had not previously submitted a grant to the PCI
Foundation. In addition the event included about 10 industry
friends, local studio students from Sacramento State and PCI
West. Networking included a dinner hosted by Jim Voss.
Next year the Professors Seminar will be hosted by
Kennesaw State University in Atlanta, Ga., and will run May
18 through 20, 2022.
PCI releases new certificate

template for certified plants
P CI-certified plants will notice some changes to their plant

certificates for 2021. During PCI’s recent audit of the
plant certification program by the International Accreditation
Service (IAS), IAS required several changes to the template

that PCI uses for its certificate. The most recognizable change PCI Foundation Board
is the separate listing of dates for PCI Policy 20 certification,
which is an annual certification, and the ISO 9001:2015 adds two trustees
certification, which is a three-year certification with an annu-
al renewal. These are now presented in separate columns.
Reference was also added, at the bottom of the certificate, to
the full description of the PCI certification categories on the
T wo industry leaders, Lloyd Kennedy
Jr. and Matt DeVoss, were recently
welcomed to the PCI Foundation’s Board
PCI website. Please contact qualityprograms@pci.org if you of Trustees.
have any questions about these changes. Kennedy is Finfrock’s executive vice pres-
ident of engineering, and he not only brings
to the board decades of industry experience
Revisions to PCI’s but he is a champion of Finfrock’s recruiting
Lloyd Kennedy Jr.
efforts to attract talent to the company. He is

Erector Certification based in Apopka, Fla.
DeVoss is president of Jackson Precast,
Program approved based in Jackson, Miss. He attended Tulane
University in New Orleans, La., where
R evisions to PCI Policy 29 on the Erector Certification

Program were approved by PCI’s Board of Directors at
its June meeting. In addition to a minor reorganization of the
he completed an undergraduate degree in
civil engineering as well as obtaining a law
degree. He is a member of PCI’s executive Matt DeVoss
document and general maintenance revisions to match existing committee and holds contractor’s licenses in
procedures, several new requirements and sections have been Mississippi, Alabama, and Arkansas.
added. A copy of the revised policy has been emailed to all cer-
tified and applicant erectors. If you have questions regarding
these changes or would like further information, please contact
qualityprograms@pci.org.
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 continuing education senior manager, Sherrie Nauden, at
snauden@pci.org or (312) 360-3215. Registration forms are available at https://www.pci.org/qc_schools.
October 11–13, 2021 Orlando, Fla.
Level I/II
February 7–10, 2022 online
Level III October 13–16, 2021 Orlando, Fla.
CFA September 7–9, 2021 Nashville, Tenn.
CCA September 10, 2021 Nashville, Tenn.
Compiled by K. Michelle Burgess (mburgess@pci.org) J

PCI’s Calendar
Events
PCI event details are subject to change. For the most current information,
visit https://www.pci.org/events.
2021 PCI Northeast Annual Meeting
September 1–2, 2021
Manchester, Vt.
Ultra-High-Performance Concrete Workshop
September 21, 2021
Loews Chicago O’Hare Hotel, Rosemont, Ill.
2021 PCI Committee Days
2021 PCI Productivity Tour
October 19–21, 2021
Charlotte, N.C.
Marketing and Technical Activities Council Meetings
January 5–7, 2022
The Biltmore Miami-Coral Gables, Coral Cables, Fla.
2022 PCI Convention at The Precast Show
March 1–5, 2022
Kansas City, Mo.
2022 PCI Productivity Tour
May 9–11, 2022
TradeWinds Island Grand Resort, St. Petersburg, Fla.
PCI Board of Directors and Committee Meetings
June 7–10, 2022
Westin New Orleans, New Orleans, La.
2022 PCI Committee Days and Technical Conference
Statement of Ownership, Management, and Circulation 13. Publication Title 14. Issue Date for Circulation Data Below Statement of Ownership, Management, and Circulation
PCI Journal
(All Periodicals Publications Except Requester Publications) (All Periodicals Publications Except Requester Publications)
1. Publication Title 2. Publication Number 3. Filing Date
July-August 2021 16. Electronic Copy Circulation Average No. Copies No. Copies of Single
PCI JOURNAL 6 6 _ 1 6 9 0 8/16/21 15. Extent and Nature of Circulation Average No. Copies No. Copies of Single Each Issue During Issue Published
Each Issue During Issue Published Preceding 12 Months Nearest to Filing Date
4. Issue Frequency 5. Number of Issues Published Annually 6. Annual Subscription Price Preceding 12 Months Nearest to Filing Date
bi-monthly six US $80, Foreign $170 a. Total Number of Copies (Net press run)
2338 2397
a. Paid Electronic Copies
7. Complete Mailing Address of Known Office of Publication (Not printer) (Street, city, county, state, and ZIP+4 ®) Contact Person
Thomas Klemens b. Total Paid Print Copies (Line 15c) + Paid Electronic Copies (Line 16a)
Precast/Prestressed Concrete Institute, 8770 W. Bryn Mawr Ave., Suite 1150, (1) Mailed Outside-County Paid Subscriptions Stated on PS Form 3541 (Include paid
Chicago, IL 60631-3517
Telephone (Include area code) distribution above nominal rate, advertiser’s proof copies, and exchange copies) 2004 2097 c. Total Print Distribution (Line 15f) + Paid Electronic Copies (Line 16a)
312-786-0300 b. Paid
8. Complete Mailing Address of Headquarters or General Business Office of Publisher (Not printer) Circulation Mailed In-County Paid Subscriptions Stated on PS Form 3541 (Include paid
(2)
(By Mail distribution above nominal rate, advertiser’s proof copies, and exchange copies)
d. Percent Paid (Both Print & Electronic Copies) (16b divided by 16c Í 100)
and
Precast/Prestressed Concrete Institute, 8770 W. Bryn Mawr Ave., Suite 1150, Chicago, IL 60631-3517 Outside Paid Distribution Outside the Mails Including Sales Through Dealers and Carriers,
the Mail) (3)
Street Vendors, Counter Sales, and Other Paid Distribution Outside USPS® I certify that 50% of all my distributed copies (electronic and print) are paid above a nominal price.
9. Full Names and Complete Mailing Addresses of Publisher, Editor, and Managing Editor (Do not leave blank)
Publisher (Name and complete mailing address)
Paid Distribution by Other Classes of Mail Through the USPS
(4) 17. Publication of Statement of Ownership
(e.g., First-Class Mail®)
Robert Risser, Precast/Prestressed Concrete Institute, 8770 W. Bryn Mawr Ave., Suite 1150, Chicago, IL 60631-3517
If the publication is a general publication, publication of this statement is required. Will be printed Publication not required.
Editor (Name and complete mailing address)
c. Total Paid Distribution [Sum of 15b (1), (2), (3), and (4)]
2004 2097 September-October 2021 issue of this publication.
in the ________________________
d. Free or (1) Free or Nominal Rate Outside-County Copies included on PS Form 3541
Thomas Klemens Precast/Prestressed Concrete Institute, 8770 W. Bryn Mawr Ave., Suite 1150, Chicago, IL 60631-3517 Nominal 18. Signature and Title of Editor, Publisher, Business Manager, or Owner Date
Rate
Managing Editor (Name and complete mailing address) Distribution (2) Free or Nominal Rate In-County Copies Included on PS Form 3541
(By Mail
and
Outside Free or Nominal Rate Copies Mailed at Other Classes Through the USPS
K. Michelle Burgess, Precast/Prestressed Concrete Institute, 8770 W. Bryn Mawr Ave., Suite 1150, Chicago, IL 60631-3517 (3)
(e.g., First-Class Mail)
the Mail)
I certify that all information furnished on this form is true and complete. I understand that anyone who furnishes false or misleading information on this form
10. Owner (Do not leave blank. If the publication is owned by a corporation, give the name and address of the corporation immediately followed by the or who omits material or information requested on the form may be subject to criminal sanctions (including fines and imprisonment) and/or civil sanctions
names and addresses of all stockholders owning or holding 1 percent or more of the total amount of stock. If not owned by a corporation, give the (4) Free or Nominal Rate Distribution Outside the Mail (Carriers or other means)
100 100 (including civil penalties).
names and addresses of the individual owners. If owned by a partnership or other unincorporated firm, give its name and address as well as those of
each individual owner. If the publication is published by a nonprofit organization, give its name and address.)
Full Name Complete Mailing Address
e. Total Free or Nominal Rate Distribution (Sum of 15d (1), (2), (3) and (4))
100 100
Precast/Prestressed Concrete Institute 8770 W. Bryn Mawr Ave., Suite 1150, Chicago, IL 60631-3517
f. Total Distribution (Sum of 15c and 15e)
2104 2197
g. Copies not Distributed (See Instructions to Publishers #4 (page #3))
234 200
h. Total (Sum of 15f and g)
2338 2397
i. Percent Paid
11. Known Bondholders, Mortgagees, and Other Security Holders Owning or Holding 1 Percent or More of Total Amount of Bonds, Mortgages, or
(15c divided by 15f times 100) 95 95
Other Securities. If none, check box None * If you are claiming electronic copies, go to line 16 on page 3. If you are not claiming electronic copies, skip to line 17 on page 3.
Full Name Complete Mailing Address
12. Tax Status (For completion by nonprofit organizations authorized to mail at nonprofit rates) (Check one)
The purpose, function, and nonprofit status of this organization and the exempt status for federal income tax purposes:
Has Not Changed During Preceding 12 Months
Has Changed During Preceding 12 Months (Publisher must submit explanation of change with this statement)
PS Form 3526, July 2014 [Page 1 of 4 (see instructions page 4)] PSN: 7530-01-000-9931 PRIVACY NOTICE: See our privacy policy on www.usps.com. PS Form 3526, July 2014 (Page 3 of 4) PRIVACY NOTICE: See our privacy policy on www.usps.com.

Put a Lid On It
800.508.2583
www.highconcreteaccessories.com
Our Members
Metromont promotes Aultman mentor for young engineers. She is heav-

ily involved in PCI and the American
to vice president of engineering Concrete Association and standards devel-
opment. Most recently, she was recognized
M etromont Corp. has promoted Suzanne Aultman from

technical manager to vice president of engineering for
its Corporate Engineering Department in Greenville, S.C.
as 2021 Engineer of the Year for the
Piedmont Chapter of the South Carolina
Society of Professional Engineers.
Aultman has been with Metromont for 18 years, mak- —Source: Metromont Corp. Suzanne Aultman
ing her mark as a source for all things engineering and a
Welcome to PCI!
Installers Supplier associates
Peska Construction Abrasives Inc.
2700 N. Fourth St. 4090 Highway 49
Sioux Falls, SD 57104 Glen Ullin, ND 58631
PeskaConstruction.com AbrasivesInc.com
(605) 334-0173 (701) 348-3610
Primary contact: William Goeken Primary contact: Russell Raad
b.goeken@peskaconstruction.com rraad@abrasivesinc.com
United Steel American Progress Group Corp.

164 School St. 150 N. Michigan Ave., 35th Floor
East Hartford, CT 06108 Chicago, IL 60601
UnitedSteel.com (312) 973-2559
(860) 289-2323 Primary contact: Mark Harrison
Primary contact: Steve Bean harrison@americanprogressgroup.com
sbean@unitedsteel.com
Insteel Wire Products
Service associates 1373 Boggs Drive
Girder Slab Technologies Mount Airy, NC 27030
10 E. Yanonali St., #22 InSteel.com
Santa Barbara, CA 93101 (336) 719-9000
Girder-Slab.com Primary contact: Randy Plitt
(888) 478-1100 rplitt@insteel.com
Primary contact: Daniel Connelly
dconnelly@girder-slab.com Supporting producer
Arabian Tiles Co.—ARTIC
Precast Design Services P.O. Box 143 Riyadh 11383
20315 151st St. NW Riyadh, KSA
Elk River, MN 55330 artic.com.sa
Precast-Design.com 00966-114567714
(612) 405-3142 Primary contact: Adel Al Haj
Primary contact: Chad Howsden adel.alhaj@artic.com.sa
chad@pds.mn

PCI’s newly certified plants and erectors
PCI recently certified the following plants and erectors. For an explanation of the certification desig-
nations, visit http://www.pci.org/Plant_Certification and http://www.pci.org/Erector_Certification.
• Alberici Constructors in Saint Louis, Mo.: S2
• Carolina Stone Setting Co. Inc. in Cary, N.C.: A, S2
• County Materials Corp. in Whitestown, Ind.: C2
• CXT Inc. in Nampa, Idaho: C1, C1A
• Oldcastle Infrastructure Inc. in Fontana, Calif.: C2
• Top Flight Steel Inc. in Rhome, Tex.: S2
Metakaolin
I T could only come from
a leader in functional
mineral development like
OPTIPOZZ is white
so there is no dark
undertone in cementi- Stronger.
end-use in the cementi-
tious industry.
OPTIPOZZ : the
flash-calcined high reactiv-
ity metakaolin for high
Burgess. tious products. Users can OPTIPOZZ is manu-
performance concrete and
expect uniformity of color factured under the strict-
cement based products.
OPTIPOZZ, the from mix to mix. est of process controls to
high-reactivity, flash- assure product uniformity. From a world leader
calcined metakaolin offer- OPTIPOZZ finishes in functional mineral
effortlesssly, giving a OPTIPOZZ is easy to
Harder.
ing superior performance technology since 1948 —
creamy texture and handle. Supplied in powder
properties compared to Burgess.
improving trowelability. form, it arrives to you in
traditional pozzolans. your choice of paper bags, Contact Burgess by phone
A mix design containing
semi-bulk bags, bulk trucks at (478) 552-2544, toll free
OPTIPOZZ cures fast, OPTIPOZZ produces
or bulk railcars. The low at 800-841-8999. Or visit
giving cement-based prod- less bleed water so slabs
dusting characeristics of the Burgess web site at
ucts significant compressive may be finished sooner.
OPTIPOZZ allow for www.burgesspigment.com.
and flexural strength early
OPTIPOZZ is ease of dispersion in the
in the curing process. Or write: Burgess Pigment
Faster.
consistent. It is not an mix design. Clean-up is
Precast producers can Company, PO Box 349,
industrial by-product but a snap.
expect to turn forms more Sandersville, GA 31082.
is processed specifically for
quickly.
ISO 9001
email: info@optipozz.com

In The News
ABC Collaborative report on lication as a key technology and approach for manufacturing
and assembly.
decarbonizing U.S. buildings The ABC Collaborative is a network of building construc-
tion, real estate, and development stakeholders, with support
highlights precast concrete from the U.S. Department of Energy, that has the goal of
accelerating the uptake of innovative, high-performance con-
T he Advanced Building Construction (ABC)

Collaborative has released a research report titled Market
Opportunities and Challenges for Decarbonizing US Buildings
struction technologies that achieve superior energy and carbon
performance, enable rapid on-site construction time lines, are
affordable to building owners and developers, and are desirable
that provides an assessment of possibilities and barriers for to building owners and users.
transforming the national buildings sector with advanced To download the report, go to https://advancedbuilding
building construction (ABC). Within the report, precast construction.org/decarbonizing-us-buildings/.
concrete is highlighted for its prefabrication and quality rep- —Source: ABC Collaborative
Industry Calendar
Event details are subject to change.

fib International Conference on Concrete Sustainability 2021
Czech Technical University, Prague, Czech Republic
ACI Fall 2021 Convention
October 17–21, 2021
Hilton Atlanta Downtown, Atlanta, Ga.
World of Concrete 2022
January 18–21, 2022
Las Vegas Convention Center, Las Vegas, Nev.
ACI Spring 2022 Convention
March 27–31, 2022
Caribe Royale Orlando, Orlando, Fla.
Post-Tensioning Institute 2022 Convention
April 24–27, 2022
Hilton La Jolla Torrey Pines, La Jolla, Calif.
2022 fib International Congress
June 12–16, 2022
Oslo, Norway
October 23–27, 2022
Hyatt Regency Dallas, Dallas, Tex.
April 2–6, 2023
Hilton San Francisco Union Square, San Francisco, Calif.
BEI-2023 “Sustainability in Bridge Engineering”
Summer 2023
National University of Singapore, Singapore
October 29–
Boston Convention Center and Westin Boston Waterfront,
November 2, 2023
Boston, Mass.
March 24–28, 2024
Hyatt Regency New Orleans, New Orleans, La.

SAVE THE DATE
MARCH 1-5, 2022
KANSAS CITY CONVENTION CENTER | KANSAS CITY, MISSOURI
pci.org/convention | #PCIConvention
Project Spotlight
Pavilion parking structure

succeeds through innovation
A new five-level underground precast concrete parking struc-

ture supports half of the new Penn Pavilion, a 17-story,
1.5 million ft2 (139,000 m2) health center in the heart of
Philadelphia, Pa., that includes 500 private patient rooms, 47
operating rooms, and a 61-room emergency department.
The parking structure, which was constructed 67 ft
(20.4 m) underground, includes 689 parking spaces, two stair
towers, a series of elevator shafts, and two precast concrete cis-
terns as well as provisions for fuel tanks, a network, locksmith,
materials management, and information technology.
Although such an underground parking structure typically
would have been constructed using cast-in-place concrete,
the engineers realized that the accelerated time line offered
by using precast concrete made that the ideal solution for the
More than 80% of the labor for Shockey Precast’s work on Penn
Pavilion’s parking structure took place remotely at the company’s facil-
ity, helping to mitigate site congestion in this densely developed West
Philadelphia, Pa., neighborhood. Photo: Hover Solutions LLC
project. Shockey Precast, based in Winchester, Va., was select-

ed as the precaster.
Impressively, Shockey provided an innovative twist to
the structure’s design: cast-in-place concrete emulation with
precast concrete elements. As a result, and to everyone’s sat-
isfaction, the structure was designed and detailed in such a
way that it meets the requirements of the applicable building
codes, as if it were constructed of monolithic cast-in-place
reinforced concrete.
In addition, to keep costs down and increase the speed of
the project, the components of the parking structure were
divided into a number of structural elements of sizes and
shapes that could be plant-fabricated, transported, and then
safely and efficiently erected on-site.
One of the reasons timeliness and efficiency were so
important on this project was that construction of the new
Precast concrete enabled a five-level underground parking structure 1.5 million ft2 (139,000 m2) steel-framed patient pavilion
for the new Penn Pavilion in Philadelphia, Pa., to support structural could not begin until the five-level underground parking struc-
loads from the steel-framed tower being constructed above without ture was complete.
the delay that would have been required using cast-in-place concrete. The use of precast concrete for the parking structure
Photo: The Shockey Precast Group. enabled an accelerated construction schedule and also meant

Precast concrete perimeter walls as well as interior precast concrete vertical and horizontal pieces were installed throughout using a crawler crane,
which was assembled within the parking structure’s excavated footprint. Photo: Hover Solutions LLC.
that shop drawings for the precast concrete could be coordi- was intermixed with other materials designed by the engineer
nated while design of the structure was being finalized. of record for the project,” says Matthew Cooper, engineering
As soon as site work was complete, erection of the precast manager with Shockey Precast. To address this challenge,
concrete components began. As a result, this fast-track con- Shockey created a colocation office, allowing its people to work
struction timetable allowed construction of the patient pavil- together with others on the project during the design phase in
ion to begin on schedule. order to arrive at the best solutions for the project.
One of the key design challenges involved transfer of the Production also posed some challenges. “Due to the heavy
structural load of the tower to the parking structure. This loading we designed the structure to bear, we added some rebar
was another reason that precast concrete made sense: Using threading and shear machines to keep up with the demand for
precast concrete components for the underground structure this project,” Cooper says.
enabled the structural load from the tower to be transferred to Transportation also required some innovative solutions.
the structural precast concrete parking structure earlier in the “There was minimal room on the project site to stage trailers
overall schedule. In addition, the topping slab could be used as for production,” he says. “Most loads required permits for
part of the diaphragm. shipping due to weight and/or size, so we retained a drop lot
Precast concrete ended up being the right choice for close to the project in order to maintain sufficient loads avail-
another reason. Because of the dense development in western able to keep erection running smoothly.”
Philadelphia, where the project is located, normal levels of There were also challenges with installation and erection
on-site labor during construction would have caused excess of the precast concrete pieces at the project site. “Due to
congestion, public safety concerns, and excess traffic. By using site logistics, we had to use multiple cranes during erection,”
precast concrete, however, more than 80% of the labor for Cooper says. “We assembled a crawler crane in the hole, utiliz-
Shockey’s portion of the work was able to take place at its ing a hydro crane. The crawler crane was used for erecting most
manufacturing facility nearly 300 mi (480 km) away. of the product for the project.” The team did leave out one bay
In all, the precast concrete portion of the project consisted of double tees and beams to allow the crawler crane to be dis-
of 1138 pieces, with a maximum piece weight of 82,000 lb mantled. “We completed installation of the remaining product
(365 kN). Specifically, this included 220,000 ft2 (20,400 m2) that we could reach with the hydro crane that was used for
of double tees, 17,000 ft2 (1580 m2) of flat slabs, 125,000 ft2 dismantling the crawler crane. Then, for the product that we
(11,600 m2) of walls, 16 stairs of various thicknesses, 71 col- couldn’t reach with the hydro crane, we utilized tower cranes,”
umns (typically 33 × 66 in. [838 × 1676 mm]), and 85 beams he says. To ensure the success of this challenging process, team
(typically 30 in. [762 mm] deep). members evaluated the complete sequence during the design
On a project of this magnitude, challenges were to be phase to ensure that product weights were kept in line with
expected. The first occurred during the design phase of the the equipment constraints.
precast concrete pieces. “The design of our precast structure —William Atkinson J

Long-term behavior of precast,
prestressed concrete sandwich panels
reinforced with carbon-fiber-reinforced
polymer shear grid
Mohamed K. Nafadi, Gregory Lucier, Tugce Sevil Yaman,

Harry Gleich, and Sami Rizkalla
P
recast concrete sandwich panels are typically used to
construct high-performance, energy-efficient build-
■ This paper presents the results of an experimental ing envelopes. These panels typically consist of two
program that studied the performance of full-scale concrete wythes separated by rigid foam insulation, such as
precast, prestressed concrete sandwich panels rein- expanded polystyrene (EPS) or extruded polystyrene (XPS).
forced with carbon-fiber-reinforced polymer shear The panels are designed to resist floor loads as well as wind
grid and subjected to 2 million reverse-cyclic lateral or seismic lateral loads while providing efficient insulation
load cycles with constant sustained axial load in to the structure. They are often fabricated with heights over
place. 45 ft (13.7 m) and widths up to 15 ft (4.6 m). Wythe thick-
ness commonly ranges from 2 to 6 in. (50.8 to 152.4 mm),
■ Six sandwich panels were constructed with continu- and overall panel thickness may be from 6 to over 12 in.
ous insulation and a carbon-fiber-reinforced polymer (304.8 mm). Longitudinal prestressing is normally provided
grid shear transfer mechanism. Three of the panels in both concrete wythes to control cracks.
were fabricated with expanded polystyrene foam
insulation, and three panels were fabricated using Insulated concrete sandwich panels may be designed as
sandblasted extruded polystyrene foam insulation. fully composite, partially composite, or noncomposite.
The degree of composite action highly depends on the type
■ From each of the panel types, one specimen was of shear connectors joining the concrete wythes. Typical
tested to failure as a control. The other two spec- shear connectors from early-generation precast concrete
imens were subjected to fatigue testing and then sandwich panels include steel-wire truss connectors, bent
tested to failure. reinforcing bars, or solid zones of concrete penetrating the
insulation wythe. Several published studies investigated the
■ The applied fatigue testing did not affect the ultimate performance of such connectors and the associated degree
performance of the panels and had a minimal effect of composite action of the panels.1–6 Although increasing the
on the composite action between the wythes. composite action between the concrete wythes increases the
structural efficiency of the panel, it can significantly reduce
the overall thermal efficiency due to the thermal bridges pro-
PCI Journal (ISSN 0887-9672) V. 66, No. 5, September–October 2021.
PCI Journal is published bimonthly by the Precast/Prestressed Concrete Institute, 8770 W. Bryn Mawr Ave., Suite 1150, Chicago, IL 60631. duced through the shear connectors. Noncomposite panels
Copyright © 2021, Precast/Prestressed Concrete Institute. The Precast/Prestressed Concrete Institute is not responsible for statements made then became more attractive due to their thermal benefits and
by authors of papers in PCI Journal. Original manuscripts and discussion on published papers are accepted on review in accordance with the
architectural features; however, noncomposite panels exhib-
Precast/Prestressed Concrete Institute’s peer-review process. No payment is offered.

ited substantially reduced structural efficiency. Although the More recently, an experimental program conducted by Kazem
typical design for such panels assumes noncomposite action, et al.19 evaluated the effect of sustained loading on the shear
test results from several studies indicate that significant shear strength of precast concrete sandwich panels connected
transfer does develop between the concrete wythes, resulting with CFRP and GFRP grids and using EPS and sandblasted
in a partial composite action.3–5 XPS insulation. The experimental program comprised three
different studies with a total of 26 test panels using different
Several studies investigated the use of fiber-reinforced poly- configurations of FRP grid and foam insulation. The research
mer (FRP) shear connection grid as a mechanism for improv- concluded that both CFRP and GFRP grid combined with
ing structural efficiency while avoiding the thermal bridges EPS and XPS foam insulation can provide a suitable shear
created by conventional means of shear transfer. Because FRP transfer mechanism for precast concrete sandwich panels
grid has very low thermal conductivity, the carbon grid can under the effects of sustained load. In addition, the research
be used to connect the concrete wythes while maintaining the findings demonstrated that the shear strength of the panels can
insulation of the panel. Several studies have also investigated be significantly increased by sandblasting the surface of the
the performance of glass-fiber-reinforced polymer (GFRP) XPS foam due to the corresponding improvement of the bond
shear connectors for insulated concrete sandwich panels.7–12 between the foam and the concrete.
These studies demonstrated that the use of GFRP increased
the thermal performance of the panels compared with panels Another study, conducted by Olsen and Maguire,20 examined
with steel or concrete connectors and maintained a high de- the performance of GFRP connectors in precast concrete
gree of composite action. sandwich panels. The study included an experimental
program to investigate the efficiency of several configura-
Frankl13 and Frankl et al.14 conducted an experimental pro- tions of GFRP connectors along with the effects of other
gram—including six full-scale precast, prestressed concrete parameters, including the type of foam and the concrete-to-
sandwich panels—to investigate flexural behavior using foam interface bond. The study concluded that the selected
carbon-fiber-reinforced polymers (CFRP) as a shear transfer connectors provided reduced strength and stiffness with
mechanism. The study also investigated the effects of sev- large wythe thicknesses and when debonded. In addition,
eral parameters, such as the type of insulation, the presence the study developed a simplified model to evaluate the shear
of solid concrete zones, panel configuration, and shear grid deformation behavior. A parametric study performed using
reinforcement ratio. Test results indicated that the failure took the developed model concluded that a triangular distribution
place at levels well above factored design loads. The study of connectors that has more connectors lumped near the
also evaluated the degree of composite action for the tested ends exhibits higher structural efficiency and that the level
panels and concluded that nearly 100% composite action can of composite action generally increases when the number of
be achieved with CFRP grid shear connections. The study also shear connectors is increased.
demonstrated that a higher composite action percentage can
be achieved using EPS insulation rather than XPS insulation. Dutta et al.21 conducted an experimental study to investigate
With either type of insulation, the use of CFRP shear grid the performance and efficiency of sandwich panels reinforced
can provide an effective shear transfer mechanism in precast, with new C-shaped GFRP pultruded channel shear connectors
prestressed concrete sandwich wall panels. in terms of the level of composite action as well as the flex-
ural strength. The shear connectors were configured in three
Using Frankl’s test results, an analytical study was carried forms: a continuous GFRP channel, discontinuous GFRP
out by Hassan et al.15 to validate design guidelines proposed channel segments, and a control conventional steel truss. Test
for precast, prestressed concrete sandwich panels reinforced results indicated that panels with continuous GFRP channels
with a CFRP grid. The study indicated good agreement be- performed the best in terms of flexural strength. The results
tween the experimentally measured strains at different load also indicated that the level of composite action was about
levels and those predicted using the proposed design guide- 50% for both continuous and discontinuous GFRP channels
lines. A simplified design chart was also proposed for panels compared with 33% for the steel truss. These findings were
with the same configuration as those tested experimentally verified by extensive finite element modeling, which also in-
to calculate the nominal flexural strength at different degrees cluded a parametric study to investigate the effects of different
of composite action. parameters believed to affect the performance of sandwich
panels reinforced with the new GFRP channels.22 The finite
Bunn,16 Sopal,17 and Hodicky et al.18 conducted several element results also revealed that using circular openings in
experimental programs to study the CFRP grid and rigid the GFRP channel web can be used to increase the efficiency
foam insulation shear transfer mechanism and to investigate of thermal insulation of the panel without affecting the struc-
several parameters believed to influence shear flow strength. tural performance.
These studies developed equations to predict the shear flow
strength provided by CFRP grid and rigid foam as affected by Despite the extensive research conducted to study the be-
the tested parameters. It was concluded that the desired level havior of precast, prestressed concrete sandwich panels
of composite action could be achieved using CFRP grid with reinforced with FRP shear grid, limited data are available
either EPS or XPS rigid foam insulation. regarding the long-term behavior of these members under the

effect of fatigue loading. This paper presents the results of an All panels were identical other than the insulation type. Each
experimental program that was conducted to characterize the panel measured a nominal 8 in. (203.2 mm) thick, 4 ft (1.2 m)
performance of full-scale precast, prestressed concrete sand- wide, and 20 ft (6.1 m) tall. Figure 1 shows the three layers of
wich panels reinforced with CFRP shear grid and subjected the typical cross section of all panels. The panel configuration
to 2 million reverse-cyclic lateral load cycles with constant included the following:
sustained axial load in place.
• 2 in. (50.8 mm) prestressed concrete outer wythe
Experimental program
• 4 in. (101.6 mm) center layer of EPS or sandblasted XPS
A series of six panels were tested, including one group of foam insulation
three panels with EPS insulation and one group of three
panels with sandblasted XPS insulation. One panel in each • 2 in. prestressed concrete inner wythe
group of three was randomly selected as the control and was
incrementally tested to failure under reverse-cyclic lateral The test specimens were labeled as follows:
loads. The remaining two panels in each group were subjected
to 2 million fully reversed lateral load cycles at 45% of their • EPS1: first specimen with EPS insulation tested to failure
design ultimate load before the failure test. Testing to 45% of
the ultimate level of load for the fatigue cycles was chosen to • EPS2: second specimen with EPS insulation subjected to
represent the lateral loading resulting from the average typical fatigue cycles before the failure test
wind speed that these panels would be subjected to over their
life spans. A constant service-level eccentric axial load was in • EPS3: third specimen with EPS insulation subjected to
place for all tests, including the lateral fatigue cycles. fatigue cycles before the failure test
Figure 1. Specimen cross section from producer shop tickets. Note: 1" = 1 in. = 25.4 mm; 1' = 1 ft = 0.305 m.

• XPS1: first specimen with sandblasted XPS insulation assemblies were placed on top, and CFRP grid was pushed
tested to failure into the wet concrete wythe beneath. With the CFRP grid
projecting up from the insulation, the top wythe welded-wire
• XPS2: second specimen with sandblasted XPS insulation reinforcement was placed. The top wythe concrete was cast,
subjected to fatigue cycles before the failure test and embedded plates for connections and lifting devices were
set. Figure 3 shows typical photographs of the panels during
• XPS3: third specimen with sandblasted XPS insulation construction.
subjected to fatigue cycles before the failure test
The specified concrete strength was 5000 psi (34.5 MPa) and
Each concrete wythe was prestressed longitudinally with two the average compressive strength of the wall panels at time of
3
∕8 in. (9.525 mm) diameter strands in addition to one layer testing was 8500 psi (58.6 MPa), as determined by compres-
of welded-wire reinforcement with W3.5 wires spaced at sion tests of three 4 × 8 in. (101.6 × 203.2 mm) cylinders cast
8 in. (203.2 mm) in the transverse direction. CFRP grid was with the test panels.
provided between the wythes to transfer shear across the rigid
insulation. Two strips of the grid were placed parallel to the Design loads and loading sequence
long axis of each panel at the locations shown in Fig. 2. A
small additional strip of CFRP grid was placed in the trans- The panels were designed for a maximum lateral wind pres-
verse panel direction to reinforce the region surrounding the sure of 42.5 lb/ft2 (2.03 kPa) due to suction in combination
lifting points for a total of 36 ft (11 m) of CFRP grid in each with a service-level axial load of 4 kip (17.8 kN). The lateral
panel. The installed CFRP grids were orthogonal and cut at wind loadings were determined according to the design wind
a 45-degree angle to develop truss action. The rigid foam in- pressures specified in American Society of Civil Engineers’
sulation was thinned in areas surrounding embedded connec- ASCE/SEI 7-1623 chapter 30, assuming a basic wind speed of
tions and lifting points; however, the insulation was continu- 150 mph (240 km/hr), building classification III, and exposure
ous from end to end and from side to side. Care was taken in category B. The tested panels were designed per the standard
the experimental design to ensure that the testing setup itself procedures of the shear grid manufacturer, as documented in
did not artificially enhance the connection between inner and International Code Council Evaluation Service (ICC-ES) re-
outer wythes. port ESR-2953,24 which includes shear flow and shear modu-
lus of the grid connectors. The axial load was applied through
Construction of wall panels a steel corbel at the top of the panel having an eccentricity
of 4 in. (101.6 mm) from the innermost surface of the inner
The panels were cast flat in long-line production forms. wythe. Lateral loads were applied in four-point bending, with
Fabrication began by placing formwork, stressing strands, the panel supported laterally at the top and bottom and loaded
and laying reinforcement for the outer wythe of concrete. equally at the quarter and three-quarters heights. The quarter
After casting a 2 in. (50.8 mm) layer of concrete for the and three-quarters heights did not exactly match the quarter
outer wythe, prefabricated foam insulation and CFRP grid and three-quarters spans because the top support was located
Figure 2. Specimen plan and profile from producer shop tickets. Note: CONC. = concrete; EA = each; INSUL. = insulation;
MIN. = minimum; RAD. = radius; TYP. = typical; W.W.F. = welded-wire reinforcement. 1" = 1 in. = 25.4 mm; 1' = 1 ft = 0.305 m.

Forms with outer wythe reinforcement installed Foam core with carbon-fiber-reinforced
polymer grid installed after casting
the outer wythe
Figure 3. Typical views of panels during fabrication.
below the top of the panel, which is common for typical roof 1000 lb (4.45 kN) increments (500 lb [2.22 kN] per loading
detailing. Quarter-height lateral loading was chosen so that jack). Table 1 summarizes the loading scheme. One load cy-
the moment and shear distributions of the applied lateral test cle was considered as taking the panel from zero lateral load
loads would closely mimic those of uniformly distributed to the specified lateral load level in both directions.
wind pressure.
Two EPS panels and two sandblasted XPS panels were
One EPS panel and one sandblasted XPS panel were random- each subjected to 2 million reverse-cyclic lateral load cycles
ly selected as the control specimens and were incrementally before failure testing (Table 1). Cycles were applied at a rate
tested to failure under reverse-cyclic lateral loads. The axial of approximately 1 Hz, for a total cycling time of approxi-
load was applied with hydraulic cylinders before lateral load- mately 23 days per panel. A constant axial load was applied
ing and was held constant using nitrogen-charged hydraulic to the panels during all fatigue loading. Each fatigue cycle
accumulators during the entire test. Lateral loads were applied induced lateral loading in the positive and negative directions
incrementally in both directions, pushing and pulling, in to a selected lateral load corresponding to 45% of the design
Table 1. Loading sequence for all static tests to failure
Loads on panel
Step Equivalent uniform
Total applied lateral load, lb Load per actuator, lb Axial load, lb
pressure on panel, lb/ft2
1 ±1000 ±500 ±12.5 4000
2 ±2000 ±1000 ±25 4000
3 ±3000 ±1500 ±37.5 4000
4 ±4000 ±2000 ±50 4000
Continue increasing each increment

5 (Total load)/2 (Total load)/20 ft/4 ft 4000
by ±1000 lb to failure
Note: 1 ft = 0.305 m; 1 lb = 0.00445 kN; 1 lb/ft2 = 0.048 kPa.

ultimate load. Such value was selected to represent the typical ed lateral translation. The top lateral support (and axial load
service-level wind load that would be expected during average connection) was attached only to the inner wythe, as would be
service (that is, the average wind speed over the life of the common in practice for the detailing of a typical panel.
panel). This load level was equivalent to 19.1 lb/ft2 (0.91 kPa),
1530 lb (6.8 kN) of the total lateral load, and 765 lb (3.4 kN) Two matching hydraulic actuators were used to apply tension
of the lateral load per load point. After fatigue testing, each and compression loads in the lateral direction. The actuators
panel was subjected to the incremental static test to failure as were always configured to produce matching loads, regard-
previously described. less of actuator stroke or panel deflection. Each actuator
was attached to the panel using two 4 ft (1.2 m) long square
Test setup loading tubes, one on each surface of the panel. The loading
tubes were bolted together through substantially oversized
The test setup was designed to enable cycling two panels holes in the panel to allow application of lateral loads in
simultaneously but independently. Two identical setups were either direction without artificially connecting the wythes.
fabricated side by side under the same supporting frame (Fig. 4 The loading tubes were separated from the panel surface by
and 5). All panels were supported at the bottom on a pin con- a thick neoprene pad on each wythe to avoid unintentionally
nection that restrained translation in all directions but allowed restraining the panel with the loading system. Axial load was
rotation. Only the inner wythe was connected to the lower applied with a hydraulic jack to the top of the panel through
connection to avoid artificially enhancing the connection be- a steel corbel welded to the embedded plate. The axial load
tween wythes. All panels were supported at the top on a slotted was applied and regulated using accumulators charged with
roller that allowed vertical translation and rotation but prevent- nitrogen in the axial load hydraulic circuit. This configura-
Figure 4. Schematic of the test setup, view of inner wythe.

Test setup with one panel Test setup with two panels
Figure 5. Views of the test setup.
tion allowed the axial load to remain virtually constant as the For the test setup as configured, extension in the lateral actu-
panel deformed laterally. The corbel was specially configured ators created positive loads while the tension in the actuators
to avoid interfering with the lateral supports. was considered negative load.
Instrumentation Potentiometers were used to monitor the lateral deflections of

the panel throughout each test at different locations (Fig. 6).
Various instruments were used during testing to moni- All deflection data presented in subsequent sections are plot-
tor panel behavior. All instruments were connected to an ted according to the given sign convention and are adjusted
electronic data acquisition system, which recorded data at to eliminate the effects of support motion. Lateral deflections
a sample rate of 1 Hz during static loading and at 10 Hz for were considered positive when they acted to place the inner
selected intervals during cyclic loading. Figure 6 shows the wythe in compression (that is, positive lateral deflection is the
instrumentation layout and sign convention for applied loads panel moving outward).
and measured deflections.
To characterize the degree of composite action of the panel,
Various load cells were used to measure the applied axial two groups of four strain gauges each were attached to the
load as well as the lateral loads. Axial load was maintained by side surfaces of the two concrete wythes. One group of four
applying constant hydraulic pressure to a calibrated hydraulic gauges was attached at midheight, and the other group of
jack. The axial load was considered positive when acting to four gauges was attached at a location 1 ft (0.305 m) below
place the inner wythe in compression (that is, the positive axi- the three-quarters height (1 ft below the upper load point).
al load acts vertically downward). One load cell was incorpo- All gauges were centered 7∕16 in. (11.1 mm) from the nearest
rated into the body of each lateral load jack. These load cells wythe surface or wythe-foam interface. Figure 7 shows the
were used to continuously monitor the applied lateral load. groups of bonded strain gauges.

Figure 6. Instrumentation layout and sign convention for applied loads and measured deflections. Note: 1 ft = 0.305 m.
In addition to the strain gauges, six reusable surface-mounted (4.79 kPa). In this group of three panels, one of the two
strain gauges, referred to as pi gauges, were used to measure fatigued panels outperformed the control panel in terms of ul-
the flexural strain at the concrete surface of the panel. Three timate load: EPS2 achieved 112.5 lb/ft2 (5.39 kPa) compared
gauges were installed on the inner panel surface and three gaug- with 100 lb/ft2 for EPS1. It is unlikely that the fatigue cycles
es on the outer panel surface. Figure 7 shows the outer-face enhanced the panel performance in any way, and it is also un-
pi gauges on a selected panel. All bonded strain gauges and pi likely that the small increase in concrete compressive strength
gauges were attached to the panel after fatigue testing to avoid (due to panel aging) that probably occurred during the fatigue
fatiguing the gauges themselves. All measured strains were cycles improved ultimate failure performance. At 100 lb/ft2
considered positive in tension and negative in compression. of applied lateral load (with 4 kip [17.8 kN] of axial load in
place), the panels exceeded their design load of 42.5 lb/ft2
Results and discussion (2.03 kPa) by a factor of 2.35.
Table 2 summarizes the test results. For all tests, the negative The tested XPS panels all failed at the equivalent of 175 lb/ft2
segment of a given lateral load cycle (the segment with the (8.38 kPa) of applied lateral pressure, which was higher than
inner wythe in tension) was completed before the positive that achieved by the EPS panels. This can be attributed to the
segment of that cycle was attempted. Due to the effects of the fact that sandblasted XPS has greater bond strength, shear
eccentric axial load, the bending moment created in the panel strength, and stiffness than EPS, evidenced by a nominal
during the positive segment of each cycle was greater than shear flow design value of 450 lb/in. (50.8 N/m) for XPS and
the moment created by the same lateral load on the negative CFRP grid compared with 270 lb/in. (30.5 N/m) for EPS and
segment. All four panels subjected to fatigue loading survived CFRP grid. The XPS panels achieved an ultimate lateral load
2 million lateral load cycles without any visible signs of degra- that was more than four times their 42.5 lb/ft2 (2.03 kPa) de-
dation. These four panels were then tested to failure using the sign load. All tested panels were designed using the standard
incremental static loading procedure outlined in the previous prescribed design methods of the CFRP grid manufacturer,
section. A constant 4 kip (17.8 kN) eccentric axial load was so these high overstrengths against shear transfer mechanism
applied to the corbel at the top of each panel during all phases failure would also be typical of production panels. Part of the
of loading, including fatigue cycles. The axial load application excess capacity can likely be attributed to measured con-
was regulated to remain constant even as the panel deformed crete strength exceeding that specified by design (8500 psi
under lateral load. The equivalent lateral design wind pressure [58.6 MPa] compared with 5000 psi [34.5 MPa], respec-
was 42.5 lb/ft2 (2.03 kPa). All six panels sustained applied tively); however, the effect of modestly increased concrete
lateral loads well in excess of their design values. strength on panel capacity is likely minimal. Both EPS and
XPS panels exceeded their design levels by ratios greater
The results indicate that the tested EPS panels all failed when than the increase of nominal flexural or shear strength due
the applied lateral load was greater than or equal to 100 lb/ft2 to the increased concrete compressive strength. The selected

Strain gauges Surface-mounted strain gauges
(pi gauges)
Figure 7. Photographs of instrumentation. Note: 1 ft = 0.305 m.
fatigue regimen did not seem to negatively affect the ultimate This can be attributed to the fact that the upper support was
performance of the panels in any way. It is noted that the about 1 ft (0.305 m) below the top face of the panel, and the
applied fatigue loading of 45% of the design ultimate load locations of the applied loads were not fully symmetric with
ended up being about 19% and 11% of the actual ultimate respect to the supports (Fig. 6). Hence, the location of the
load of the EPS and XPS panels, respectively, due to the panel maximum moment was consistently shifted to the lower load-
overstrength. These levels of overstrength were achieved with ing point, whether the load was applied in pulling or pushing
standard design methods and therefore would be typical of directions. All panels were visually uncracked at the start of
production panels designed with the same methods. failure testing, including panels that had been subjected to the
2 million fatigue cycles. The following sections describe the
Cracking patterns and failure modes cracking pattern and failure mode for each tested panel.
Observed cracking in all panels primarily took the form of EPS1-control Horizontal cracks were first observed on
horizontal flexure cracking in the zone between the load the inner wythe of the EPS control panel during the pulling
points. Figure 8 shows examples of typically observed cracks, segment of a total applied load cycle of 8000 lb (35.6 kN).
failure modes, and a typical CFRP grid failure. Critical cracks The test was terminated because the panel could not sustain
often developed at or near failure at the lower loading point. this load level and continued to displace laterally under the

Table 2. Summary of results for all panels
Total ultimate Total ultimate

Failure
Panel applied lateral equivalent uniform Failure mode
cycle
load, lb pressure, lb/ft2
Global flexure failure accompanied by loss of shear

4000 lb/jack
EPS1-control 8000 (pull) 100 transfer after reaching 4000 lb/jack pulling (did not
cycle (pull)
sustain this level).
Global flexure and shear failure at lower lateral load

4500 lb/jack point with loss of shear transfer. Survived 4500 lb/jack
EPS2-fatigued 9000 (pull) 112.5
cycle (push) pulling. Failed at 4000 lb/jack while pushing towards
4500 lb/jack on subsequent segment.
Global flexure and shear failure at lower lateral load

4000 lb/jack point with loss of shear transfer. Survived 4000 lb/jack
EPS3-fatigued 8000 (push) 100
cycle (push) pulling. Reached 4000 lb/jack pushing but could not
sustain this load.
7000 lb/jack Global flexural failure after reaching 7000 lb/jack pulling.
XPS1-control 14,000 (pull) 175
cycle (pull) Did not sustain this load level.
Extensive global flexural behavior followed by ultimate

7000 lb/jack
XPS2-fatigued 14,000 (push) 175 flexure and shear crack at the lower load point. Reached
cycle (push)
7000 lb/jack pushing but did not sustain this level.
Global flexural failure at the lower lateral load point. Sus-

7000 lb/jack
XPS3-fatigued 14,000 (pull) 175 tained 7000 lb/jack pulling, failed at 6500 lb/jack while
cycle (push)
pushing towards 7000 lb/jack on subsequent segment.
Note: EPS1-control = first specimen with expanded polystyrene insulation tested to failure; EPS2-fatigued = second specimen with expanded polysty-
rene insulation subjected to fatigue cycles before the failure test; EPS3-fatigued = third specimen with expanded polystyrene insulation subjected to
fatigue cycles before the failure test; XPS1-control = first specimen with sandblasted extruded polystyrene insulation tested to failure; XPS2-fatigued =
second specimen with sandblasted extruded polystyrene insulation subjected to fatigue cycles before the failure test; XPS3-fatigued = third specimen
with sandblasted extruded polystyrene insulation subjected to fatigue cycles before the failure test. 1 lb = 0.00445 kN; 1 lb/ft2 = 0.048 kPa.
constant applied load. Significant relative displacement was lateral load point at a total applied load of 8000 lb (35.6 kN)
observed at the interface between the inner concrete wythe in the pushing segment, which the panel reached but could
and foam core at and after failure. This behavior suggests not sustain.
that the failure took place due to a loss of composite action
between the wythes that led to a global flexure failure. XPS1-control Panel XPS1 demonstrated significant global
flexural action before failure, including extensive horizon-
EPS2-fatigued Similar to EPS1, no cracks were observed on tal cracking between the load points and below the lower
EPS2 before failure. As the failure developed, a large flexure load point. Horizontal cracks were obvious on the inner and
crack appeared underneath the lower loading tube. This crack outer wythes at the cycle of a total applied load of 8000 lb
was accompanied by concrete crushing on the opposite face. (35.6 kN). Horizontal cracks continued to grow in width and
The panel sustained a total applied load of 9000 lb (40 kN) to increase in number before failure. Unlike the EPS panels,
on the pulling segment in the cycle segment before failure. the XPS1 panel was able to continue resisting total applied
A global flexure and shear failure due to loss of composite loads up to 14,000 lb (62.2 kN) at a pulling segment, at which
action then took place at the lower lateral load point at a a global flexural failure took place.
total applied load of 8000 lb (35.6 kN) while pushing toward
9000 lb on the subsequent pushing segment. XPS2-fatigued Cracking in XPS2 progressed in much the
same way as for XPS1. Horizontal cracks developed and
EPS3-fatigued EPS3 behaved in a similar fashion to EPS2, spread on both faces as loading increments increased. The
as expected. The panel remained visually uncracked until panel exhibited extensive global flexural behavior followed
just before failure. Horizontal cracks developed at the lower by an ultimate flexure and shear crack at the lower load point
load point (Fig. 8). A secondary horizontal crack developed at a total applied load of 14,000 lb (62.2 kN) in the pushing
between the lower lateral load point and the support, likely segment, which the panel reached but could not sustain.
due to a partial loss of composite action at failure. Failure in
the shear transfer mechanism then triggered an immediate XPS3-fatigued XPS3 behaved in much the same way as
global flexure and shear failure that took place at the lower XPS2, as expected. The failure mode was due to a large flex-

Figure 9. Measured lateral deflections at midheight for the
EPS panels. Note: Only the first and last cycles are shown for
better clarity. EPS = expanded polystyrene; EPS1-control =
first specimen with expanded polystyrene insulation tested
to failure; EPS2-fatigued = second specimen with expanded
polystyrene insulation subjected to fatigue cycles before the
failure test; EPS3-fatigued = third specimen with expanded
polystyrene insulation subjected to fatigue cycles before the
failure test.
Third test specimen with Third test specimen with

expanded polystyrene sandblasted extruded
insulation (EPS3) polystyrene insulation
(XPS3)
Figure 10. Measured lateral deflections at midheight for the

Typical CFRP grid failure sandblasted XPS panels. Note: Only the first and last cycles
are shown for better clarity. XPS = extruded polystyrene;
XPS1-control = first specimen with sandblasted XPS insulation
tested to failure; XPS2-fatigued = second specimen with
Figure 8. Cracking pattern and failure mode of test speci- sandblasted XPS insulation subjected to fatigue cycles before
mens. Note: CFRP = carbon-fiber-reinforced polymer. the failure test; XPS3-fatigued = third specimen with sand-
blasted XPS insulation subjected to fatigue cycles before the
ure and shear crack at the lower load point (Fig. 8). The panel failure test.
could sustain up to a total applied load of 14,000 lb (62.2 kN)
on the pulling segment, followed by a global flexural failure better clarity, the figures show only the first and last cycles.
at the lower lateral load point at 13,000 lb (57.8 kN) of total The overall performance of all three panels was quite similar.
applied load while pushing toward the same 14,000 lb target. The initial stiffness of control panel EPS1 and fatigued panel
EPS2 are quite similar. Interestingly, the initial stiffness of
Measured lateral deflections fatigued panel EPS3 slightly exceeds that of EPS1 and EPS2,
likely due to random variations in the manufacturing process.
Figures 9 and 10 compare the lateral deflections measured at In addition, the residual deflections (hysteresis) observed
the midheights of the EPS and XPS panels, respectively. For during the static cycles of EPS3 are relatively less than those

from equivalent cycles for EPS1 and EPS2. This finding im- Measured strains
plies that panel EPS3 had a better bond between concrete and
foam than did panels EPS1 and EPS2. It is not reasonable to Two groups of four strain gauges each were used to monitor
conclude that fatigue cycling enhanced the EPS3 bond in any the strains across the panel thickness at midheight and at 1 ft
way, so it is assumed that this better bond was due to natural (0.305 m) below the upper load point. The measured strains
variability in the manufacturing process. Panels EPS1, EPS2, were then used to plot the strain profiles across the thickness
and EPS3 were manufactured at the same time from the same for each panel at different load levels to characterize the de-
material inputs, and no obvious differences in the production gree of composite action.
of these three panels were observed. Therefore, it is concluded
that the fatigue cycles did not influence the final load-deflec- Figure 11 shows strain profiles plotted for each panel at a se-
tion behavior of the EPS panels. lected load level, where a total applied lateral load of 3400 lb
(15.1 kN) was pushing the panel outward at 1 ft (0.305 m)
Figure 10 shows the measured lateral deflections at the mid- below the upper load and at midheight of each panel. The load
heights of the XPS panels. The behaviors of the control panel level represents an equivalent uniform pressure on the panel
XPS1 and the fatigued panels XPS2 and XPS3 could not be of 42.5 lb/ft2 (2.03 kPa), which is the design lateral wind
distinguished from this plot. The overall behaviors appeared pressure. Despite the small values of the measured strains,
to be identical for all practical purposes, and the initial stiff- which were close to the accuracy of the instrumentation, the
ness matched across the three tests. It is concluded that the recorded strains at the outer and inner surfaces were clearly
fatigue cycles did not affect the final load-deflection behavior in opposite directions as the inner wythe was in compression
of the tested XPS panels. while the outer wythe was in tension. The results indicate
Figure 11. Measured strains across the panel thickness at total applied load of 3400 lb (pushing segment) at upper and mid-
height groups of strain gauges. Note: Measured strains are in microstrain. EPS1 = first specimen with expanded polystyrene
insulation tested to failure; EPS2 = second specimen with expanded polystyrene insulation subjected to fatigue cycles before the
failure test; EPS3 = third specimen with expanded polystyrene insulation subjected to fatigue cycles before the failure test; XPS1
= first specimen with sandblasted extruded polystyrene insulation tested to failure; XPS2 = second specimen with sandblasted
extruded polystyrene insulation subjected to fatigue cycles before the failure test; XPS3 = third specimen with sandblasted ex-
truded polystyrene insulation subjected to fatigue cycles before the failure test. 1 lb = 0.00445 kN.

that the strain profiles for all of the panels were nearly linear
except one location at the upper gauge group in panel EPS2,
where each wythe appeared to act independently. These
results suggest that all of the panels were generally able to
exhibit a relatively high degree of composite action up to
the design load. Furthermore, it could be concluded that the
applied fatigue cycling before the failure test had a minimal
effect on the composite action between the wythes.
These results indicate that in the EPS panels, a loss of

composite action likely triggered flexural failure and in the
sandblasted XPS panels, a loss of composite action was likely
triggered by flexural failure. Although the findings of Frankl
et al.14 indicate that a higher degree of composite action can
be achieved using EPS insulation than standard XPS insula-
tion, these results indicate that the sandblasted XPS panels
could maintain a higher degree of composite action up to fail-
ure compared with EPS panels. This can be mainly attributed Figure 12. Measured concrete strains at inner and outer faces
to the beneficial effect from sandblasting the surface of the of panel EPS3. Note: EPS3 = third specimen with expanded
XPS foam, which is expected to improve the bond between polystyrene insulation subjected to fatigue cycles before the
failure test. 1 lb = 0.00445 kN.
the concrete wythes and the foam and, consequently, increase
the degree of composite action and overall flexural and shear
strengths of the panels. design load of 42.5 lb/ft2 (2.03 kPa). The tested sand-
blasted XPS panels failed at the equivalent of 175 lb/ft2
Figure 12 shows an example of measured strains at both (8.38 kPa) of applied lateral pressure, equivalent to over
the inner and outer concrete faces for panel EPS3. The 4.0 times their design ultimate load of 42.5 lb/ft2.
measured strains were generally small in value and symmet-
rical with respect to the vertical axis. The load versus strain • All four panels subjected to fatigue survived 2 million
behavior was fairly linear up to the failure at all locations lateral load cycles without any visible signs of degrada-
except near the lower loading point at both the inner and tion. For the EPS panels, the two fatigued panels slightly
outer faces, where flexural and shear cracks initiated before outperformed the control panel in terms of ultimate load
failure. This behavior is consistent with the visual observa- (EPS2 achieved 112.5 lb/ft2 [5.39 kPa]) and initial stiff-
tions during the test, which did not reveal any cracks until ness (EPS3 exhibited a greater initial stiffness). There-
just before the failure. fore, it is concluded that the effects of fatigue at the se-
lected levels on the ultimate load-deflection performance
Conclusion are less significant than are the effects of variability in the
manufacturing process. For the XPS panels, the load-de-
This paper documents testing of six 20 × 4 ft (6.1 × 1.2 m) flection behaviors of the three panels were indistinguish-
precast, prestressed concrete sandwich panels constructed able up to failure, indicating a high level of consistency
with continuous insulation and a CFRP grid shear transfer in the manufacturing process and foam-to-concrete bond
mechanism. All panels were identical except for foam type, for these panels.
and all panels were cast together on the same long-line
prestressing bed. Three of the six panels were fabricated • All panels demonstrated a high degree of composite ac-
with EPS foam insulation, and the remaining three panels tion through load levels well in excess of their 42.5 lb/ft2
were fabricated using sandblasted XPS foam insulation. (2.03 kPa) design load; however, composite action was
For each group of three panels, one was randomly selected lost at failure. It is likely that a loss of composite action
and tested to failure as a control and two others were each triggered failure in the EPS panels and that global failure
subjected to 2 million lateral load cycles equivalent to 45% of triggered loss of composite action in the sandblasted XPS
their ultimate lateral design capacity, in combination with a panels. The sandblasted XPS panels showed a higher
service-level axial load. After fatigue testing, all panels were level of composite action than the EPS panels, as would
tested to failure. The conclusions are as follows: be expected by the larger stiffness of XPS compared with
EPS and the strong bonding characteristics of sandblasted
• All tested panels were designed with standard commer- XPS. This behavior can be attributed to the beneficial
cial methods using the manufacturer’s recommended val- effects of sandblasting the surface of the XPS foam,
ues for shear flow, and all sustained applied lateral loads which is expected to improve the bond between the con-
well in excess of their design values. The EPS panels all crete wythes and the foam and consequently increase the
failed when the applied lateral load was greater than or degree of composite action and the overall flexural and
equal to 100 lb/ft2 (4.79 kPa), which is 2.35 times their shear strengths of the panels.

Acknowledgments cast Prestressed Concrete Sandwich Panels with NU-Tie.”
Final report. University of Nebraska–Lincoln.
The authors would like to acknowledge AltusGroup for
sponsoring these tests and for designing, producing, and 11. Naito, C., J. Hoemann, M. Beacraft, and B. Bewick. 2012.
transporting the test panels. The authors are also grateful “Performance and Characterization of Shear Ties for Use
to Metromont Corp. and its crews in Charlotte, N.C., for in Insulated Precast Concrete Sandwich Wall Panels.”
fabricating the test specimens. In addition, the authors would Journal of Structural Engineering 138 (1): 52–61.
like to thank the staff at the Constructed Facilities Laboratory
at North Carolina State University for their help throughout 12. Woltman, G., D. Tomlinson, and A. Fam. 2013. “Investi-
the experimental program. Finally, the authors would like to gation of Various GFRP Shear Connectors for Insulated
acknowledge Hamid Kazem, the former postdoctoral fellow Precast Concrete Sandwich Wall Panels.” Journal of
at North Carolina State University, for his assistance with the Composites for Construction 17 (5): 711–721.
experimental program.
13. Frankl, B. 2008. “Structural Behavior of Insulated Precast
References Prestressed Concrete Sandwich Panels Reinforced with
CFRP Grid.” MSc thesis, Department of Civil, Construc-
1. Gleich, H. 2007. “New Carbon Fiber Reinforcement tion and Environmental Engineering, North Carolina
Advances Sandwich Wall Panels.” Structure Magazine State University, Raleigh, NC.
(April): 61–63.
14. Frankl, B. A., G. W. Lucier, T. K. Hassan, and S. H. Riz-
2. PCI Committee on Precast Sandwich Wall Panels. 1997. kalla. 2011. “Behavior of Precast, Prestressed Concrete
“State of the Art of Precast/Prestressed Sandwich Wall Sandwich Wall Panels Reinforced with CFRP Shear
Panels.” PCI Journal 42 (2): 92–134. Grid.” PCI Journal 56 (2): 42–54.
3. Pessiki, S., and A. Mlynarczyk. 2003. “Experimental 15. Hassan, T. K., and S. H. Rizkalla. 2010. “Analysis and
Evaluation of Composite Behavior of Precast Concrete Design Guidelines of Precast, Prestressed Concrete,
Sandwich Wall Panels.” PCI Journal 48 (2): 54–71. Composite Load-Bearing Sandwich Wall Panels Rein-
forced with CFRP Grid.” PCI Journal 55 (2): 147–162.
4. Bush, T. D., and G. L. Stine. 1994. “Flexural Behavior of
Composite Prestressed Sandwich Panels.” PCI Journal 16. Bunn, W. G. 2011. “CFRP Grid/Rigid Foam Shear
39 (2): 112–121. Transfer Mechanism for Precast, Prestressed Concrete
Sandwich Wall Panels.” MSc thesis, Department of Civil,
5. Lee, B., and S. Pessiki. 2007. “Design and Analysis of Construction and Environmental Engineering, North
Precast, Prestressed Concrete Three-Wythe Sandwich Carolina State University, Raleigh, NC.
Wall Panels.” PCI Journal 52 (4): 70–83.
17. Sopal, G. J. 2013. “Use of CFRP Grid as Shear Transfer
6. Lee, B., and S. Pessiki. 2008. “Experimental Evaluation Mechanism for Precast Concrete Sandwich Wall Panels.”
of Precast, Prestressed Concrete, Three-Wythe Sandwich PhD diss., Department of Civil, Construction and Envi-
Wall Panels.” PCI Journal 53 (2): 95–115. ronmental Engineering, North Carolina State University,
Raleigh, NC.
7. Salmon, D. C., A. Einea, M. K. Tadros, and T. D. Culp.
1997. “Full Scale Testing of Precast Concrete Sandwich 18. Hodicky, K., G. Sopal, S. Rizkalla, T. Hulin, and
Panels.” ACI Structural Journal 94 (4): 354–362. H. Stang. 2015. “Experimental and Numerical Investiga-
tion of the FRP Shear Mechanism for Concrete Sandwich
8. Lameiras, R., J. Barros, I. B. Valente, and M. Azenha. Panels.” Journal of Composites for Construction 19 (5):
2013. “Development of Sandwich Panels Combining 04014083.
Fibre Reinforced Concrete Layers and Fibre Reinforced
Polymer Connectors. Part I: Conception and Pull-Out 19. Kazem, H., W. G. Bunn, H. M. Seliem, S. H. Rizkalla,
Tests.” Composite Structures 105: 446–459. and H. Gleich. 2015. “Durability and Long Term Behav-
ior of FRP/Foam Shear Transfer Mechanism for Concrete
9. Lameiras, R., J. Barros, M. Azenha, and I. B. Valente. Sandwich Panels.” Construction and Building Materials
2013. “Development of Sandwich Panels Combining 98: 722–734.
Fibre Reinforced Concrete Layers and Fibre Reinforced
Polymer Connectors. Part II: Evaluation of Mechanical 20. Olsen, J. T., and M. Maguire. “Shear Testing of Precast
Behaviour.” Composite Structures 105: 460–470. Concrete Sandwich Wall Panel Composite Shear Con-
nectors.” in The PCI Convention and National Bridge
10. Maximos, H. N., W. A. Pong, M. K. Tadros, and L. D. Conference: Proceedings, March 3–6, 2016, Nashville,
Martin. 2007. “Behavior and Design of Composite Pre- Tennessee. Chicago, IL: PCI.

21. Dutta, D., A. Jawdhari, and A. Fam. 2020. “A New
Studded Precast Concrete Sandwich Wall with Embedded
Glass-Fiber-Reinforced Polymer Channel Sections: Part
1, Experimental Study.” PCI Journal 65 (3): 78–99.
22. Jawdhari, A., and A. Fam. 2020. “A New Studded Precast

Concrete Sandwich Wall with Embedded Glass-Fiber-
Reinforced Polymer Channel Sections: Part 2, Finite
Element Analysis and Parametric Studies.” PCI Journal
65 (4): 51–70.
23. ASCE (American Society of Civil Engineers). 2017. Min-

imum Design Loads for Buildings and Other Structures.
ASCE/SEI 7-16. Reston, VA: ASCE.
24. ICC-ES (International Code Council Evaluation Service).

2017. “C-Grid Shear Connectors.” Report ESR-2953.
Brea, CA: ICC-ES.

About the authors Abstract
Mohamed K. Nafadi, PhD, is an This paper documents the testing of six 20 ft × 4 ft ×

assistant professor of structural 8 in. (6.1 m × 1.2 m × 203.2 mm) precast, prestressed
engineering at Assiut University in concrete sandwich panels constructed with continuous
Assiut, Egypt. He is a former rigid insulation and a carbon-fiber-reinforced poly-
visiting assistant professor at the mer grid shear transfer mechanism. All panels were
Department of Civil, Construction, identical except for foam type and were cast together
and Environmental Engineering at on the same prestressing bed. Three of the six panels
North Carolina State University (NCSU) in Raleigh. were fabricated with expanded polystyrene (EPS) foam
insulation, and the remaining three panels were fabri-
Gregory Lucier, PhD, is a research cated using sandblasted extruded polystyrene (XPS)
associate professor in the Depart- foam. For each group of three panels, one was tested to
ment of Civil, Construction, and failure as a control and two others were cycled 2 mil-
Environmental Engineering and lion times to 45% of their design ultimate load before
manager of the Constructed failure testing. The tested EPS panels all failed when
Facilities Laboratory at NCSU. the applied lateral load was greater than or equal to
100 lb/ft2 (4.79 kPa), which is 2.35 times their design
Tugce Sevil Yaman, PhD, is an load of 42.5 lb/ft2 (2.03 kPa). The tested XPS panels
assistant professor of structural all failed at the equivalent of 175 lb/ft2 (8.38 kPa) of
engineering at Mersin University applied lateral pressure, which is more than 4.0 times
in Mersin, Turkey. She is a their design load of 42.5 lb/ft2. All four panels subject-
former postdoctoral fellow in the ed to fatigue survived 2 million reverse-cyclic lateral
Department of Civil, Construc- load cycles without any visible signs of degradation.
tion, and Environmental Engi-
neering at NCSU. https://doi.org/10.15554/pcij66.5-01
Harry Gleich, PE, FPCI, FACI, is Keywords

the vice president of engineering
for Metromont Corp. in Green- Carbon-fiber-reinforced polymer grid, composite ac-
ville, S.C. He is the chair of the tion, concrete wythe, fatigue, sandwich panel.
PCI Technical Activities Council,
chair of the PCI Research and Review policy
Development Council, and past
chair of the PCI Precast Insulated Wall Panels Commit- This paper was reviewed in accordance with the
tee. He is also the former chair of ACI Committee 533, Precast/Prestressed Concrete Institute’s peer-review
Precast Panels, and former chair of ACI Committee process.
550, Precast Concrete Structures.
Reader comments
Sami Rizkalla, PhD, FPCI, FACI,
FASCE, FIIFC, FEIC, FCSCE, is Please address any reader comments to PCI Journal
a Distinguished Professor Emeri- editor-in-chief Tom Klemens at tklemens@pci.org or
tus in the Department of Civil, Precast/Prestressed Concrete Institute, c/o PCI Jour-
Construction, and Environmental nal, 8770 W. Bryn Mawr Ave., Suite 1150, Chicago, IL
Engineering at NCSU. 60631. J

Influence of hollow-core wall panels
on the cyclic behavior of different
types of steel framing systems
Parsa Monfaredi, Mehdi Nazarpour, and Abdoreza S. Moghadam
P
recast concrete hollow-core panels are common
components in modern buildings. Because of ease
and time savings with installation and finish, they are
useful in residential, commercial, warehouse, and indus-
trial buildings as exterior or interior partitions. In addition,
hollow-core panels are used as nonstructural components in
lateral-load-bearing systems.
Only a few studies have evaluated the seismic performance

■ This paper presents the results of an experimental of hollow-core wall panels. For example, Hamid and Ghani1
program that evaluated the seismic behavior of steel carried out an experimental study on seismic behavior of
frames with hollow-core wall panels under reversed hollow-core walls under biaxial lateral cyclic loading. In
cyclic loading and discusses the effects of the their study, two wall specimens were detailed with steel
hollow-core panels. armoring at the base-to-foundation interfaces, including
supplementary unbonded post-tensioned prestress, fuse
■ Three half-scale, single-story, single-bay steel frames bars, and mechanical energy dissipators. Bora et al.2 used a
were built and tested. One test specimen was a bare slotted-bolted friction joint to avoid brittle wall or anchorage
moment frame with a rigid connection frame, and the failure in thin hollow-core precast concrete panels. Holden
other two had hollow-core wall panels. Of the two et al.3 discussed the armoring details based on rocking
hollow-core specimens, one had a rigid connection behavior. Perez et al.4 introduced the seismic design require-
and the other had a pinned connection frame. ments for precast concrete walls with additional details such
as shear connectors and spiral reinforcement.
■ The test results indicated that hollow-core panels
within the frame could provide additional stiffness Previous studies have demonstrated that single hollow-core
and strength and have a positive impact on the over- walls are capable of resisting substantial lateral loads, de-
all seismic response of the structure. spite their lack of transverse shear reinforcement, if connec-
tion details are modified.5 There is a significant uncertainty
PCI Journal is published bimonthly by the Precast/Prestressed Concrete Institute, 8770 W. Bryn Mawr Ave., Suite 1150, Chicago, IL 60631. regarding the behavior of hollow-core wall panels under
Copyright © 2021, Precast/Prestressed Concrete Institute. The Precast/Prestressed Concrete Institute is not responsible for statements made lateral loading without any modification to the connection
details,6 and further research has been needed to investigate

the effect of these panels on lateral load capacity, failure surrounding moment-resistant steel frame. Figure 1 illustrates
modes, and the energy dissipation capacity of the steel frame. the geometry and dimensions of the test specimens, and
Table 1 summarizes the characteristics of the specimens.
The aims of this study were to evaluate how hollow-core
wall panels behaved in steel frames and present experimental Each specimen was defined by the type of frame and type of
results that could lead to the effective recognition of hollow- connection:
core wall behavior under cyclic lateral loading.
• Bare frame with rigid connection (BF-RC): Specimen
Research significance BF-RC was the control, or reference, specimen and con-
sisted of a bare frame with rigid reduced-beam-section
Despite the findings of the previous studies, a firm conclusion connections.
cannot be drawn regarding the influence of hollow-core panels
on the seismic performance of steel frames. The behavior of • Vertical panels with rigid connection (VP-RC): Specimen
the infill walls is associated with several factors, including VP-RC had two vertical, hollow-core panels in a rigid
the rigidity of the beam-to-column connection. In particular, connection frame, or moment frame.
in gravity frames, the hollow-core panels have unidentified
response to lateral loading that had not been examined in the • Vertical panels with pinned connection (VP-PC): Spec-
literature. imen VP-PC had two vertical, hollow-core panels in a
pinned connection frame, or gravity frame.
To investigate the in-plane behavior of steel frames with rigid
or even pinned connections of beams to columns, three iden- For specimen VP-RC and specimen VP-PC, the panels were
tical half-scale steel frames were built and tested in the same separated from the steel column by 14 mm (0.55 in.) verti-
manner. To examine the influence of inclusion of hollow-core cal gaps. Specimen VP-PC represents a single-bay infilled
panels, the steel moment-resisting frame and the gravity frame in a conventional steel frame with shear walls so that
frame infilled with panels were compared with a frame with the beam-to-column connections were not rigid and the steel
no infill walls. This study offers experimental results focused frame was assumed to resist only the gravity loads.
on overall response, modes of failure, energy dissipation,
and stiffness-degrading behavior of the frames. The results Hollow-core panels were manufactured with longitudinal
of this study can be used in modifying and redesigning the cores and two layers of high-strength bonded pretensioning
hollow-core walls so that they can be considered the structural strands, filled with high-strength and very-low-slump con-
walls in regions with moderate to high seismic activity. crete, and formed by extruders. Transverse reinforcement was
not included. Box-section steel columns and built-up steel
Experimental program plate beams were constructed by a steel fabricator and then
assembled into a frame with the hollow-core panels. Longitu-
Specimens dinal fillet welds joined four steel plates to form the steel box
columns (Fig. 1).
Three identical half-scale, single-story, single-bay steel
frames were designed and built in a laboratory setting. One Materials
assembly was a bare moment frame, while the others had
wall units. All specimens were 1974 mm (77.7 in.) long × Standard test methods were used to determine the structural
1446 mm (56.9 in.) high. Infill panels consisted of 1260 × 880 steel properties and the concrete compressive strength accord-
× 150 mm (49.6 × 34.6 × 5.9 in.) hollow-core panels within a ing to ASTM E8/E8M7 and ASTM C39/C39M,8 respectively.
Table 1. Summary of specimens
Specimen
Specimen Type of frame Type of connection Type of panels
dimensions, mm
BF-RC Bare steel frame Rigid n/a 1974 × 1446
Steel moment Hollow-core with longitudinal

VP-RC frame with vertical Rigid cores and two layers of pre- 1974 × 1446
panels stressed strands
Hollow-core with longitudinal

Steel gravity frame
VP-PC Pinned cores and two layers of pre- 1974 × 1446
with vertical panels
stressed strands
Note: BF-RC = bare frame with rigid connection; n/a = not applicable; VP-PC = vertical panels in pinned connection frame; VP-RC = vertical panels in
rigid connection frame. 1 mm = 0.0394 in.

Beam and column details
Specimen VP-RC and hollow-core wall section
Figure 1. Dimensions and geometry of test specimens. Note: All dimensions are in millimeters. PL = plate; R = radius;
VP-RC = vertical panels in rigid connection frame. 1 mm = 0.0394 in.

ASTM A416/A416M9 testing was performed on the strands. A constant low axial load was applied to the specimens by
Table 2 summarizes the material properties. circular solid steel rods. The axial compression load in each
column was 80 kN (18 kip). In addition, four 160 mm (6.3 in.)
Test setup and instrumentation deep channels were attached to the beams and columns to
prevent out-of-plane movement of the panels. To ensure that
Figure 2 shows the schematic arrangement of the experimental in-plane loading was imposed, lateral restraints were set up at
setup with an in-plane actuator attached to the reaction frame. the top corners of the specimens.
Each column was fixed to the reaction frame by four anchored
bolts. Lateral load was applied by a 1000 kN (225 kip) actua- The experiments were conducted in drift control. In a dis-
tor, and the force was measured by an in-series load cell. Each placement/drift control test, the displacement is the indepen-
specimen was loaded laterally with forces applied through the dent variable and the load reaction is the dependent variable.
loading beam, which was a stiffened 220 mm (8.7 in.) deep Eu- The drift ratio was calculated by dividing the difference
ropean wide-flange beam IPB 220. A lateral force was applied between the displacements at the top and bottom of the steel
at the center of the loading beam, which was located 1945 mm moment frame by the column height. Two full cycles were ap-
(76.6 in.) above the reaction frame. plied at each target drift level during the test. Lateral displace-
ments of the specimens were recorded with linear variable
Table 2. Material properties differential transformers and an image processing system that
tracked the position of 70 points referred to as markers. The
Steel Concrete Strand color pattern matching algorithm and the mean shift tracking
Seven-wire steel
algorithm were applied in image processing technique. All
Fy = 293 MPa fc = 58 MPa markers were placed on a 200 × 200 mm (7.9 × 7.9 in.) grid
strand
on the panels, columns, and beams.
1860 MPa Grade LR
Fu = 420 MPa w/c = 0.4
ASTM A416 The loading procedure was limited by the range of amplitude of
the hydraulic actuator, which was 120 mm (4.7 in.) in both the
E = 205 GPa Weight = 2403 kg/m3 Diameter = 8.35 mm
positive direction (pulling) and the negative direction (push-
Note: E = modulus of elasticity of steel; fc = concrete compressive ing). Therefore, the applied displacements were the same for
strength; Fu = ultimate tensile strength of steel; Fy = yield strength of corresponding cycle numbers for all specimens, but the applied
steel; LR = low relaxation; w/c = water-cement ratio. 1 mm = 0.0394 in.; loads varied. In other words, loading tests were stopped at the
1 MPa = 0.145 ksi; 1 GPa = 145 ksi; 1 kg/m3 = 1.6875 lb/yd3. maximum possible displacement of the hydraulic actuator,
which corresponded to an ultimate drift ratio of 8.3%.
Figure 2. Illustration of the experimental setup showing specimen VP-RC. Note: VP-RC = vertical panels in rigid connection
frame.

Loading history
All specimens were tested under quasi-static cyclic lateral

loading. The in-plane lateral load was a series of displace-
ment-controlled cycles in pull (positive displacement) and
push (negative displacement) imposed by the actuator. The
loading history is shown in Fig 3. Each complete load cycle
consisted of one half cycle in each direction. This nominal
loading history is similar to that specified by the Federal
Emergency Management Agency’s (FEMA’s) Interim Testing
Protocols for Determining the Seismic Performance Charac-
teristics of Structural and Nonstructural Components.10
Test results and discussion
Overall response
Figure 4 plots the responses of specimens BF-RC, VP-RC,

and VP-PC based on applied load versus drift ratio. Table 3 Specimen BF-RC
summarizes the results. A comparison of load capacity
showed that specimen VP-RC resisted higher loads than the
other specimens; on average, its load capacity was 40% great-
er than that for specimen BF-RC in both directions (34% and
46% in the positive and negative direction, respectively). The
peak load applied to specimen VP-PC exceeded the BF-RC
specimen peak load by approximately 12%. From these re-
sults, we can conclude that the lateral load capacity of a frame
is enhanced by hollow-core wall panels.
Figure 4 illustrates that hysteresis loops for frames with rigid

connection exhibited less pinching than the conventional
gravity frame with pinned connection. The pinching was most
pronounced in specimen VP-PC due to the lack of a lateral-
load-resisting system. Even though specimen BF-RC had
more-stable and less-pinched hysteretic loops, its load-bearing Specimen VP-RC
capacity remained nearly constant after a drift ratio of 4%.
A cyclic degradation of strength occurred in the frames with
hollow-core panels such that the ultimate load was less than
the peak load described in Table 3.
Specimen VP-PC
Figure 4. Experimental hysteresis curves. Note: BF-RC = bare

frame with rigid connection; VP-RC = vertical panels in rigid
Figure 3. Loading history for each specimen in the experi- connection frame; VP-PC = vertical panels in pinned connec-
mental program. Note: 1 mm = 0.0394 in. tion frame. 1 mm = 0.0394 in.; 1 kN = 0.225 kip.

Table 3. Peak loads, ultimate loads, and measured drift ratios for each specimen
Specimen Loading direction Peak load, kN δd, % Ultimate load, kN δu, %
+ 491 8.3 491 8.3

BF-RC
– –503 –8.3 –503 –8.3
+ 658 8.3 658 8.3

VP-RC
– –735 –5.9 –687 –8.3
+ 542 5.9 538 8.3

VP-PC
– –572 –3.0 –406 –8.3
Note: BF-RC = bare frame with rigid connection; VP-PC = vertical panels in pinned connection frame; VP-RC = vertical panels in rigid connection frame;
δd = drift ratio at peak load; δu =ultimate drift ratio at the last testing cycle. 1 kN = 0.225 kip.
The VP-RC specimen was able to maintain sufficient strength about 50% more than those for specimen BF-RC. As for the
at large deformations (and therefore had high ductility) and ultimate drift ratio, hollow-core panels in specimen VP-RC
resisted high loads at the ultimate drift ratio. The loading had severe damage, which led to a significant reduction of
had to be stopped due to the limits of the hydraulic actuator damage in its surrounding steel frame compared with the
stroke. Nevertheless, specimen VP-RC had a greater chance other specimens’ steel frames.
of additional load resistance compared with specimen BF-RC.
Specimen VP-RC showed a rocking behavior under cyclic
Modes of failure lateral loading (Fig. 6). In the final testing cycle, the maxi-
mum movements of the panels at the bottom corner were 15
Significant differences in the failure modes were observed and 95 mm (0.6 and 3.7 in.) in the horizontal and vertical
between the bare frame and the infilled frames. In this regard, directions, respectively; these values were determined using
the types of damage were divided into two categories: damage image processing techniques (Fig. 5). A similar behavior was
to the structural components and damage to the nonstructural observed in specimen VP-PC.
components, which are related to the steel frame and hollow-
core panel, respectively. Table 4 presents definitions of the Corner crushing of concrete panels occurred due to the rock-
failure modes and the corresponding loads and drift ratios. ing behavior of the wall panels, which led to the high stress
concentrations at each corner of the compression diagonal. As
Figure 5 illustrates the observed failure modes and damage the drift ratio increased, corner crushing became more pro-
for each specimen. The grid of markers for the image pronounced. Hollow-core panels of specimen VP-PC contributed
cessing system can also be seen in Fig. 5. A comparison of to the load-bearing system from the beginning of loading.
the failure mechanisms revealed the advantage of hollow-core Consequently, compared with specimen VP-RC, compression
panels in the postponement of plastic hinge formation and struts in specimen VP-PC formed earlier and, in turn, corner
reduction of damage severity in higher drift ratios. In relation crushing occurred in that specimen at a lower load.
to the failure of beam-to-column rigid connections, the peak
load and corresponding drift ratio for specimen VP-RC were As noted, specimen BF-RC failed in modes of structural
Table 4. Summary of failure modes, peak applied load, and drift ratio
Specimen BF-RC Specimen VP-RC Specimen VP-PC

Element Sign Failure mode Load, Drift Load, Drift Load, Drift
kN ratio, % kN ratio, % kN ratio, %
Fracture of weld along

(a) –412 –2.1 –686 –4.2 n/a n/a
Steel frame beam-to-column connection
(b) Column base connection failure –486 –5.8 633 5.7 –565 –3.4
Corner crushing of concrete

(c) n/a n/a –568 –3.0 379 3.0
Hollow-core panels
wall panel Concrete crushing between
(d) n/a n/a 658 8.2 539 8.2
two panels
Note: BF-RC = bare frame with rigid connection; n/a = not applicable; VP-PC = vertical panels in pinned connection frame; VP-RC = vertical panels in
rigid connection frame. 1 kN = 0.225 kip.

Specimen BF-RC
Specimen VP-RC
Specimen VP-PC
Figure 5. Observed failure modes and damage for specimen BF-RC, specimen VP-RC, and specimen VP-PC. Note: (a) = fracture
of weld along beam-column connection; (b) = column base connection failure; BF-RC = bare frame with rigid connection; (c) =
corner crushing of concrete panels; (d) = concrete crushing between two panels; VP-PC = vertical panels in pinned connection
frame; VP-RC = vertical panels in rigid connection frame.

Figure 6. Rocking behavior of the VP-RC specimen at the ultimate drift ratio. Note: All dimensions are in millimeters. VP-RC =
vertical panels in rigid connection frame. 1 mm = 0.0394 in.
components and experienced lower load capacity, whereas the brittleness: the greater the difference is, the more brittle the
identical frame with hollow-core panels (VP-RC) had panel-re- behavior will be. For example, specimen VP-PC showed a
lated failure modes and resulted in higher load-carrying capac- substantial difference (150%) between two backbone curves
ity. This means that the hollow-core panels contributed to the due to the behavior of the pinned connection and the severe
reduction of the structural damage of the steel moment frame. damage to the column base.
Backbone curve Table 5 summarizes the values of the peak loads and the
corresponding drift ratios obtained from both procedures.
Figure 7 shows the experimental backbone curves for the Specimen VP-PC showed the greatest difference in peak loads
three specimens, which were derived from the American
Society of Civil Engineers/Structural Engineering Institute Table 5. Peak load and corresponding drift ratio
(ASCE/SEI) 41-17, Seismic Evaluation and Retrofit of Exist- for the backbone curves
ing Buildings,11 and ASCE/SEI 41-06, Seismic Rehabilitation
Peak load, kN
of Existing Buildings.12 As specified in ASCE/SEI 41-17, Drift Difference,
the peak points of the first cycle at each displacement incre- Specimen ASCE/ ASCE/
ratio, % %
ment were connected to form the backbone curves, which SEI 41-06 SEI 41-17
are shown by solid black piecewise linear lines in Fig. 7. In
5.9 474 474 0
FEMA guidance on the seismic rehabilitation of buildings13,14 BF-RC
and ASCE/SEI 41-06, the backbone curve is drawn through –5.9 –463 –486 5
the intersection of the first cycle curve for the deformation
5.9 521 633 21
step i with the second cycle curve of the deformation step VP-RC
(i-1) for all steps i. The dashed lines are representative of the –5.8 –538 –735 37
procedure proposed in ASCE/SEI 41-06.
5.8 340 542 59
VP-PC
There were differences in the values and trends of the two –5.7 –199 –498 150
curves (Fig. 7). Regarding the trends, implementation of the
Note: ASCE = American Society of Civil Engineers; BF-RC = bare frame
procedures suggested by ASCE/SEI 41-06 led to the appear-
with rigid connection; SEI = Structural Engineering Institute; VP-PC =
ance of severe cyclic degradation of strength in the backbone
vertical panels in pinned connection frame; VP-RC = vertical panels in
curve of the VP-PC specimen. The differences between
rigid connection frame. 1 kN = 0.225 kip.
two cyclic backbone curves can be used as an indicator of

between the two backbone curves. Therefore, the behavior of
specimen VP-PC was more brittle than the behavior of other
specimens. In contrast, the backbone curves were almost the
same in specimen BF-RC, so this specimen’s behavior was
less brittle than the behavior in the other specimens.
Figure 8 presents the backbone curves of the specimens

derived from ASCE/SEI 41-17.11 The maximum strength
of specimen BF-RC was 66% and 89% of the maximum
strengths of specimens VP-RC and VP-PC, respectively.
Moreover, the strength of specimen VP-RC was up to 48%
greater than the corresponding values in specimen VP-PC.
The difference between the behaviors of specimens VP-RC
and VP-PC was mainly attributed to the change in the con-
nection's rigidity of the surrounding frames. In other words,
Specimen BF-RC pinned connections, not rigid connections, led to less contri-
bution of infill in the steel frame.
Energy dissipation
The ability to dissipate energy is one of the important pa-

rameters for evaluating the seismic performance of structural
walls. The energy dissipation is calculated on the basis of area
enclosed by the load-displacement hysteresis loops. Figure 9
illustrates the stacked value of energy dissipation capacity
versus drift ratio.
At drift ratios less than or equal to 2.1%, the energy dissi-

pation capacity of specimen VP-RC was similar to that of
specimen BF-RC. This similarity is due to the gaps between
panels and columns. Consequently, dissipating energy was
limited by the moment frame. At drift ratios greater than
Specimen VP-RC
2.1%, gradual closing of the vertical gaps between panels
Specimen VP-PC
Figure 7. Cyclic backbone curves obtained from ASCE/SEI 41-

17, Seismic Evaluation and Retrofit of Existing Buildings, and Figure 8. Comparison of the backbone curves obtained from
ASCE/SEI 41-06, Seismic Rehabilitation of Existing Buildings. ASCE/SEI 41-17, Seismic Evaluation and Retrofit of Existing
Note: ASCE = American Society of Civil Engineers; BF-RC = Buildings. Note: ASCE = American Society of Civil Engineers;
bare frame with rigid connection; SEI = Structural Engineering BF-RC = bare frame with rigid connection; SEI = Structur-
Institute; VP-PC = vertical panels in pinned connection frame; al Engineering Institute; VP-PC = vertical panels in pinned
VP-RC = vertical panels in rigid connection frame. 1 kN = connection frame; VP-RC = vertical panels in rigid connection
0.225 kip; 1 mm = 0.0394 in. frame. 1 kN = 0.225 kip; 1 mm = 0.0394 in.

Stiffness-degrading behavior
Figure 10 plots the relationship of deduced values of

cyclic stiffness to drift ratios. At the initial loading cycle, the
hollow-core panels demonstrated no contribution in initial
stiffness. Therefore, the initial stiffness values of specimens
VP-RC and BF-RC were nearly the same (approximately
17 kN/mm [96.9 kip/in.]). By comparison, the initial stiffness-
es of specimens VP-RC and BF-RC were 25% greater than
that of specimen VP-PC.
At drift ratios equal to or less than 2.1%, the overall stiffness

of the system was only associated with the steel frame mem-
bers; this was an expected finding due to the gaps between
panels and columns. Thus, for these drift ratios, the stiffness
values of the VP-RC and BF-RC specimens exhibited the
same pattern. At drift ratios greater than 2.1%, the gaps in
Figure 9. Relationship of the stacked value of cumulative specimen VP-RC gradually closed and ultimately led to
energy-dissipation capacity to drift ratio. Note: BF-RC = bare stiffness that was 44% greater than the stiffness of specimen
frame with rigid connection; VP-PC = vertical panels in pinned BF-RC. Because of the failure modes, gradual degradation in
connection frame; VP-RC = vertical panels in rigid connection
frame. 1 kN-m = 8.85 kip-in. overall stiffness occurred in all specimens.
and columns, formation of compression struts, and rocking As for specimen VP-PC, because of the pinned connec-
behavior of hollow-core wall panels in specimen VP-RC tion, its initial stiffness was 20% less than that of the other
caused a slight increase in energy dissipation comparison with specimens. The cyclic stiffness of specimen VP-PC was up
specimen BF-RC. to 11% greater than that of specimen BF-RC when the drift
ratio was in the range of 3.5% to 7.0%, where failure of the
The overall energy-dissipation capacity of the VP-RC spec- column base connection in specimen VP-PC did not cause
imen (270,745 kN-mm [2396 kip-in.]) was approximately significant stiffness degradation. Clearly, the hollow-core
4% and 53% higher than the energy-dissipation capacities of panels were effective in preventing severe degradation of
specimens BF-RC and VP-PC, respectively (Fig. 9). For spec- stiffness.
imen VP-PC, the columns and hollow-core panels both par-
ticipated in energy dissipation when the drift ratio range was One of the major effects of hollow-core panels is a greater
between 2.1% and 5.9%. Subsequently, at drift ratios greater stiffness value compared with a bare frame when the drift
than 5.9%, the energy dissipation value remained constant due ratio is between 2% and 8%. The greater stiffness value is
to severe damage to the specimen’s column bases and panels. due to the rocking behavior of the hollow-core panels and
In fact, the frame was no longer able to dissipate energy and a better load transfer mechanism. Steel frames infilled with
the main energy-dissipating element was the hollow-core
panels.
For drift ratios less than or equal to 5.9%, the over-

all energy-dissipation capacity of the VP-PC specimen
(118,605 kN-mm [1050 kip-in.]) was 20% less than that of
specimen BF-RC and 23% less compared to specimen VP-
PC. This finding demonstrates the considerable effects of the
inclusion of hollow-core walls. As noted, the hollow-core
panels contributed to the energy-dissipation mechanism in
the gravity frame (VP-PC) and postponed the frame’s main
damages until the drift ratio exceeded 5.9%.
In conclusion, the presence of hollow-core panels increased

the frames’ energy-dissipating capability. These types of
panels can act as dissipation devices, strongly reducing the
damage in the structural elements, even though they are as-
sumed to be nonstructural walls. Hence, hollow-core panels Figure 10. Relationship of cyclic stiffness response to drift
could help improve seismic behavior of the steel moment ratio. Note: BF-RC = bare frame with rigid connection; VP-PC
frame. = vertical panels in pinned connection frame; VP-RC = vertical
panels in rigid connection frame. 1 kN/mm = 5.71 kip/in.

hollow-core panels behave as a braced frame with the panels significant amount of the normalized energy dissipation when
forming diagonal compression struts. the drift ratio was less than or equal to 5%. This improved
behavior was caused by the inclusion of hollow-core panels.
Less structural stiffness results in a longer natural period of Finally, when the drift ratio was greater than 5%, the normal-
vibration and, consequently, lower seismic force demand.15 ized dissipated energy of specimen VP-PC decreased due to
In contrast, greater initial stiffness causes more seismic force severe damage in columns and panels.
demand, which is undesirable. This is one of the reasons to
have gaps between the infills and the structural elements. The Idealized backbone
gaps are to prevent the impact of the panels on the increment
of initial stiffness. Following that, the same initial stiffness of Idealized backbone curves were obtained based on the
both bare and infilled frames could be considered an advan- proposed procedures of ASCE/SEI 41-17.11 Because of the
tage of these gaps. loading limitation of the hydraulic jack, which had a maxi-
mum displacement of 120 mm (4.7 in.) (target drift ratio of
Normalized dissipated energy 8.3%), the experimental tests did not enter the degradation
phase; therefore, the idealized curves are bilinear. Figure 12
To compare the reduced energy dissipation per cycle in this illustrates the bilinear idealization of the backbone response
study, the approach proposed by Kakaletsis and Karayannis curve as a force-displacement relationship for the specimens.
was applied.16 This approach had previously been used by The first line segment of the idealized force displacement
researchers such as Tasnimi and Mohebkhah17 and Emami and curve begins at the origin and has a slope equal to the effec-
Mohammadi.18 The energy-dissipation capacity per cycle was tive lateral stiffness.
normalized by the peak-to-peak displacement 2Δ for that cycle,
which was plotted against the imposed drift ratio (Fig. 11). Table 6 presents the key parameters used to derive the
2Δ refers to the lateral displacement of the specimen in each idealized backbone curves. The effective lateral stiffness Ke,
complete cycle, which is in both directions. The distribution of applied load at yielding Vy, applied load at ultimate strength
plastic hinges in the steel frame and other damage to the system Vd, lateral displacement at yielding Δy, lateral displacement at
caused the dissipating energy to gradually increase. As a result, ultimate strength Δd, drift ratio at yielding δy, and drift ratio at
the normalized energy dissipation of the specimens increased as ultimate strength δd are shown for both positive and negative
the drift ratio increased. The values of normalized energy dissi- loads. The effective (secant) stiffness was calculated as the
pation of the BF-RC and VP-RC specimens were very similar slope of the line joining a yielding point on the idealized
when drift ratios were in the range of 1.0% to 5.0% (Fig. 11). curve to the origin.
However, after this stage, the normalized dissipated energy in
specimen VP-RC was more than that of specimen BF-RC due In comparison to the bare frame, the specimens with
to the presence of the infill panel in specimen VP-RC. hollow-core panels experienced greater strength and stiffness
capacity in the plastic range. This benefit of using hollow-core
Although steel frames with pinned connections have low panels can improve the seismic performance level of the struc-
energy-dissipation capacity, specimen VP-PC experienced a ture. As pointed out earlier, hollow-core walls did not contrib-
Figure 11. Comparison of the ratio of energy dissipation to

displacement 2Δ per cycle against the drift ratio. Note: BF-RC Figure 12. Idealized backbone curves. Note: BF-RC = bare
= bare frame with rigid connection; VP-PC = vertical panels frame with rigid connection; VP-PC = vertical panels in pinned
in pinned connection frame; VP-RC = vertical panels in rigid connection frame; VP-RC = vertical panels in rigid connection
connection frame. 1 kN-mm/mm = 0.225 kip-in/in. frames. 1 mm = 0.0394 in.; 1 kN = 0.225 kip.

Table 6. Key parameters of each specimen for idealization backbone curves
Specimen Ke, kN/mm α1 Vy, kN Vd, kN Δy, mm Δd, mm δy, mm δd, mm
12.9 0.05 430 491 33.4 120 2.31 8.30

BF-RC
–17.6 0.02 –469 –503 –26.6 –120 –1.84 –8.30
11.8 0.22 440 658 37.2 120 2.57 8.30

VP-RC
–16.3 0.22 –545 –735 –33.4 –85.8 –2.31 –5.93
11.6 0.28 362 542 30.8 85.8 2.13 5.93

VP-PC
–14.1 0.12 –523 –565 –37.2 –61.2 –2.57 –4.23
Note: BP-RC = bare frame with rigid connection; Ke = effective lateral stiffness of the specimen; Vd = applied load at ultimate strength; Vy = applied load
at yielding; VP-PC = vertical panels in pinned connection frame; VP-RC = vertical panels in rigid connection frame; α1 = positive post-yield slope ratio
equal to the positive post-yield stiffness divided by the effective stiffness; δd = drift ratio at ultimate strength; δy = drift ratio at yielding; Δd = lateral dis-
placement at ultimate strength; Δy = lateral displacement at yielding. 1 mm = 0.394 in.; 1 kN = 0.225 kip; 1 kN/mm = 5.71 kip/in.
ute to initial stiffness of the frame in the elastic range due to the • The presence of hollow-core panels resulted in improved
gaps between panels and columns. Subsequently, the responses structural performance by transferring the failure events
of the specimens were nearly identical up to yield. Therefore, from the steel frame to the panels. This demonstrates
the yielding point of both the VP-RC and VP-PC specimens that the panels are not, as they are often considered to be,
were almost the same as the yielding point of specimen BF-RC. nonstructural components.
After this stage, the presence of hollow-core panels prevented
the propagation of failures in the steel frame by rocking behav- • Separating panels from the steel column with 14 mm
ior. Thus, the slopes of second segments of idealized curves (0.55 in.) vertical gaps caused an identical initial stiff-
(post yield) of the VP-RC and VP-PC specimens were, respec- ness of infilled and bare frames. One of the major effects
tively, about 4 and 5 times that of specimen BF-RC. of hollow-core panels was a greater stiffness value
compared with the bare frame when the drift ratio was
In order to highlight the advantages of hollow-core panels on between 2% and 8%. The greater stiffness was caused by
the overall seismic behavior of steel frames, a comparison with the panels’ contribution to the load-bearing system, the
conventional infills would be beneficial. Conventional infills, rocking behavior of the hollow-core panels, and a better
such as masonry infill walls, must be properly connected with load transfer mechanism.
the surrounding frame; however, the interaction between the
infill wall and the frame may or may not be beneficial for the • The contribution of hollow-core panels in lateral load
seismic behavior of the structure.18 One of the disadvantages of response of the steel frames when the drift ratio was
this interaction is the greater initial stiffness of the frame, which greater than 2% led to a higher load-bearing capacity of
leads to increased seismic demands. This may have an effect on the frames. Compared with specimen BF-RC, specimen
the elastic behavior of the frame and change the yielding point VP-RC resisted loads up to 40% higher on average in
area. This behavior of masonry-infilled frames in earthquake both directions.
conditions was pointed out by Mohammadi and Emami.19 In
contrast, using hollow-core walls with a gap between panel and • Using hollow-core walls with a 14 mm (0.55 in.) gap
column resulted in an inconsequential change in the yielding between panel and column resulted in an inconsequential
point and improved plastic behavior of the frame. change in yielding point of the bare moment frame; a
significant increment of the second line segment of the
It should be noted that the gap between the panel and column idealized backbone curve, which represents a positive
is one of the most effective parameters to improve structural post-yield slope (α1Ke), increasing the lateral load capac-
performance levels; however, further research must be carried ity up to 46%; and overall improved plastic behavior of
out to investigate the effective value of the gaps. the frame.
Conclusion • When the drift ratio was less than or equal to 2%, the
steel frame was the main element of dissipating energy.
This paper described the effect of hollow-core infills on the When the drift ratio exceeded 2%, the effect of the frame
cyclic behavior of both steel moment and gravity frames. The on the dissipation energy mechanism decreased gradually
behavioral characteristics of the specimens were quantified as the drift ratio and frame damage increased, and the
with an emphasis on lateral load capacity, ductility, strength energy-dissipation contribution of the hollow-core panels
degradation attributes, hysteretic energy dissipation, and po- increased. When the drift ratio was greater than 5.9%, the
tential failure modes. The following conclusions were drawn hollow-core walls exhibited a significant improvement in
from this investigation: energy-dissipation capacity.

• The difference between the overall stiffness and strength 9. ASTM Subcommittee A01.05. 2012. Standard Spec-
of the infilled frame with pinned connection (specimen ification for Steel Strand, Uncoated Seven-Wire for
VP-PC) and the bare moment frame with rigid connec- Prestressed Concrete. ASTM A416/A416M-12. West
tions (specimen BF-RC) was not considerable. This Conshohocken, PA: ASTM International.
finding demonstrates the significant effect that the type of
connections has on stiffness and load-bearing capacity in 10. FEMA (Federal Emergency Management Agency).
steel frames. In summary, the stiffness and strength of the 2007. Interim Testing Protocols for Determining the
infilled frame with rigid connections (specimen VP-RC) Seismic Performance Characteristics of Structural and
were, respectively, up to 43% and 28% greater than the Nonstructural Components. FEMA 461. Washington,
corresponding values in specimen VP-PC. DC: FEMA.
This experimental work determined that using hollow-core 11. ASCE (American Society of Civil Engineering)/SEI
infills can enhance the seismic performance of steel frames sub- (Structural Engineering Institute). 2017. Seismic Evalua-
jected to the large deformations caused by severe earthquakes, tion and Retrofit of Existing Buildings. ASCE/SEI 41-17.
despite the fact that they are regarded as nonstructural elements. Reston, VA: ASCE/SEI.
References 12. ASCE/SEI. 2007. Seismic Rehabilitation of Existing

Buildings. ASCE/SEI 41-06. Reston, VA: ASCE/SEI.
1. Hamid, N. H., and K. D. Ghani. 2013. “Seismic Behav-
ior of a Precast Hollow Core Wall under Biaxial Lateral 13. FEMA. 1997. NEHRP Guidelines for the Seismic Re-
Cyclic Loading.” WIT Transactions on the Built Environ- habilitation of Buildings. FEMA 273. Washington, DC:
ment 134: 815–825. FEMA.
2. Bora, C., M. G. Oliva, S. D. Nakaki, and R. Becker. 2007. 14. FEMA. 2000. Prestandard and Commentary for the Seis-
“Development of a Precast Concrete Shear-Wall System mic Rehabilitation of Buildings. FEMA 356. Washington,
Requiring Special Code Acceptance.” PCI Journal 52 (1): DC: FEMA. https://www.nehrp.gov/pdf/fema356.pdf.
122–135.
15. Sharbatdar, M. K., and M. Saatcioglu. 2009. “Seismic
3. Holden, T., J. Restrepo, and J. B. Mander. 2003. “Seismic Design of FRP Reinforced Concrete Structures.” Asian
Performance of Precast Reinforced and Prestressed Con- Journal of Applied Sciences 2 (3): 211–222.
crete Walls.” Journal of Structural Engineering 129 (3):
286–296. 16. Kakaletsis, D. J., and C. G. Karayannis. 2008. “Influ-
ence of Masonry Strength and Openings on Infilled R/C
4. Perez, F. J., S. Pessiki, and R. Sause. 2004. “Seismic De- Frames under Cycling Loading.” Journal of Earthquake
sign of Unbonded Post-tensioned Precast Concrete Walls Engineering 12 (2): 197–221.
with Vertical Joint Connectors.” PCI Journal 49 (1): 58–79.
17. Tasnimi, A. A., and A. Mohebkhah. 2011. “Investiga-
5. Hamid, N. H. A., and J. B. Mander. 2006. “Experimen- tion on the Behavior of Brick-Infilled Steel Frames with
tal Study on Bi-lateral Seismic Performance of Precast Openings, Experimental and Analytical Approaches.”
Hollow Core Wall Using Shaking Table.” In Proceedings Engineering Structures 33 (3): 968–980.
of the 10th East Asia-Pacific Conference on Structural
Engineering and Construction (EASEC 2010) (109–114). 18. Emami, S. M. M., and M. Mohammadi. 2016. “Influence
Bangkok, Thailand: Asian Institute of Technology. of Vertical Load on In-plane Behavior of Masonry In-
filled Steel Frames.” Earthquakes and Structures 11 (4):
6. Nazarpour, M., P. Monfaredi, and A. S. Moghadam. 609–627.
2019. “Experimental Evaluation of Hollow-Core Wall
Orientation in Steel Moment Frame.” PCI Journal 64 (3): 19. Mohammadi, M., and S. M. M. Emami. 2019. “Multi-bay
92–103. and Pinned Connection Steel Infilled Frames; an Exper-
imental and Numerical Study.” Engineering Structures
7. ASTM Subcommittee E28.04. 2009. Standard Test Meth- 188: 43–59.
ods for Tension Testing of Metallic Materials. ASTM
E8/E8M-09. West Conshohocken, PA: ASTM Interna- Notation
tional.
E = modulus of elasticity of steel
8. ASTM Subcommittee C09.61. 2001. Standard Test
Method for Compressive Strength of Cylindrical Concrete fc' = concrete compressive strength
Specimens. ASTM C39/C39M-01. West Conshohocken,
PA: ASTM International. Fu = ultimate tensile strength of steel

Fy = yield strength of steel
i = cycle number
Ke = effective lateral stiffness
Vd = applied load at ultimate strength
Vy = applied load at yielding
α1 = positive post-yield slope ratio equal to the positive

post-yield stiffness divided by the effective stiffness
δd = drift ratio at ultimate strength
δu = drift ratio at ultimate load
δy = drift ratio at yielding
Δ = lateral displacement in each half cycle
Δd = lateral displacement at ultimate strength
Δy = lateral displacement at yielding

Parsa Monfaredi is a structural Hollow-core precast concrete panels are widely used
engineer in the Department of in commercial, industrial, and warehouse buildings
Structural Engineering at the as exterior or interior partitions. Because these infill
International Institute of Earth- walls are considered to be nonstructural elements,
quake Engineering and Seismolo- their interaction with the surrounding frame during
gy (IIEES) in Tehran, Iran, where an earthquake has been mostly neglected. This paper
he also received his MSc in the describes experimental research that evaluated the
field of precast concrete seismic behavior of different types of steel frames with
hollow-core wall panels and steel structures. hollow-core infill under reversed cyclic loading and
discusses the effects of the hollow-core panels. Three
Mehdi Nazarpour received his identical half-scale steel frames were built and tested
PhD from the IIEES Department in the same manner. A steel moment-resisting frame
of Structural Engineering in and a gravity frame with hollow-core panels were
Tehran in 2019. The main fields of compared with a frame with no infill walls. The test
his research include precast results indicated that under moderate to high shaking
concrete wall panels; high- intensity, hollow-core panels rocked within the frame
strength concrete; squat reinforced could provide additional stiffness, strength, and energy
concrete shear walls; and repair, dissipation to the bare frame, as well as better flexibil-
rehabilitation, and retrofitting of structures. ity and ductility. A comparison of failure mechanisms
revealed the advantage of hollow-core panels in the
Abdoreza S. Moghadam received postponement of plastic hinge formation and reduction
his BS in civil engineering in 1987 of structural damage severity at higher drift ratios. This
and MS in structural engineering study shows that hollow-core walls can have a positive
in 1991 from Tehran University impact on the overall seismic response of the structure,
and his PhD in earthquake despite the fact that they are regarded as nonstructural
engineering from McMaster elements.
University in Hamilton, ON,
Canada, in 1999. His research interests include https://doi.org/10.15554/pcij66.5-02
earthquake engineering, evaluation and design of tall
buildings, development of building codes, seismic Keywords
retrofitting of structures, and effects of three-dimen-
sional modeling in building evaluation and design. He Failure mode, hollow-core, multipanel hollow-core
is currently an associate professor at IIEES in Tehran. wall, quasi-static cyclic loading, seismic behavior.
Review policy
This paper was reviewed in accordance with the

Precast/Prestressed Concrete Institute’s peer-review
process.
Reader comments
Please address any reader comments to PCI Journal

editor-in-chief Tom Klemens at tklemens@pci.org or
Precast/Prestressed Concrete Institute, c/o PCI Jour-
nal, 8770 W. Bryn Mawr Ave., Suite 1150, Chicago, IL
60631. J

Seismic design and analysis
of precast concrete
buckling-restrained braced frames
Shane Oh, Yahya C. Kurama, Jon Mohle, and Brandt W. Saxey
B
uckling-restrained braced frames are a type of lateral
force-resisting system currently used primarily for
steel buildings in moderate and high seismic zones.
These structures resist lateral loads using buckling-restrained
braces placed diagonally and connected to the beams and
columns of the frame in each story. Although buckling-
restrained braced frames are visually similar to conventional
■ The study described in this paper investigated the concentrically braced frames, the unique characteristics of
lateral load behavior and design of precast concrete buckling-restrained braces result in distinct behavior under
buckling-restrained braced frames and the feasibility seismic loads. Buckling-restrained braces are typically com-
of their use in seismic regions. posed of a high-ductility steel core plate surrounded by a
concrete- or grout-filled steel tube. Under compressive loads,
■ Thirty-two precast concrete braced-frame archetypes the concrete- or grout-filled tube prevents buckling of the
were designed, and nonlinear numerical models of steel core plate (also known as the yielding core) to provide
the structures were developed. an axial strength of the brace in compression that is similar
to the axial strength to the brace in tension. This character-
■ Nonlinear static pushover analyses and incremental istic creates stable and nearly symmetric hysteretic load-de-
dynamic time-history response analyses were per- formation behavior with large energy dissipation, allowing
formed, and the analysis results were used to evalu- the yield strength of the steel core to dictate the design of the
ate the seismic performance of the archetypes. brace rather than the critical buckling load of the brace.1–5
■ This paper provides a recommended seismic de- Extensive research on steel buckling-restrained braced
sign procedure and recommended seismic perfor- frames has demonstrated that properly designed and detailed
mance factors for precast concrete buckling-re- frames concentrate damage during a seismic event in the
strained braced frames and suggests topics for yielding region of the braces, while the beams and columns
future research. essentially behave elastically.3,6–9 These findings led to the
codification of steel buckling-restrained braced frames for
PCI Journal is published bimonthly by the Precast/Prestressed Concrete Institute, 8770 W. Bryn Mawr Ave., Suite 1150, Chicago, IL 60631. use in the United States beginning in the 2005 edition of
Copyright © 2021, Precast/Prestressed Concrete Institute. The Precast/Prestressed Concrete Institute is not responsible for statements made the American Society of Civil Engineers’ Minimum Design
Loads for Buildings and Other Structures (ASCE 7-05),10

with a larger response modification coefficient R of 8 com- models are subjected to pushover analyses and incremental
pared with other braced-frame systems (for example, a dynamic time-history response analyses as defined within
response modification factor R of 6 for special concentrically the methodology. The dynamic analyses include the use of a
braced frames). Consequently, buckling-restrained braced prescribed ground-motion record set. Finally, the FEMA P695
frames have become the lateral system of choice for many methodology outlines a systematic evaluation of the analysis
steel structures in seismic regions, where they are associated results based on the uncertainty and collapse performance of
with significant reductions in costs as well as stable ductile the system.
lateral load behavior of the frame.
FEMA P695 requires extensive material, component, con-
Despite the popularity of buckling-restrained braced frame nection, and system testing for characterizing the behavior of
systems in steel construction, they have rarely been used in the proposed system and for calibrating the analysis models.
concrete structures, in large part due to limited research and Because these extensive experimental data are not currently
lack of codification. A few studies from outside the United available for precast concrete buckling-restrained braced-
States have investigated the use of buckling-restrained braces frame structures, the study described in this paper is limited
in reinforced concrete frames;11–14 however, these studies to the relatively small amount of experimental information
have focused primarily on seismic retrofit applications rather available to date.
than new construction. To the best of the authors’ knowl-
edge, no United States–based research on precast concrete Overview of archetypes
buckling-restrained braced frames has been published, and
only one experimental study (Guerrero et al.15) on the seismic This section describes the 32 precast concrete buckling-
behavior of these structures has been published worldwide. restrained braced-frame archetypes that were designed for
Consequently, practical implementation of precast concrete evaluation based on the FEMA P69519 methodology. Although
buckling-restrained braced frames has been rare, with limited all archetypes were designed with the same uniformly dis-
applications in international projects and only one building tributed gravity loads and material properties, various seismic
application in the United States.16 design categories (SDCs), building plans, numbers of braced
frames, brace configurations, and numbers of stories were
In an effort to address this research gap, this paper numerical- considered to span the expected design space of the proposed
ly investigates the lateral load behavior and design of precast structural system. Once established, these archetypes were
concrete buckling-restrained braced frames for potential then organized into performance groups in accordance with
feasibility in seismic regions. To this end, 32 precast concrete FEMA P695. The precast concrete beam and column mem-
braced-frame archetypes were designed, and nonlinear numer- bers were designed using deformed steel reinforcement with
ical models of these structures were developed using the Open no prestressing, considering details that emulate monolithic
System for Earthquake Engineering Simulation (OpenSees)17 cast-in-place reinforced concrete structures. Jointed (also
structural analysis platform. The numerical model was vali- referred to as “nonemulative”) precast concrete buckling-re-
dated using the results presented in Guerrero et al. and also by strained braced-frame structures were not included in this
comparing the OpenSees analyses with the results obtained study, but these types of precast concrete systems should be
from a second structural analysis platform, DRAIN-2DX.18 investigated in the future.
After the model was deemed suitable based on this validation,
nonlinear static pushover analyses and incremental dynam- Archetype design space
ic time-history response analyses were performed on the
32 archetypes. Ultimately, the analysis results were used to Two SDCs were used for this study, SDC Dmax and SDC Dmin,
evaluate the seismic performance of the archetypes and the as described by the spectral acceleration values provided in
seismic performance factors used in their design. FEMA P69519 Tables 5-1A and 5-1B. While the structures
evaluated for SDC Dmax were expected to be more critical,
To develop useful results grounded in a rational basis, this SDC Dmin was also considered for a limited number of designs
study followed many of the procedures described in the 2009 to capture any unexpectedly critical scenarios. To minimize
Federal Emergency Management Agency report Quantifi- structural overstrength and produce lower-bound designs,
cation of Building Seismic Performance Factors (FEMA the archetypes designed for SDC Dmin included fewer braced
P695),19 which provides a methodology to formalize the deter- frames within their building plans.
mination of seismic performance factors (for example, the
response modification coefficient) for new proposed lateral Figure 1 shows the archetype space, which consisted of three
force-resisting systems. This methodology includes several different symmetric building footprints. All building plans
steps to identify the range of application for the proposed sys- had an area of about 30,000 ft2 (2800 m2). The first repre-
tem and accurately assess the seismic collapse risk. The first sented an office building with 15 ft (4.6 m) story heights, the
step is to develop and design a set of archetypes that span the second represented an industrial building with 25 ft (7.6 m)
range of expected applications, where an archetype is defined story heights, and the third represented an alternate industrial
as a prototypical representation of the system. Second, non- building layout with 15 ft story heights. The office building
linear models are developed for each archetype. Third, these plan also included three different braced-frame layouts, which

42 ft 21 ft 42 ft 7 bays at 42 ft each = 294 ft 7 bays at 35 ft each = 245 ft
50 ft 35 ft 50 ft
1: SDC Dmax office building N
with accidental torsion
42 ft 21 ft 42 ft
4: SDC Dmax industrial building

with 25 ft story heights
7 bays at 30 ft each = 210 ft

Building perimeter
5 bays at 30 ft each = 150 ft

Gridlines
2: SDC Dmax office building
Braced frame
without accidental torsion
42 ft 21 ft 42 ft
3: SDC Dmin office building 5: SDC Dmax industrial building

with 15 ft story heights
Figure 1. Building and braced-frame plan layouts. Note: SDC = seismic design category. 1 ft = 0.305 m.
considered different levels of accidental torsion effects and and axial loads transferred to the beams and distributes the brace
different numbers of braced frames. The first office layout yielding across multiple stories.21 In contrast, single-diagonal
was arranged with the braced frames placed toward the core braces result in high axial forces in the beams and chevron
of the building to introduce accidental torsion effects per braces generate high bending moments in the beams. Therefore,
the 2016 edition of Minimum Design Loads and Associated these brace configurations were evaluated to capture the most
Criteria for Buildings and Other Structures (ASCE 7-16).20 critical conditions in the FEMA P695 methodology. The brace
The second layout had the same number of braced frames in angle was also considered an important parameter in the design
each direction, but accidental torsion effects were eliminated space; specifically, the different frame span lengths and story
from design by placing the east-west braced frames along the heights resulted in archetypes with brace angles ranging from
perimeter of the building plan. The third layout was designed 35.5 to 45.0 degrees from horizontal.
for SDC Dmin using a significantly reduced number of braced
frames arranged to eliminate accidental torsion effects. The The range of archetypes used in the study included one-,
industrial building layouts were both designed for SDC Dmax, two-, three-, four-, six-, and nine-story frames, with building
with braced frames at the exterior to eliminate accidental tor-
sion effects. The building layouts without accidental torsion
effects were expected to result in more critical FEMA P695
evaluations because these layouts were designed for lower
seismic forces.
Three different buckling-restrained brace elevation configura-

tions were investigated in this study: single diagonal, alternating
single diagonal (also known as zigzag), and chevron. Figure 2
presents these configurations within two-story frame archetypes. Single diagonal Zigzag Chevron
Given the large variety of possible arrangements, brace config-
urations deemed to be unlikely in precast concrete structures
or less critical based on the FEMA P695 procedures were not
included. For example, multistory X-bracing tends to be less Figure 2. Brace configurations for two-story frame
archetypes.
critical because it minimizes the unbalanced vertical loading

heights ranging from 15 to 135 ft (4.6 to 41.1 m). Based on of 46 ksi (317 MPa), respectively, based on section 5.5 of
preliminary results, archetypes taller than nine stories (taller the American Institute of Steel Construction’s (AISC’s) third
than 135 ft) were less critical in the FEMA P695 methodol- edition of the Seismic Design Manual22 and common industry
ogy and were also deemed less likely to be implemented in practice. The design yield strength of deformed reinforcing
precast concrete practice. Therefore, no archetypes taller than steel fsy was 80 ksi (552 MPa) and the design compressive
nine stories were included. strength of concrete fc' was 6 ksi (41.4 MPa). Given the large
design axial tension forces in the beams and columns, the
In the remainder of this paper, each archetype is labeled with use of Grade 80 (552 MPa) rather than Grade 60 (414 MPa)
a four character identifier, where the first character is the lay- reinforcing bars was necessary to minimize the sizes of these
out number (see Fig. 1), the next two characters indicate the members while satisfying design requirements for maximum
brace configuration (see Fig. 2), and the last character is the reinforcement ratios (see the “Design of Archetypes” section
number of stories. For example, archetype 1SD3 is a 3 story in this paper).
frame with single-diagonal braces in building plan layout 1.
Performance groups
Gravity loads
Table 2 shows the archetype designs grouped into nine per-
All archetypes were designed using the average distributed formance groups for system evaluation per FEMA P695.19
dead loads D and live loads L listed in Table 1. The total The frame designs within each performance group shared
average roof and floor dead loads were taken as 160 lb/ft2 similar characteristics expected to influence the results of the
(7660 N/m2), including a precast concrete double-tee-beam seismic evaluation. For this study, the performance groups
flooring system with a 4 in. (100 mm) thick cast-in-place top- were determined based on brace configuration, seismic
ping. The roof and floor average live loads were taken as 20 and design category, and fundamental building period domain
100 lb/ft2 (960 and 4790 N/m2), respectively. (short or long). FEMA P695 typically requires at least three
archetypes for each performance group, though groups with
Design material properties fewer than three archetypes are allowed if having three or
more alternate designs within a performance group is not
For the design of all archetypes, the yield strength of the brace considered feasible.
steel core was assumed to have the typical range of 42 ± 4 ksi
(290 ± 28 MPa), corresponding to minimum yield strength Design of archetypes
fymin of 38 ksi (262 MPa), and maximum yield strength fymax
This section describes the procedures used to design the
archetype braced-frame structures used in the investigation.
Table 1. Assumed overall average gravity loads
The design method was based on the equivalent lateral
Dead loads force procedure from ASCE 7-1620 and followed the Amer-
ican Concrete Institute’s Building Code Requirements for
Average load per
Contribution Structural Concrete (ACI 318-19) and Commentary (ACI
roof/floor area, lb/ft2
318R-19)23 for the design of the precast concrete beams and
Double-tee flooring 50 columns. Several applicable design requirements and rec-
ommendations for steel buckling-restrained braced frames
Topping slab 45
were also adopted, particularly with respect to the design
Beams and columns 25 of the braces and the resulting design forces on the beams
and columns, referencing AISC’s Seismic Design Manu-
Spandrels/exterior cladding 15 al,22 Specification for Structural Steel Buildings (ANSI/
Partition loads 15 AISC 360-16),24 and Seismic Provisions for Structural Steel
Buildings (ANSI/AISC 341-16),25 as well as the Structural
Buckling-restrained braces 5 Engineers Association of California’s (SEAOC’s) Structural/
Miscellaneous 5 Seismic Design Manual.26
Total dead load 160 Based on preliminary designs, trial values of the required
Live loads
seismic performance factors were chosen as follows: response
modification coefficient R of 8, deflection amplification factor
Location
Average load per Cd of 8, and system overstrength factor Ω0 of 2.5. These values
roof/floor area, lb/ft2 were then verified in the final step of the FEMA P69519 eval-
uation. The selected response modification coefficient R and
Roof 20
system overstrength factor Ω0 values are the same as those for
Floor 100 steel buckling-restrained braced frames, but the deflection am-
plification factor Cd of 8 is greater than the value of 5 specified
Note: 1 lb/ft2 = 47.9 N/m2.
for steel buckling-restrained braced frames in ASCE 7-16.

Table 2. Performance group summary
Grouping criteria
Performance Design load level

Number of archetypes
group number Brace configuration Period domain
Seismic design
Gravity
category
Single diagonal
1 Dmax Short 3 (1, 2, and 3 stories)
(with torsion)
2 Short 5 (1, 2, and 3 stories)

Dmax
3 Single diagonal Long 4 (6 and 9 stories)
4 Dmin Short 3 (1, 2, and 3 stories)

Typical
Dmax
6 Chevron Long 3 (4, 6, and 9 stories)
7 Dmin Short 3 (1, 2, and 3 stories)

Zigzag Dmax
9 Long 3 (4, 6, and 9 stories)
Figure 3 presents a summary flowchart of the design proce- system was not conducted. Furthermore, because this study
dure; subsequent sections of this paper describe each com- evaluated the overall behavior of the braced-frame system,
ponent of the flowchart. The design procedure focuses on the the brace-to-frame connections are not addressed. To this end,
lateral force-resisting braced frames, not the entire building it is implicitly assumed that the brace-to-beam and brace-to-
structure. Consequently, detailed design of the gravity load column connections would be designed to remain essentially
linear-elastic under the maximum brace forces, following
capacity-based design procedures.
Equivalent lateral force procedure
The ASCE 7-1620 equivalent lateral force procedure was used

to determine the lateral forces for the design of the archetype
frames. Table 3 shows the short-period design spectral accel-
eration parameter SDS and 1-second design spectral acceler-
ation parameter SD1, taken for SDC Dmax and SDC Dmin, per
FEMA P695.19
FEMA P695 defines the fundamental period T used for design

and analysis as
T = CuTa
Table 3. Design spectral acceleration parameters
Seismic design
SDS SD1
category
Dmax 1.0 0.60
Dmin 0.50 0.20
Note: SD1 = design spectral response acceleration parameter at 1-second

period; SDS = design spectral response acceleration parameter at short
Figure 3. Archetype design flowchart. periods.

where α = angle of brace relative to horizontal
Cu = coefficient for upper limit on the calculated period nb = number of braces in the story being designed
from ASCE 7-16 Table 12.8-1
This calculation for the brace axial force NQE conservatively
Ta = approximate fundamental period from ASCE 7-16 assumed that only the braces would carry lateral forces, with
section 12.8.2.1 no contribution from beam and column moment frame action
(similar to Design Example 3 from the SEAOC Structural/
Assuming comparable vibration characteristics, the approx- Seismic Design Manual26 and section 5.5 of the AISC Seismic
imate fundamental period Ta was calculated using the coeffi- Design Manual22 for steel buckling-restrained braced frames).
cients specified for steel buckling-restrained braced frames in The brace axial force NQE values were then increased to
ASCE 7-16 Table 12.8-2. Based on the design spectrum and account for second-order effects using the approximation
this fundamental period, the total seismic base shear force provided in ANSI/AISC 360-1624 appendix 8. Finally, the
VELF was determined using ASCE 7-16 Eq. (12.8-1), with the factored brace design axial force Nu was calculated based on
seismic response coefficient based on ASCE 7-16 section ASCE 7-16 load combinations. Per AISC 341-1625 section
12.8.1.1 and the seismic weight taken as 1.0D (which was F4.3, the braces were assumed not to carry any gravity loads
assumed to be the same at each floor level, including the roof, to ensure that the beam and column members of the frame
as shown in Table 1) per ASCE 7-16 section 12.7.2. These were designed for the full gravity loads in the event of loss of
calculations used a seismic importance factor Ie of 1 with braces (for example, due to fire loading). As such, the factored
Risk Category I or II, assuming that the office and industrial brace design forces under load combinations 6 and 7 were
buildings included in the archetype space represented low risk calculated as
to human life in the event of failure (ASCE 7-16 Table 1.5-1).
This choice was made to result in more-critical archetypes for Nu = ρNQE
the FEMA P695 evaluation.
where
Next, the total seismic base shear force was distributed
between the buckling-restrained braced frames in each of the ρ = redundancy factor, taken as 1.0 based on ASCE
two primary directions of the building. Because the braced 7-16 section 12.3.4
frames in each direction were assumed to be the same, the lat-
eral stiffnesses of these frames were also identical; and thus, Next, the yielding core areas of the braces were calculated
the total seismic base shear was divided evenly between the using the area-based approach described in the AISC Seismic
frames in each direction. The base shear forces were increased Design Manual. With this approach, the required brace core
as necessary to account for accidental torsion effects based on area was determined based on the lowest expected steel yield
the procedures outlined in Paulay and Priestley,27 assuming strength fymin. Thus, including a capacity reduction factor φ of
the sum of the frame stiffnesses in one direction to be equiv- 0.9, the minimum required steel core area of each brace was
alent to the sum of the frame stiffnesses in the orthogonal calculated as
direction. As permitted by ASCE 7-16 section 12.8.4.2, the
building plans with braced frames on the perimeter (layouts 2 Asc,min = Nu/(φfymin)
through 5 in Fig. 1) resulted in designs without any accidental
torsion effects. Once distributed to each individual frame, the The resulting ranges of brace yielding (core) areas and yield-
base shear force was then distributed vertically over the height ing lengths over the height of each archetype design are listed
of the structure at each floor and roof level, per ASCE 7-16 in Table 4, where the required areas have been rounded up to
section 12.8.3. the next 0.10 in.2 (64.5 mm2) increment to achieve realistic
designs with minimal overstrength.
Brace design
After the yielding area of each brace was designed, the
The buckling-restrained braces were designed based on the adjusted brace forces were determined based on the highest
brace axial forces NQE from the ASCE 7-1620 equivalent lateral expected steel core yield strength fymax for use in the design of
force procedure and the expected yield strength of the yield- the beams and columns, following a capacity-based design
ing region of the braces. The brace forces in each story were approach. The adjusted brace forces were calculated accord-
calculated as ing to ANSI/AISC 341-16 section F4.2a as
NQE = Vstory/[nbcos(α)] BRBT = ωRy fymaxAsc
where BRBC = βωRy fymaxAsc
Vstory = shear force in story being designed where

Table 4. List of archetypes with corresponding range of brace yielding (core) areas and brace yielding lengths
Design configuration
Performance Archetype design iden- Range of brace Range of brace
group tification number Number of Seismic design yielding areas, in.2 yielding lengths, in.
stories category
1SD1 1 4.9 197
1 1SD2 2 Dmax 6.5 to 9.9 170 to 188
1SD3 3 6.2 to 12.2 164 to 198
2SD1 1 3.8 186
2SD2 2 5.1 to 7.6 186 to 192
2 2SD3 3 Dmax 4.8 to 9.4 182 to 193
4SD1 1 6.1 382
4SD2 2 6.3 to 9.2 336 to 395
1SD6 6 4.7 to 14.7 158 to 190
1SD9 9 4.0 to 16.6 129 to 198

3 Dmax
2SD6 6 3.6 to 11.3 176 to 201
2SD9 9 3.1 to 12.6 154 to 215
3SD1 1 2.9 197
4 3SD2 2 Dmin 2.7 to 4.0 195 to 197
3SD3 3 2.3 to 4.6 193 to 198
2CC1 1 2.9 202
2CC2 2 3.8 to 5.9 191 to 199
5 2CC3 3 Dmax 3.6 to 7.1 186 to 199
5CC1 1 3.3 311
5CC2 2 3.5 to 5.0 302 to 310
2CC4 4 3.2 to 3.7 184 to 201
6 2CC6 6 Dmax 2.7 to 8.5 178 to 202
2CC9 9 2.3 to 9.6 163 to 196
3CC1 1 1.5 208
7 3CC2 2 Dmin 1.4 to 2.0 206 to 208
3CC3 3 1.2 to 2.3 205 to 209
2ZZ2 2 5.1 to 7.6 186 to 192
8 2ZZ3 3 Dmax 4.8 to 9.4 182 to 193
4ZZ2 2 6.3 to 9.2 336 to 395
2ZZ4 4 4.3 to 10.1 180 to 194
9 2ZZ6 6 Dmax 3.6 to 11.3 178 to 196
2ZZ9 9 3.1 to 12.5 155 to 215
Note: CC = chevron brace configuration; SD = single-diagonal brace configuration; ZZ = zigzag brace configuration. 1 in. = 25.4 mm; 1 in.2 = 645 mm2.

BRBT = brace force in tension the adjusted brace forces and the beam design axial forces in
this study, as described in the following paragraphs.
β = adjustment factor for brace force in compression
For the single-diagonal brace configuration, the beam design
ω = strain-hardening adjustment factor axial forces were calculated using the adjusted brace forces
directly above and below the beam being designed. Assuming
Ry = expected yield-strength adjustment factor account- that the earthquake-induced shear force in the building could
ing for material variability be evenly transferred from both ends of the frame, the axial
force in each beam was calculated as the average horizontal
Asc = area of the steel core component of the two adjusted brace forces. In this calcula-
tion, the beam tensile axial force demand corresponded to the
BRBC = brace force in compression direction of lateral loading with the braces in compression,
whereas the beam compressive axial force demand corre-
For preliminary design, the compression force adjustment fac- sponded to the loading direction with the braces in tension.
tor β was assumed as 1.1 and the strain-hardening adjustment
factor ω was assumed as 1.4. Because material variability A similar procedure was followed for the beam design axial
was already accounted for by designing the brace areas based forces with braces in the zigzag configuration, using the ad-
on the minimum yield strength fymin, while using the maxi- justed brace forces directly above and below the beam. How-
mum yield strength fymax for the adjusted brace forces (used to ever, in this configuration, one brace will be in tension while
design the rest of the frame), the expected brace yield strength the other is in compression. Therefore, the beam axial force
adjustment factor Ry was equal to 1 in all of the adjusted brace demands were calculated conservatively as the difference
force calculations. between the horizontal components of the absolute adjusted
brace forces directly above and below the beam. The axial
The design flowchart in Fig. 3 shows that the adjusted brace forces were calculated considering lateral forces acting to the
forces were revised based on the brace deformations deter- left and the right, and the largest compressive and tensile forc-
mined from an effective linear-elastic analysis of the prelimi- es from either direction were used as the axial force demands
nary frame design, which is described later in this paper. Once on each beam.
the effective linear drift analysis of each preliminary arche-
type was completed, the adjusted brace forces and resulting For the chevron brace configuration, the beam design axial
frame designs were iterated using updated values for the force demands were calculated based on the two buckling-
compression force adjustment factor β and strain-hardening restrained braces below the beam in consideration. This con-
adjustment factor ω. figuration results in a large horizontal force at the connection
between the buckling-restrained braces and the midlength of
Beam design the beam because one of the braces will be in tension while the
other is in compression. This force is carried as tension in half
The precast concrete beams of each archetype were designed of the beam length and compression in the other half, though
based on the factored axial force Pu and bending moment the exact distribution of this force between the two halves of
Mu demands from ASCE 7-1620 load combinations and the the beam depends on the load path. In addition, each half length
adjusted brace forces in tension BRBT and compression BRBC. of a beam can experience tension as well as compression,
Unlike traditional beam design, the large compressive and depending on the direction of loading. In this study, the two
tensile axial forces of the buckling-restrained braces required halves of each beam were assumed, for simplicity, to have an
the beams to carry large axial forces from earthquake ef- even tributary area, thus evenly carrying the horizontal force
fects in addition to moments and shear forces from gravity from the braces. Therefore, each beam was designed for tensile
loads. Therefore, the beams were assumed to act like column and compressive axial forces equal to one-half of the sum of the
members, and the design of the beams followed the column horizontal components of the adjusted brace forces.
requirements for special moment frames in ACI 31823 chapter
18, rather than the equivalent requirements for beams. The factored design bending moment demands Mu for the
beams were determined from both gravity loads and earthquake
Because gravity loads do not produce axial forces in beams, effects. Although Table 1 lists the average distributed dead and
the factored design axial force Pu in each beam was calcu- live loads assumed for the entire structure, some of the dead
lated from load combinations 6 and 7 based solely on the loads were not carried by the beams. Therefore, all beams were
adjusted brace forces, following the requirements of ANSI/ designed for a smaller dead load of 130 lb/ft2 (6220 N/m2) to
AISC 341-1625 section F4.3. The exact relationship between exclude the weight of the buckling-restrained braces, columns,
the beam axial forces and the adjusted brace forces depends and exterior cladding. For beams at the exterior (that is, perime-
on the seismic load path, tributary mass, collectors on either ter) of the structure, an additional vertically distributed 35 lb/ft2
side of the frame, and the distribution of forces throughout the (1700 N/m2) dead load was included to account for the exterior
entire structure. For simplicity, however, several assumptions cladding weight. The live loads listed in Table 1, reduced per
for each brace configuration guided the relationship between ASCE 7-16 section 4.7, were used for the beam design.

The gravity moments were then calculated based on the a simply supported beam with the net upward point load Fy
factored dead and live loads over the tributary width of each acting at the midlength and used in load combinations 2, 6,
beam, and the orientation of the double-tee flooring system. and 7 to find the total factored design moments for each beam.
For the archetypes with single-diagonal and zigzag brace For frames with single-diagonal and zigzag brace configu-
configurations, the flooring system was assumed to run rations, the braces were assumed to be pin connected at the
perpendicular to the beam on both sides (that is, the floor and beam-to-column joints such that the brace forces resulted in
roof double tees were assumed to be framing into the beams), no significant bending moments on the beams.
thus transferring dead and live loads onto the beams. For the
chevron brace configuration, the double tees were assumed The beams were designed for the combined factored axial
to run parallel to the frame. Therefore, each beam was only force Pu and bending moment Mu demands for each load
designed for dead and live loads from the beam self-weight, combination. Because each beam was designed based on
weight of topping slab directly above the beam width, exterior the requirements for columns in special moment frames in
cladding on perimeter beams, and live load directly above the ACI 318 chapter 18, the longitudinal reinforcement ratio was
beam width. These different assumptions for the orientation kept between 1% and 6% (ACI 318 section 18.7.4.1). The
of the flooring system were made to evaluate effects of gravity large axial tensile forces and the maximum reinforcement
load variations on the design and performance of the beams. limit of 6% tended to generate excessively large member sizes
when using Grade 60 (414 MPa) reinforcement. Therefore,
Different boundary conditions were considered to determine Grade 80 (552 MPa) reinforcing steel was used consistently
the largest positive and negative beam bending moment de- instead. For simplicity, all beams were designed as rectangu-
mands. For gravity loads, the maximum negative moments at lar sections with a 4 in. (100 mm) thick cast-in-place topping
the beam ends were calculated assuming fixed end supports, slab placed to act compositely on top of the beam (Fig. 4).
while the maximum positive moment at the midlength was For configurations with the floor and roof double tees oriented
calculated assuming simply supported boundary conditions perpendicular to the braced frame, the beams were designed
(similar to section 5.5 of the AISC Seismic Design Manual22 as T beam sections with an effective topping slab flange width
for steel buckling-restrained braced frames). For braces in the per ACI 318 Table 6.3.2.1 and eight no. 6 (19M) reinforcing
chevron configuration, an additional negative beam moment bars assumed within this topping slab width. For configura-
due to gravity loads was calculated at the brace location (that tions with the floor and roof system running parallel to the
is, midlength of beam) assuming a simply supported two-span braced frame, the effective width of the topping slab was
continuous beam (similar to Design Example 3 in the SEAOC limited to the width of the beam, with only two no. 6 rein-
Structural/Seismic Design Manual26). forcing bars assumed within this slab width (Fig. 4). Because
a full design of the floor and roof system was not conducted,
For the chevron brace configuration, beam bending moments the number and size of the topping slab reinforcing bars were
also develop from earthquake effects because the brace selected based on typical industry designs.
forces directly below the beam, one in compression and
the other in tension, generate a net upward point load Fy of The beams were designed for each factored axial-moment
( )
BRBC − BRBT sin (α ) at the beam midlength (section 5.5 (Pu-Mu) load combination pair using interaction diagrams
in the AISC Seismic Design Manual and Design Example 3 generated in MATLAB. Fig. 5 shows a representative beam
from the SEAOC Structural/Seismic Design Manual). The interaction diagram. The interaction diagrams considered both
moments caused by this point load were calculated assuming positive and negative bending, as well as compressive and ten-
Effective topping slab width Effective topping slab width
No. 6 reinforcing bars (8 total) No. 6 reinforcing bars

topping
4 in.
Nonbuckling
reinforcing
bars, typical
Ties/hoops,
Buckling reinforcing typical
bars, typical
Beam (perpendicular floor system) Beam (parallel floor system) Column
Figure 4. Sample beam and column cross sections in braced-frame archetypes. Note: no. 6 = 19M; 1 in. = 25.4 mm.

Beam interaction diagram Column interaction diagram
3000 2000
Design axial strength, kip
Design axial strength, kip

2000 Positive bending
1000 Negative bending
Load case 2 demands
1000 Load case 6 demands
0 Load case 7 demands
0
-1000 -1000
-500 -250 0 250 500 750 0 100 200 300 400
Design moment strength, kip-ft Design moment strength, kip-ft
Figure 5. Sample axial-moment strength interaction diagrams for beam and column design. Note: 1 kip = 4.45 kN; 1 kip-ft = 1.356
kN-m.
sile axial forces, and the contribution of the assumed effective also considered by assuming that the columns at each end of
topping slab width and reinforcement to the axial-moment the beam carried 0.5Fy as axial tension.
strength was included. In generating the interaction diagrams
for design, the stress-strain behavior of the reinforcement AISC 341-1625 section F4.3 allows column bending mo-
was idealized as elastic, perfectly plastic. The design of each ments from seismic effects to be neglected, assuming that the
beam was considered to be satisfied if all applicable load portion of story shear resisted by these moments is generally
combination pairs fell within the interaction diagram with small (Kersting et al.21). As such, only the moment demands
minimal overstrength so as to result in critical archetypes for from gravity loads were used in column design. The proce-
the FEMA P69519 evaluation. dure to calculate these moment demands was based on the
SEAOC Structural/Seismic Design Manual,26 where the beam
Finally, the design of each beam was checked for shear end moments from gravity loads (assuming fixed-fixed beam
requirements. Although a full shear reinforcement design end boundary conditions) are distributed to the connecting
was not performed, the ACI 318 section 22.5.1.2 limits for columns. This distribution assumes points of zero moment at
the maximum allowable shear strength based on material the column base (above the foundation) and at the midheight
strengths and the dimensions of each member were checked. of each upper story (in other words, each story except the
The corresponding beam shear force demands were calculated first story) and constant shear force along the column height
based on ACI 318 Fig. 18.6.5 to ensure that the maximum al- between those points.
lowable shear strength was not exceeded. Per ACI 318 section
18.7.6.1.1, the shear demand was checked against the strength Similar to the beams, the columns were designed using axial-
over the range of the factored design axial forces. The shear moment strength interaction diagrams generated in MATLAB.
design requirements often governed the beam dimensions, Figure 5 shows a representative column interaction diagram.
resulting in beam widths greater than the corresponding beam Because each column was symmetric, these interaction
depths to satisfy shear demands without significantly increas- diagrams only considered positive bending moments. The col-
ing the beam moment strengths. umn longitudinal reinforcement percentages were kept within
the range of 1% to 6%, and Grade 80 reinforcement was used
Column design to minimize the column sizes. For simplicity, longitudinal re-
inforcing bars were only placed around the section perimeter,
The columns were designed for the combined factored axial and the column reinforcing bars over each story height were
force Pu and bending moment Mu demands from ASCE 7-1620 designed to be the same size. Per typical precast concrete
load combinations 2, 6, and 7. The axial force demands due to industry practices, the column dimensions were changed only
gravity loads were calculated by multiplying the factored dead every third story.
and live loads by the tributary area for each column. Earth-
quake effects caused both axial compressive and tensile force ACI 31823 section 18.7.3.2 enforces strong column–weak
demands in the columns, considering equivalent lateral forces beam behavior for special reinforced concrete frames by
in each direction of the frame. These demands were calculated requiring that
using the vertical components of the adjusted brace forces in
all of the braces above the column being designed. For the ΣMnc ≥ (6/5)ΣMnb
chevron brace configuration, the net upward force Fy due to
the unequal adjusted brace forces at the beam midlength was where

ΣMnc = sum of nominal moment strengths of the columns deformations from this step were also used to iterate the
framing into each joint adjusted brace forces (by updating the adjustment factor for
brace force in compression β and strain hardening adjust-
ΣMnb = sum of nominal moment strengths of the beams ment factor ω) and update the design of the beams and col-
framing into the same joint umns accordingly. To ensure accurate drift analysis results at
the equivalent lateral force level, several effective stiffness
This requirement was indirectly satisfied (that is, without parameters were used to represent each beam, column, and
specifically considering ACI 318 section 18.7.3.2) for most of brace member linear elastically. Each brace was modeled
the columns in each archetype structure; however, some of the as a single element connected at the frame work points,
column sections in the upper two or three stories of the taller assumed to be at the intersecting centroids of the beam and
archetypes did not satisfy this requirement. Because it was column members (Fig. 6), with an area equal to the yielding
deemed important to design critical structures with minimal area and a stiffness modification factor (greater than 1.0) to
overstrength for the FEMA P69519 evaluation, the column sec- account for the added stiffness from the much stiffer end re-
tion sizes and/or reinforcement amounts were not increased to gions of the brace. These brace stiffness modification factors
achieve ΣMnc ≥ (6/5)ΣMnb. were recalculated after iteration, as necessary, ranging from
stiffness increases of 35% to 90%, depending on the brace
Each column design was also checked to meet shear require- size and geometry. The ends of each brace were assumed to
ments. Similar to the beams, a full shear reinforcement design be pinned into each work point node, thus transferring only
was not performed for the columns; however, the ACI 318 axial forces along the brace axis.
maximum shear force demands were calculated to ensure
that the maximum allowable shear strength limits were not The beam and column members were modeled with the axial
exceeded. Unlike beam design, shear requirements never and flexural stiffness reduction factors shown in Table 5
governed the column dimensions. based on the gross area Ag and gross moment of inertia
Ig of each member. These factors are based on ACI 31823
Effective linear-elastic drift analysis Table 6.6.3.1.1(a) and Table 6.6.3.1.1(b), which provide
model area and moment of inertia reduction factors for effective
linear-elastic analysis at factored load levels. However,
After the preliminary design of all frame members was some modifications were necessary because the ACI 318
completed, an effective linear-elastic equivalent lateral force effective stiffness factors do not account for the increased
pushover analysis for each archetype was conducted to cracking and stiffness reduction expected to occur due to
check that allowable story drift limits per ASCE 7-1620 were the large axial tension forces in the beams and columns of
satisfied. As described in the flowchart in Fig. 3, the brace buckling-restrained braced frames. The recommended mod-
Bay width
Column 2 Threaded Column 2
region fiber
Duct element
Grout Grouted
Story height
Grout pad splice Beam fiber

connector element
A A
Beam
Rigid
zones
Work point
Story height
Connecting
reinforcing bars Column 1
Column 1 fiber element
Confined
Duct concrete
fibers Work point
Grout Beam and column outline
Unconfined
Connecting concrete Buckling-restrained brace
reinforcing bar fibers Beam/column gross
Steel centroid axes
reinforcement
Section A-A fibers
Assumed connection details Connection modeling details Frame modeling details
Figure 6. Illustration of beam-to-column connection region assumptions and frame modeling details.

Table 5. Effective area and moment of inertia By repeating this process until convergence, a different
reductions for beam and column members effective drift model was created for each ASCE 7-16 load
combination, and the largest story drift values were taken as
Axial force Area Moment of inertia
the governing values. Among load combinations 2, 6, and 7,
Tension 0.5Ag 0.25Ig load combination 7 typically governed the drift results.
⎛ Ast ⎞ ⎛ Mu P ⎞ * After all governing effective linear-elastic story drifts θe were

Compression 1.0Ag ⎜ 0.80 + 25 ⎟ ⎜1− − 0.5 u ⎟ Ig determined, the corresponding inelastic drifts θ were calculat-
⎝ Ag ⎠⎝
Pu
h Po⎠
ed using a proposed deflection amplification factor Cd of 8 for
Note: Ag = gross area of beam or column section, neglecting reinforce-
this system, as
ment; Ast = total area of longitudinal reinforcement in beam or column
section; h = depth of beam or column section; Ig = gross moment of
θ = Cdθe
inertia of beam or column section, neglecting reinforcement; Mu =
factored design moment of beam or column; Po = nominal compression
These inelastic drift θ values were then compared with the
(uniaxial) strength of beam or column at zero eccentricity; Pu = factored
requirements of ASCE 7-16 Table 12.12-1, which prescribes
design axial force of beam or column (positive for compression and
a maximum allowable story drift of 2.5% for structures with
negative for tension).
four stories or fewer and 2% for all other structures. If the
*See the American Concrete Institute’s Building Code Requirements for
ASCE 7-16 drift limits were exceeded, all column cross-section
Structural Concrete (ACI 318-19) and Commentary (ACI 318R-19) Table
dimensions over the entire structure height were scaled up until
6.6.3.1.1(b); not to exceed 0.875Ig or be taken less than 0.35Ig.
the drift requirements were met. No changes were made in the
beam sizes because preliminary analysis results showed that the
axial deformations of the columns typically controlled the drifts
ification factors were calibrated to match the deformations from the effective linear-elastic model. Using this approach,
of the effective linear-elastic drift model to a more detailed only the nine-story archetypes were drift controlled. All other
nonlinear inelastic model at the design equivalent lateral archetypes were controlled by strength design.
force level (see the next section of this paper, “Nonlinear
Numerical Modeling”). Finally, the effective linear-elastic drift model results were
used to iterate the brace designs. ANSI/AISC 341-1625 section
Because the beam and column effective moment of inertia F4 requires buckling-restrained braces to be designed for
reduction factors for compression varied depending on the deformations corresponding to a prescribed 2% story drift or
factored design axial force Pu and bending moment Mu, the twice the inelastic story drift (that is, 2θ), whichever is larger.
effective drift calculations were conducted iteratively. The The brace deformations were used to adjust the brace over-
ASCE 7-16 requirements for drift calculations also resulted in strength factors (adjustment factor for brace force in com-
the following additional steps in this iterative process: pression β and strain hardening adjustment factor ω) and the
stiffness modification factors used in design.
1. Determine the fundamental period of the preliminary
effective linear-elastic drift model from modal analysis. Common practice in steel buckling-restrained brace design
calculates equivalent brace deformations from story drifts
2. Use this fundamental period to update the equivalent using a simplified shear frame model, where the beam and
lateral forces. Because the period of the effective drift column members are assumed to be axially rigid. Even
model was usually longer than the capped period used in though these assumptions may work reasonably well for steel
the force-based member design, this step typically result- frames, where the member stiffnesses are similar in tension
ed in lower equivalent lateral forces for drift checks. and compression, significant axial elongations can occur in
concrete beam and column members due to reduced effective
3. Perform a pushover analysis on the effective drift model axial stiffnesses from cracking. Specifically, the beams and
under the updated equivalent lateral forces. This anal- columns of the archetype precast concrete buckling-
ysis included gravity loads based on the applicable restrained braced frames showed large axial elongations, thus
ASCE 7-16 load combinations, and P-Δ effects were contributing significantly to the effective story drifts. As such,
included per ASCE 7-16 section 12.8.7. each brace was designed for the larger of the following two
deformations to avoid excessive overestimation of the brace
4. Determine the beam and column axial forces and bending deformations:
moments from the effective drift analysis at the equiva-
lent lateral force level. • deformations in the brace elements of the effective
linear-elastic drift model at the equivalent lateral force
5. Use these axial loads Pu and moments Mu to revise the level (corresponding to effective linear-elastic story drift
moment of inertia reductions for members in compres- θe) multiplied by the deflection amplification factor Cd
sion per Table 5, and update the effective drift model
accordingly. • equivalent brace deformations at 2% story drift based

on shear frame assumptions (that is, assuming beam and two steel material models for the longitudinal (that is, axial-
column members are axially rigid) flexural) reinforcing bars and two material models for the
concrete. Figure 7 shows the stress-strain relationships for
In this approach, equivalent brace deformations calculated the reinforcing steel and typical unconfined concrete and
based on shear frame assumptions with twice the inelastic brace materials.
story drifts (that is, 2θ) from the effective drift model were
deemed unreasonable for precast concrete frames and, there- The two reinforcing steel material models included a buckling
fore, were not used to design the braces. After calculating reinforcing bar model and a nonbuckling reinforcing bar mod-
the brace deformations, the corresponding brace overstrength el. Both of these materials were modeled using the Steel4 uni-
and stiffness modification factors were updated as needed in axial material type in OpenSees,17 together with the MinMax
consultation with the brace manufacturer, and the iterative material limit. Steel4 is based on the Giuffré-Menegotto-Pinto
process was repeated. Once all brace factors converged and Steel02 material model, but it provides additional parameters
the ASCE 7-16 drift requirements were met, the design was to allow better simulation of the nonlinear cyclic steel stress-
considered complete. strain relationship. The MinMax material was applied to the
Steel4 parent material, allowing simulation of steel rupture
Nonlinear numerical modeling in tension and buckling in compression by reducing the steel
stress to zero after a user-defined strain limit was exceeded.
This section describes the detailed nonlinear numerical mod- Both buckling and nonbuckling reinforcing bar materials
eling and model validation of the precast concrete braced- were assumed to rupture at a tensile strain of 0.06, based on
frame system that was used in the FEMA P69519 evaluation. reversed-cyclic reinforcing bar test results by Aragon et al.28
The models were developed using the OpenSees17 structural In addition, the buckling reinforcing bar material was as-
analysis platform. sumed to buckle when the compression strain exceeded 0.004,
which was the assumed crushing strain for the surrounding
Element and material models unconfined cover concrete, indicating possible loss of lateral
support to the reinforcing bar from the concrete.
Figure 6 shows the assumed arrangement for the connections
between the beams and columns, with grouted vertical steel The following criteria were considered to determine which
reinforcing bars connecting the precast concrete columns reinforcing bars would be modeled as buckling and which
through the beams, leaving a precast concrete joint at the would be modeled as nonbuckling bars in each cross section.
interface of each column connecting to the beam. Therefore, ACI 31823 section 18.7.5 requires a transverse seismic hook or
each beam was modeled from work point node to work point the corner of hoop reinforcement supporting every longitudinal
node, whereas the columns were modeled with rigid end reinforcing bar in a cross section when Pu > 0.3Ag fc' .
zones within the beam-column joints. The work point nodes Because most of the beams and columns had large axial
were assumed to be at the intersecting centroids of the beam compression Pu demands, nearly all of the longitudinal rein-
and column members. forcing bars in each section were supported by seismic hooks
or the corner of hoop reinforcement (Fig. 4). Previous studies
To account for nonlinear axial-flexural behavior, each beam (Brown and Kunnath,29 Kunnath et al.,30 and Mander et al.31)
and column member was modeled as a fiber element using have shown that reinforcing bars are less likely to buckle if the
150 8 100
Cyclic
6 𝑓𝑓𝑓𝑓𝑐𝑐𝑐𝑐′ Backbone curve
75 50
4
Stress, ksi
Stress, ksi
Stress, ksi
0 2 Ec 0
0.2𝑓𝑓𝑓𝑓𝑐𝑐𝑐𝑐′
0
-75 7.5 𝑓𝑓𝑓𝑓𝑐𝑐𝑐𝑐′ -50
Cyclic
Cyclic -2 (𝑓𝑓𝑓𝑓𝑐𝑐𝑐𝑐′ in psi Compression
Monotonic units) Tension
-150 -4 -100
-0.01 0.01 0.03 0.05 0.07 -0.004 0 0.004 0.008 -0.02 -0.01 0 0.01 0.02
Strain, in./in. Strain, in./in. Strain, in./in.
Buckling reinforcement Unconfined concrete Buckling-restrained brace

with MinMax effect
Figure 7. Buckling reinforcement stress-strain curves with MinMax effect, unconfined concrete stress-strain curve, and buckling-
restrained brace under cyclic loading calibrated to backbone curve. Note: Ec = initial concrete stiffness; fc = specified
(design) compressive strength of concrete in psi units. 1 ksi = 6.89 MPa.

clear space between the transverse support hooks or hoops is place reinforced concrete structures due to different detailing
less than 6db, where db is the nominal diameter of the longitu- and response of precast concrete members and connections.41
dinal reinforcing bar. This requirement was met for every beam However, the plastic hinge lengths calculated based on the
and column based on the transverse reinforcement spacing equations in Table 6 were deemed appropriate for the purpos-
from shear design and longitudinal reinforcing bar sizes from es of this study because the archetype structures investigated
axial-flexural design. Thus, only the reinforcing bars in a beam were intended to emulate monolithic cast-in-place reinforced
or column with Pu ≤ 0.3Ag fc' and not located at the corner of a concrete structures. In addition, because the intent of the reg-
hoop or seismic hook were considered to buckle (Fig. 4). ularization of the concrete post-peak compressive stress-strain
relationships is to minimize the sensitivity of the analysis
The two concrete material models included unconfined and results to the selected plastic hinge length (that is, critical
confined concrete, represented using the Concrete01 and Con- integration length), models with concrete stress-strain rela-
crete02 material types in OpenSees. The unconfined concrete tionships regularized based on different assumed plastic hinge
material was used in the unconfined regions of each beam lengths are expected to result in similar performance.33–36
and column, which included the concrete cover, as well as the
beam tributary topping slab. This model assumed a maximum Both Concrete01 and Concrete02 material models in
concrete compressive strength fc' of 6 ksi (41 MPa), the same OpenSees use a uniaxial Kent-Scott-Park nonlinear com-
as the concrete strength used in design. The confined concrete pressive stress-strain relationship. Linear unloading and
material represented the confined regions of the members reloading stiffness, equal to the initial stiffness Ec, was
bounded by the centerlines of the transverse hoop reinforce- assumed for cyclic loading (Fig. 7). The maximum compres-
ment. The maximum compressive strength of the confined sive strength, strain at maximum compressive strength, re-
concrete was calculated using the method by Mander et al.32 sidual strength, and strain at residual strength are defined by
based on the spacing and arrangement of the transverse rein- the user, whereas the initial concrete stiffness Ec is implicitly
forcement from ACI 318 section 22.5 requirements. calculated by OpenSees based on the user-defined maximum
compressive strength and strain at maximum compressive
The post-peak compressive stress-strain relationships of the strength. Because these material models do not allow the
unconfined and confined concrete were determined follow- user to specify the concrete initial stiffness explicitly, the
ing the regularization process developed by Coleman and strain at maximum compressive strength was calculated to
Spacone,33 Pugh et al.,34 Vásquez et al.,35 and Pozo et al.36 to result in an initial concrete stiffness of Ec = 57,000 f c′ ,
minimize the sensitivity of the softening concrete models with fc' in psi units per ACI 318.
(that is, with reducing stress beyond the maximum strength
point) to the assumed critical integration lengths of the fiber The only difference between Concrete02 and Concrete01 is
beam and column elements. The residual post-peak concrete that Concrete01 has zero tensile strength, whereas Concrete02
strength was assumed to be 20% of the maximum compres- includes the tensile strength of the concrete. With Concrete02,
sive strength. The regularization of the concrete post-peak the user defines the tensile strength and tension softening
compressive stress-strain relationships was performed using stiffness, in addition to the parameters used in Concrete01.
a plastic hinge length Lp taken as the average plastic hinge The concrete in the beams was assumed to have tensile
length calculated from the five equations listed in Table strength (Concrete02) equal to 7.5 f c′ (with fc' in psi units
6,27,37–40 which was prescribed as the critical integration length per ACI 318), with the tensile stress conservatively assumed
of each element. Note that plastic hinge lengths for precast to drop immediately to zero after the initiation of cracking,
concrete members can differ from those of monolithic cast-in- thus ignoring any gradual tension softening. In comparison,
Table 6. Plastic hinge length equations
Equation Units of measure Reference
0.032z
Lp = 0.5d + Meters Corley (1966)
d
Lp = 0.5d + 0.05z Any consistent length unit Mattock (1967)
Lp = 0.08z + 6db Any consistent length unit Priestley and Park (1987)
Lp = 0.08z + 0.022dbfsy Millimeters, megapascals Paulay and Priestley (1992)
0.1fsy db
Lp = 0.05z + Millimeters, megapascals Berry et al. (2008)
fc′
Note: d = distance from extreme compression fiber to centroid of longitudinal reinforcement; db = nominal longitudinal reinforcing bar diameter;
fc = specified (design) compressive strength of concrete; fsy = specified (design) yield strength of reinforcing steel; Lp = plastic hinge length;
z = distance from critical section of beam or column to point of contraflexure (assumed as ½ element length). 1 mm = 0.0394 in.; 1 m = 3.281 ft;
1 MPa = 0.145 ksi.

the column concrete was assumed to have no tensile strength on tributary widths. The column gravity loads were modeled
(Concrete01), thus conservatively simulating the nonmono- as point loads at the work point nodes at each floor and roof
lithic (precast concrete) joints at the column ends (Fig. 6). level. These loads were calculated from the column tributary
areas and the loads in Table 1 but excluded the loads already
The buckling-restrained braces were modeled as nonlinear applied to the beams in the model.
truss elements pinned between work point nodes without
rigid end zones (Fig. 6) rather than with separate yielding and Because all of the gravity loads in a building (that is, not just
nonyielding regions of the brace. To properly account for the the loads tributary to the braced frames) contribute to second-
different regions of each brace with a single truss element, the order P-Δ effects, a leaning column was connected to each
element had an area equal to the yielding region of the brace frame model to capture the second-order effects of the gravity
(Table 4), and a stiffness modification factor was used to ac- loads not already applied on the frame. This leaning column
count for the increased stiffness from the relatively rigid end was pinned at the base and pinned at every three stories based
connection regions of each brace. The stiffness modification on typical precast concrete column lengths in practice, with a
factors ranged from 1.35 to 1.90, depending on the geometry pin-ended horizontal rigid link connection to the frame at each
and yielding area of the brace. The equivalent backbone axial floor and roof work point node (Fig. 8). The additional building
force–deformation curve for the work point–to–work point gravity loads were calculated as the total column axial loads in
truss element was developed from the backbone curve for the the gravity system of the building (that is, columns not part of
yielding region and the equivalent areas for the nonyielding the archetype braced frames). These loads were then summed
end regions of each brace. Cyclic characteristics (kinematic and applied as point loads to the leaning column at each floor
and isotropic hardening) were then selected such that the and roof level of the model. The leaning column was modeled
element behavior under cyclic loading would match the back- as an elastic element with an assumed area equal to the sum of
bone curves (Fig. 7). the cross-sectional areas and 70% of the sum of the moments
of inertia of the gravity system columns based on ACI 31823
Frame modeling Table 6.6.3.1.1(a). Because the gravity-resisting system of the
building was not explicitly designed, each gravity load column
Each archetype braced frame was simulated in OpenSees17 us- was assumed to be 24 × 24 in. (610 × 610 mm) based on typical
ing the aforementioned nonlinear elements and materials. The precast concrete practices.
models had fixed column bases and fixed beam-to-column
connections, but the column concrete was modeled with zero For dynamic modeling, seismic masses were assigned to all
tensile strength (Concrete01), as noted in the previous section, of the work point nodes on the frame. As required by FEMA
to simulate the precast concrete joints. Tributary dead and live P695,19 the total seismic mass for each floor and roof level of
gravity loads were applied to the beams and columns of each the entire building was calculated using the median gravity
archetype frame using the expected median gravity load com- load combination of 1.05D + 0.25L, with live load reductions
bination of 1.05D + 0.25L, as required by FEMA P695,19 with per ASCE 7-1620 section 4.7. The total building seismic mass
live load reductions per ASCE 7-1620 section 4.7. The gravity was distributed equally between the braced frames in each di-
loads on each beam were applied as distributed loads based rection, assuming rigid floor and roof diaphragms. Finally, the
Released end
Buckling-restrained
brace
Rigid link
Leaning column
Single diagonal and zigzag Chevron
Figure 8. Braced-frame model elevation depicting leaning column and superimposed frames.

frame seismic mass at each floor and roof level was distribut-
ed equally between the two work point nodes at that level.
Depending on the brace configuration, the archetypes were

modeled with one or two frames. The archetypes with chevron
braces were modeled using a single braced frame to represent
the building response. For the archetypes with zigzag and
single-diagonal brace configurations, the lateral load behavior
of a single frame was slightly different (that is, asymmet-
ric) for the two in-plane loading directions. This difference
stemmed from the asymmetric layout of the braces over the
height of the frames. However, the overall building behavior
was expected to remain essentially symmetric because an
even number of frames were placed within the building plan Work point node
with alternating brace orientations to avoid the asymmetry of Other node
a single frame. To simulate the expected symmetric behavior Beam and column outline
of the building as a whole, rather than the asymmetric behav- Buckling-restrained brace
ior of a single frame, each zigzag and single-diagonal braced Beam/column gross centroid axis
archetype was modeled using two superimposed frames Rigid zone
(Fig. 8). These frames were identical except for the orienta-
tion of the braces. The analyses were conducted by subjecting
the corresponding work point nodes (at the beam-to-column Figure 9. Modeling assumptions for precast concrete buck-
connections) of the two frames to the same lateral displace- ling-restrained braced frame tested by Guerrero et al. (2018).
ment history, assuming the presence of a rigid diaphragm at
each floor and roof level of the building. Accordingly, the material for the cast-in-place regions of each member. The
seismic masses and gravity loads for these models were also properties of the confined concrete regions were calculated us-
doubled for analysis. ing the procedures of Mander et al.32 based on the arrangement
of the reinforcement in the test specimens. Both beams and col-
Nonlinear model validation umns were modeled with rigid end zones and concrete tensile
strength (Concrete02) because the beam-to-column connections
The OpenSees17 nonlinear modeling approach described in the of the specimens in the study by Guerrero et al. did not have the
previous section was validated using available experimental precast concrete joints assumed for the archetypes herein. The
and numerical results from Guerrero et al.15 and comparable brace elements were placed at 20 degrees from horizontal, with
numerical models developed using a second analysis plat- rigid end zones to represent the connections of each brace to
form, DRAIN-2DX.18 The study by Guerrero et al. compared the beam elements of the frame (Fig. 9).
the dynamic properties and seismic responses of two four-
story precast concrete frame test specimens, with and without These models were used to conduct nonlinear static pushover
buckling-restrained braces. The beam and column members and dynamic analyses of the test specimens. Because the physi-
and connections of the specimens used a combination of cal specimens were not subjected to pushover tests, the numer-
precast concrete and cast-in-place concrete, designed with ical base shear versus story drift behaviors from the models in
practices used in Mexico. Braces were placed in a zigzag this study were compared with the numerical pushover curves
configuration over the four stories, with the ends of each presented in Guerrero et al., which were also developed using
brace laterally offset from the beam-to-column connections OpenSees (Fig. 10). For dynamic analyses, the numerical re-
such that the braces were attached solely to the beams. The sults were compared with the experimentally measured roof dis-
Ministry of Communications and Transportation building site placement time history of the braced-frame specimen subjected
east-west ground motion (SCT-EW) record from the 1985 to the SCT-EW record of the 1985 Michoacán earthquake.
Michoacán, Mexico, earthquake was used in the study. The
structures, constructed at one-third scale, were subjected to In addition, Fig. 11 compares the numerical peak dynam-
low-intensity white noise and the SCT-EW record of the 1985 ic response envelopes over the height of the braced-frame
Michoacán earthquake at different scaling factors (up to two specimen with the experimental results from Guerrero et al.
times the original record intensity) using a shake table. The These comparisons show that the analyses conducted fol-
study also included numerical analyses of the test specimens. lowing the numerical modeling approach for the archetypes
provided a good match to the measured peak story drift and
The OpenSees models of the test specimens from Guerrero et floor displacement data for the test specimen. The discrep-
al. were developed following an approach similar to that used ancies for the peak absolute velocities and accelerations
for the archetype frames. The material models for the concrete, were generally larger, which may be expected for nonlinear
reinforcing steel, and braces were based on the properties dynamic modeling because they are higher-order responses.
reported by Guerrero et al., including an additional Concrete02 However, because the peak drift and displacement responses

0.8 0.2
Base shear/seismic weight
Roof displacement, in.

0.6 0.1
0.4 0
OpenSees (this paper) with BRBs

0.2 OpenSees (this paper) without BRBs -0.1
OpenSees (Guerrero et al.) with BRBs OpenSees (this paper) with BRBs
OpenSees (Guerrero et al.) without BRBs Experimental (Guerrero et al.) with BRBs
0.0 -0.2
0.000 0.010 0.020 40 50 60 70 80
Maximum story drift ratio Time, sec
Figure 10. Comparisons of pushover curves (left) and dynamic roof displacement response (right, subjected to the Ministry of
Communications and Transportation building site east-west ground motion record at 100%) with results from Guerrero et al.
Note: BRB = buckling-restrained brace. 1 in. = 25.4 mm.
4 4 4 4
3 3 3 3
Floor or roof
Floor or roof
Floor or roof
Floor or roof
2 2 2 2 OpenSees (this
paper)
Experimental
1 1 1 1 (Guerrero et al.)
0 0 0 0
0 0.001 0.002 0 0.1 0.2 0 2 4 0 0.5 1
Story drift Displacement, in. Absolute velocity, ft/sec Absolute acceleration, g
Figure 11. Comparisons of absolute peak dynamic response envelopes of braced-frame test specimen (subjected to the Ministry
of Communications and Transportation building site east-west ground motion record at 100%) with results from Guerrero et al.
Note: g = gravitational acceleration. 1 in. = 25.4 mm; 1 ft = 0.305 m.
0.8 0.2
Base shear/seismic weight
Roof displacement, in.
0.6 0.1
0.4 0
OpenSees with BRBs
0.2 OpenSees without BRBs -0.1
DRAIN-2DX with BRBs OpenSees with BRBs
DRAIN-2DX without BRBs DRAIN-2DX with BRBs
0.0 -0.2
0.000 0.010 0.020 40 50 60 70 80
Maximum story drift ratio Time, sec
Figure 12. Comparisons of OpenSees and DRAIN-2DX model pushover curves (left) and dynamic roof displacement response
(right, (subjected to the Ministry of Communications and Transportation building site east-west ground motion record at 100%)
of the test specimens from Guerrero et al. Note: 1 in. = 25.4 mm.

are the most important dynamic results for the FEMA P69519 pushover curves for archetype 2SD9 in Fig. 13). Because the
evaluation, the validations of the modeling approach herein column dimensions of the drift-controlled archetypes were
demonstrated the suitability of the models for the purposes of increased to meet the ASCE 7-16 drift requirements (see the
the study described in this paper. “Design of Archetypes” section in this paper), the longitu-
dinal reinforcement ratios in these structures were relative-
For further validation, the OpenSees analyses of the test speci- ly low, around 2%, whereas most of the other archetypes
mens were compared with results obtained using nonlinear mod- had reinforcement ratios around 4% to 5%. Therefore, the
els in a second structural analysis program, DRAIN-2DX. All stiffness reduction factors in Table 5 may be better calibrated
elements in DRAIN-2DX were also modeled with fiber sections, to column reinforcement ratios of around 4% to 5%. This
and the concrete and steel materials were modeled to match the discrepancy, however, was deemed acceptable given that the
OpenSees models as closely as possible. Figure 12 shows that errors were not large and that only five of the archetypes were
the results from the two modeling programs were nearly identi- drift controlled during design. It may be possible to eliminate
cal for the nonlinear pushover analyses as well as the dynamic drift-controlled designs by using prestressed beam and col-
response history analyses of the specimens. These comparisons umn members to reduce cracking; however, this design option
provided further confidence for the use of the OpenSees models was not investigated.
in the FEMA P695 study described in this paper.
Archetype performance evaluation
Linear-elastic drift model validation and results
In addition to the FEMA P69519 evaluation of the archetypes, Each archetype braced frame was subjected to static push-
the OpenSees17 nonlinear braced-frame model was also used over and dynamic response time-history analyses using the
to validate the equivalent lateral force level deformations validated nonlinear OpenSees17 numerical model described
of the effective linear-elastic drift model used in design. To previously in this paper. These analyses and the evaluations
perform this validation, both the nonlinear model and the of the results followed the FEMA P69519 methodology. The
effective linear-elastic drift model for each archetype were pushover analysis results were used to determine the system
subjected to the same equivalent lateral forces used in design, overstrength factor Ω0 and period-based ductility μT, where-
and the resulting roof drifts were compared. Because the as the dynamic analysis results were evaluated with respect
gravity loads from ASCE 7-1620 load combination 7 governed to the response modification coefficient R and deflection
the drift design from the effective linear-elastic drift models, amplification factor Cd used in design. The results for the 32
the corresponding nonlinear models were also subjected to the archetypes are summarized in Table 7. The procedures used
same gravity loads from load combination 7. for analysis, including the dynamic response parameters and
collapse criteria used in the seismic evaluation, are described
Figure 13 shows that the stiffness reduction factors in Table 5 next, followed by discussions of the results in Table 7.
resulted in good or reasonable estimations of the nonlinear
roof drifts at the equivalent lateral force level for three select- Dynamic response parameters
ed archetypes. The effective stiffness models of the archetypes and collapse criteria
that were drift controlled during design (all of these cases
were for nine-story archetypes) tended to underestimate the The results of the FEMA P69519 procedure rely heavily on the
nonlinear model drifts at the equivalent lateral force level (see response parameters determined from the numerical analyses
750 750 750
Nonlinear
Base shear, kip
500
Base shear, kip
500 500
Base shear, kip
Effective linear-
elastic
250 250 250 Equivalent lateral
force level
0 0 0
0 0.25 0.5 0.75 1 0 0.25 0.5 0.75 1 0 0.25 0.5 0.75 1
Roof drift, percent Roof drift, percent Roof drift, percent
Figure 13. Comparison of effective linear-elastic drift model and nonlinear model monotonic pushover curves for archetypes
1SD3 (left), 2CC3 (center), and 2SD9 (right). Note: CC = chevron brace configuration; SD = single-diagonal brace configuration. 1
kip = 4.45 kN.

Table 7. Summary of pushover and dynamic analysis results
Overstrength and
Archetype design Design configuration Acceptance check
Performance collapse parameters
identification
Group Number Seismic design ACMR Pass/
number Ω0 μT ACMR
of stories category limit† fail
1SD1 1 2.94 7.35 1.89 1.56 Pass
1SD2 2 Dmax 2.29 7.55 2.55 1.56 Pass
1
1SD3 3 2.19 8.43 2.65 1.56 Pass
Performance group mean 2.47 7.78 2.36 1.96 Pass
2SD1 1 2.97 7.01 1.61 1.56 Pass
2SD2 2 2.30 8.28 1.78 1.56 Pass
2SD3 3 Dmax 2.09 8.79 1.73 1.56 Pass
2
4SD1 1 2.78 7.61 2.42 1.56 Pass
4SD2 2 2.24 9.57 2.32 1.56 Pass
1SD6 6 2.14 7.34 3.02 1.56 Pass
1SD9 9 2.13 7.11 2.15 1.56 Pass
Dmax
3 2SD6 6 2.04 7.64 2.13 1.56 Pass
2SD9 9 2.31 7.50 1.86 1.56 Pass
3SD1 1 2.79 7.65 2.31 1.56 Pass
3SD2 2 Dmin 2.34 8.23 2.67 1.56 Pass
4
3SD3 3 2.28 7.77 2.89 1.96 Pass
2CC1 1 2.04 8.49 1.77 1.56 Pass
2CC2 2 1.97 7.84 2.01 1.56 Pass
2CC3 3 Dmax 1.89 8.22 2.07 1.56 Pass
5*
5CC1 1 2.05 8.97 2.55 1.56 Pass
5CC2 2 2.00 9.11 2.32 1.56 Pass
2CC4 4 1.87 8.87 2.24 1.56 Pass
2CC6 6 Dmax 1.78 8.14 2.59 1.56 Pass
6
2CC9 9 1.76 7.34 2.68 1.56 Pass
3CC1 1 2.27 9.03 1.86 1.56 Pass
3CC2 2 Dmin 1.94 9.74 2.82 1.56 Pass
7
3CC3 3 1.76 10.72 2.95 1.56 Pass
2ZZ2 2 2.42 7.99 2.20 1.56 Pass
2ZZ3 3 Dmax 2.11 8.68 2.13 1.56 Pass
8*
4ZZ2 2 2.14 9.61 2.66 1.56 Pass
2ZZ4 4 2.02 8.46 2.07 1.56 Pass
2ZZ6 6 Dmax 1.96 7.88 2.14 1.56 Pass
9
2ZZ9 9 2.14 7.93 2.03 1.56 Pass
Note: ACMR = adjusted collapse margin ratio; CC = chevron brace configuration; SD = single-diagonal brace configuration; ZZ = zigzag brace configura-
tion; μT = period-based ductility; Ω0 = system overstrength factor.
*These performance groups were evaluated using a less-conservative definition of maximum brace ductility (see "Dynamic Response Parameters and
Collapse Criteria" section of paper).
† The ACMR limit listed for the individual archetypes is for 20% collapse probability, and the ACMR limit listed for the performance group mean is for
10% collapse probability. Per FEMA P695, the value of the response modification coefficient R used in design is deemed acceptable when both of the
following conditions are satisfied: the average ACMR for the archetypes in each performance group is greater than or equal to the 10% ACMR limit and
the ACMR value for each archetype is greater than or equal to the 20% ACMR limit.

and the collapse criteria used in the seismic evaluation. The limits based on measured brace performance from minimum
collapse criteria include failure types that are directly simu- ANSI/AISC 341-16 requirements. In this revised definition,
lated during the numerical analysis, as well as nonsimulated brace failure was based on the total ductility range of the brace
failure types determined during the postprocessing of the rather than the absolute ductility in either direction (that is,
analysis results. The failure types considered for the braced- compression or tension) of loading. This ductility range was
frame archetypes in this study were brace failure, gravity load calculated as the difference between the overall maximum and
system failure, shear failure of the precast concrete beams and minimum brace ductility over its entire loading history. For ex-
columns, and axial-flexure failure of the beams and columns. ample, consider a brace with a maximum ductility limit of ±17
established from ANSI/AISC 341-16 cyclic qualification testing.
Brace failure Brace failure was the most common collapse If this brace experienced a tension ductility demand of +25 and
criterion governing the FEMA P69519 evaluation results, a compression ductility demand of -5 during an earthquake, the
which may be expected because braces are the primary corresponding total maximum ductility demand range would
yielding lateral-force-resisting members of this system. In be calculated as 25 – (-5) = 30. In this example, the maximum
this criterion, the structure was deemed to have failed when a brace ductility demand of 25 exceeds the maximum ductility
brace exceeded the limit for its maximum ductility μmax or its limit of 17; however, the brace is deemed not to have reached
maximum cumulative inelastic ductility μc during the dynamic failure because the total range of ductility demand (30) remains
analysis. These brace ductility parameters were defined as under the maximum ductility range limit of 34, calculated as
twice the maximum ductility limit of 17 (that is, maximum duc-
μmax = Δmax/Δby tility range of 17 – [-17] = 34). The archetypes in performance
groups 5 and 8 were all evaluated based on this less-conservative
and collapse criterion (Table 7).
μc = ∑Δinelastic/Δby To ensure that the revised definition of maximum brace duc-

tility failure criterion was still conservative, the most critical
where braces from the archetypes were further checked for fatigue
failure. This was done by evaluating the strain history of a brace
Δmax = maximum brace deformation over the entire dynamic analysis (from the FEMA P695 study)
with a fatigue life model (between strain and number of cycles
Δinelastic = inelastic brace deformation to failure) developed through testing by the brace manufacturer
and the University of California San Diego, as documented in
Δby = brace deformation at yield Saxey et al.42 This analysis provided a damage index for each
brace, calculated based on standard rainflow counting of the
The maximum brace ductility limits used in this study ranged strain reversals of the brace over the duration of the ground
from 17 to 20 in both compression and tension based on the motion; a damage index at or above 100% would indicate brace
performance of braces that have been tested in accordance failure. The results showed a maximum damage index of only
with the cyclic qualification requirements in ANSI/AISC 16% for even the most critical cases from the archetypes in this
341-1625 section K3. The cumulative ductility limit was study. Thus, the braces were deemed very unlikely to fail under
conservatively set as 200, which is the minimum cumulative the FEMA P695 ground-motion record set before the collapse
ductility required for cyclic qualification testing, though the criteria outlined in this paper, and the revised definition of max-
performance of braces tested in accordance with ANSI/AISC imum brace ductility failure was deemed adequate.
341-16 would support much higher values.
Note that because maximum brace deformation limits are
The maximum cumulative inelastic ductility limit of the braces based on the strain capacities of the yielding region, braces
did not control the seismic performance of the archetypes. In with shorter yielding lengths tend to be less deformable and
comparison, the maximum ductility limit governed the primary could reach failure earlier than longer braces. Thus, frame
failure mode. Because the maximum ductility limit values used dimensions that result in shorter brace yielding lengths than
in the evaluation were based on the performance of braces tested those investigated in this paper (Table 4) may be more suscep-
in accordance with minimum ANSI/AISC 341-16 requirements, tible to collapse.
these limits also represented conservative (lower) values for
actual expected brace ductility capacities. In other words, the Gravity load system failure Gravity load system failure
actual ductility capacity of a brace under a ground motion record was deemed to occur at a prescribed story drift of 5% for any
was expected to exceed the performance of the brace from story. This requirement is consistent with other FEMA P695
ANSI/AISC 341-16 cyclic qualification testing because the test- studies (for example, Bruneau et al.43 and Tauberg et al.44),
ing procedures require extremely rigorous strain cycles that are where 5% drift was expected to correspond to extensive dam-
not likely to occur under typical earthquake events. Therefore, a age to a structure.
revised (less conservative) collapse criterion for brace ductility
was used for a few of the archetypes that did not initially pass Shear failure of beams and columns Shear failure was
the FEMA P695 requirements when using maximum ductility checked by comparing the maximum shear force demands

in the beam and column members against the corresponding ultimate roof drift was determined from the roof displacement
shear strengths calculated using ACI 31823 section 22.5.5. at the predicted failure of the frame, usually corresponding to
The analysis results showed that this failure criterion did not the maximum ductility limit of the braces (see the “Dynamic
govern the collapse of any of the archetypes. Response Parameters and Collapse Criteria” section in this
paper). Effective yield roof drift was calculated as the linear
Axial-flexural failure of beams and columns Exces- elastic roof drift of the structure at the maximum base shear
sive axial-flexural damage to the beams and columns of strength Vmax. Based on the results of Table 7, period-based
each frame was also considered as a collapse criterion. This ductility μT exceeded 6 for all of the archetypes, and the sys-
excessive damage was defined as the crushing of concrete in tem had an average ductility of 8.2.
any beam or column over a depth equal to 25% of the beam
or column depth, or buckling or fracture of the extreme layer System response modification
of longitudinal reinforcement in any beam or column. These coefficient R
criteria did not govern the collapse of the archetypes.
The response modification coefficient R of 8 used in the
System overstrength factor Ω0 design of the archetypes was evaluated based on the results of
and period-based ductility μT the nonlinear dynamic response history analyses. Per FEMA
P695,19 these analyses were conducted according to a mod-
The nonlinear static pushover analyses provided data on the ified incremental dynamic analysis procedure (Vamvatsikos
system overstrength factor Ω0 and period-based ductility μT and Cornell45), where each model was subjected to a set of
for each archetype frame. Per FEMA P695,19 the vertical ground motion records at increasing (scaled) intensities until
distribution of the lateral forces at the floor and roof levels the structure was deemed to reach collapse. FEMA P695 pre-
was determined based on the seismic mass and the ordinate of scribes a set of 22 far-field ground motion pairs (44 records
the fundamental mode for the frame model at each level. As total) curated from the Pacific Earthquake Engineering
an example, Fig. 14 shows the monotonic pushover analysis Research Center Next-Generation Attenuation database46 to
curve for archetype 1SD2. Although not required by FEMA assess collapse. These ground motion records were normal-
P695, the behavior of the structure under reversed-cyclic lat- ized by their respective peak ground velocities, and then the
eral loads is also shown. For each archetype, the overstrength entire record set was collectively multiplied with increasing
factor was calculated as the ratio of the maximum base shear scaling factors under the requirements of the FEMA P695
strength Vmax (from a monotonic analysis) to the design base methodology. Specifically, FEMA P695 characterizes ground
shear force VELF. Based on the results in Table 7, the over- motion intensity based on the spectral acceleration at the fun-
strength factor for the precast concrete buckling-restrained damental period of the archetype, with the important distinc-
brace system Ω0 was calculated as 2.5, taken as the largest tion that this intensity is defined collectively by the median
average overstrength factor out of all the performance groups, spectral acceleration of the entire record set (Fig. 15), rather
rounded up to the next half-unit interval. than by the different spectral accelerations of the individual
records in the set.
Period-based ductility μT was calculated as the ratio of the
ultimate roof drift δu to the effective yield roof drift δy,eff, The fundamental period T used to determine the median in-
where roof drift was calculated as the lateral displacement tensity is based on the maximum allowed ASCE 7-1620 funda-
at the roof level divided by the height of the archetype. The mental period CuTa for each archetype, not the period comput-
1500 1500
Vmax = 1220 kip 1000
Base shear, kip
Base shear, kip
1000 500
0
VELF = 532 kip
500 -500
Tangent line
Pushover curve
Design base shear -1000
Maximum base shear
0 -1500
0.0 2.0 4.0 -4.0 -2.0 0.0 2.0 4.0
δy,eff = 0.38% Roof drift, percent δu = 2.88% Roof drift, percent
Figure 14. Monotonic pushover curve (left) showing parameters for calculation of overstrength and ductility and cyclic pushover
curve (right) for archetype 1SD2. Note: SD = single-diagonal brace configuration; VELF = design base shear force; Vmax = maximum
base shear strength; δu = ultimate roof drift; δy,eff = effective yield roof drift. 1 kip = 4.45 kN.

10 3.0
Median record set intensity, g

ŜCT =
Spectral acceleration, g
2.15g
1 2.0
0.1 1.0
Individual records
Median spectrum
0.01 0.0
0.01 0.1 1 10 0.0 1.0 2.0 3.0 4.0
Period, sec Maximum story drift, percent
Figure 15. Acceleration response spectra of the normalized Federal Emergency Management Agency’s Quantification of Building
Seismic Performance Factors, FEMA P695, record set (left) and sample incremental dynamic analysis results for archetype 1SD1
(right). Note: g = gravitational acceleration; SD = single-diagonal brace configuration. ŜCT = median collapse intensity.
ed from a numerical model of the structure using eigenvalue period T for the archetype, is intended to account for ground
analysis. According to FEMA P695, normalizing and then motions that differ from the ASCE 7-16 design spectrum.
collectively scaling the records leads to a range of spectral ac-
celerations across the record set at the fundamental period of To evaluate the response modification coefficient R of 8 used in
the structure in a manner that maintains overall record-to-re- design, the collapse performance of each archetype was assessed
cord variability while eliminating inherent differences from based on the ACMR value from the incremental dynamic
factors such as event magnitude and distance to source. analysis procedure. For this evaluation, FEMA P695 Table 7-3
provides minimum ACMR limits corresponding to 10% and
Per FEMA P695, the lowest intensity (that is, lowest medi- 20% collapse probabilities based on the total system collapse
an spectral acceleration of the collective record set) that is uncertainty βTOT calculated using the following equation.
deemed to cause collapse of an archetype under half (22) of
2 2 2 2 2
the records was taken as the median collapse intensity ŜCT. βTOT = βRTR + βDR + βTD + βMDL
This approach can be represented in an incremental dynamic
analysis response plot of median spectral acceleration and where
maximum story drift, as shown for archetype 1SD1 in Fig. 15.
In this figure, each gray line represents the maximum story βRTR = record-to-record variability = 0.20 ≤ 0.1 + 0.1μT ≤
drift response of the archetype model under one ground mo- 0.40
tion record at increasing intensities, where each point rep-
resents the dynamic analysis result from one record at one in- βDR = uncertainty in design requirements
tensity. Failure of the archetype (based on the aforementioned
collapse criteria) during a ground motion event is usually βTD = test data uncertainty
manifested in the graph as the flattening of the response curve
under increasing intensities. The ŜCT value corresponds to the βMDL = modeling uncertainty
median intensity at which the structure is deemed to have col-
lapsed under half of the records. After ŜCT is determined, the Values for the total system collapse uncertainty βTOT range from
corresponding collapse margin ratio (CMR) is calculated as 0 to 1. The expected uncertainty values in design requirements,
test data, and modeling are all based on a qualitative assessment
CMR = ŜCT/SMT of each uncertainty and FEMA P695 Tables 3-1, 3-2, and 5-3.
In this study, all three uncertainties were rated as “good,” corre-
where sponding to βDR = βTD = βMDL = 0.20. Based on these results, the
total system collapse uncertainty βTOT is 0.53, and the corre-
SMT = maximum considered earthquake intensity from the sponding minimum ACMR limits were calculated as 1.56 and
response spectrum at the fundamental period of the 1.96 for 20% and 10% collapse probabilities, respectively.
structure T per ASCE 7-1620 section 11.4.4
Per FEMA P695, the value for the response modification
The adjusted collapse margin ratio (ACMR) is then calculated coefficient R used in design is deemed acceptable when both
by multiplying the CMR with a spectral shape factor (SSF) of the following conditions are satisfied:
as defined by FEMA P695 Tables 7-1a or 7-1b. The SSF, de-
termined based on period-based ductility μT and fundamental • The average ACMR for the archetypes in each perfor-

mance group is greater than or equal to the 10% ACMR earthquake spectrum level at the fundamental period of
limit. the archetype CuTa, where the design earthquake intensity
is determined from FEMA P695 based on the seismic
• The ACMR value for each archetype is greater than or design category (SDC).
equal to the 20% ACMR limit.
2. Conduct nonlinear dynamic history analyses of the struc-
Table 7 shows that the calculated individual and average ture under each scaled record.
ACMR values for all of the archetypes were above the ACMR
limits based on 10% and 20% collapse probabilities. The 3. Find the maximum roof drift demand from each record.
response modification coefficient R value of 8 used in design
was therefore deemed acceptable for the precast concrete 4. Take the median value across all of the maximum roof
buckling-restrained braced-frame system. drift demands as the maximum dynamic roof drift de-
mand at the design earthquake intensity δDBE.
System deflection amplification factor Cd
The roof drift of the effective linear-elastic drift model
According to FEMA P695,19 the system deflection amplifica- under equivalent lateral forces δELF was determined for each
tion factor Cd is calculated as archetype with the gravity loads changed to the FEMA P695
median load combination of 1.05D + 0.25L. The purpose
Cd = R/BI of using the effective linear-elastic drift model rather than
the nonlinear model was for the resulting roof drift δELF
where to be consistent with the value used in the design of each
archetype. The results of these calculations are outlined in
BI = numerical coefficient for effective damping βI from Table 8, showing that the deflection amplification factor Cd
Table 18.7-1 of ASCE 7-1620 across all archetypes ranged from 3.9 to 9.8. The mean Cd
values for the different performance groups ranged from 4.9
βI = effective damping due to inherent dissipation of en- to 7.5, and thus, the results of this alternate method support-
ergy by the structure, at or just below the effective ed the recommended system level deflection amplification
yield displacement of the seismic-force-resisting factor Cd value of 8.
system (ASCE 7-16 section 18.7.3.2)
Note that although a deflection amplification factor Cd of 8
FEMA P695 assumes that the structure will have inherent is recommended to encompass a wide range of applications
damping at 5% of critical. Therefore, the numerical coefficient for the proposed precast concrete braced-frame system,
for effective damping BI is 1.0 for the system, and deflection archetypes with a higher number of stories typically resulted
amplification factor Cd = R = 8. in lower values of the deflection amplification factor Cd. For
example, all archetypes with six stories or more had deflec-
An explicit calculation of the deflection amplification factor Cd tion amplification factor Cd values close to 5 (Table 8). Thus,
for each archetype was also performed to justify the value of future work may consider separating the precast concrete
8. This alternate calculation of deflection amplification factor buckling-restrained braced-frame system based on the number
Cd was based on a method performed by other FEMA P695 of stories or height to allow for the use of a lower deflection
studies, including Bruneau et al.43 and Tauberg et al.44 In this amplification factor Cd value for taller buildings. This would
method, the deflection amplification factor Cd is calculated as be beneficial for design because taller buildings are more like-
ly to be controlled by drift, and thus, more efficient designs
Cd = δDBE/δELF can be achieved using a lower deflection amplification factor
Cd value for these structures.
where
Conclusion
δDBE = maximum dynamic roof drift demand at design
earthquake intensity This study numerically evaluated the seismic design of pre-
cast concrete buckling-restrained braced frames as a novel
δELF = roof drift of effective linear-elastic drift model primary lateral-force-resisting structure using the FEMA
under equivalent lateral forces P69519 methodology. A set of 32 braced-frame archetypes
were designed to represent the range of key parameters
The value of the maximum dynamic roof drift demand at the expected to govern the seismic performance of the system,
design earthquake intensity δDBE for each archetype was deter- including the brace configuration (chevron, single diagonal,
mined using the following steps: and zigzag), building height, and building layout. Nonlinear
static pushover analyses of the structures were conducted
1. Scale the median spectral acceleration of the FEMA to determine the system overstrength factor Ω0. In addition,
P695 collective ground-motion record set to the design incremental dynamic analyses were conducted under a set of

Table 8. Calculated deflection amplification factors
Archetype design Design configuration

Performance group Cd
identification number Number of stories Seismic design category
1SD1 1 8.7
1SD2 2 Dmax 7.3
1
1SD3 3 6.4
Performance group mean 7.5
2SD1 1 8.6
2SD2 2 6.7
2SD3 3 Dmax 6.0
2
4SD1 1 6.5
4SD2 2 6.8
1SD6 6 4.8
1SD9 9 5.1
Dmax
3 2SD6 6 4.8
2SD9 9 5.0
3SD1 1 9.2
3SD2 2 Dmin 5.9
4
3SD3 3 3.9
2CC1 1 8.6
2CC2 2 6.8
2CC3 3 Dmax 6.5
5
5CC1 1 7.0
5CC2 2 7.6
2CC4 4 6.6
2CC6 6 Dmax 5.4
6
2CC9 9 4.3
3CC1 1 9.8
3CC2 2 Dmin 6.1
7
3CC3 3 4.4
2ZZ2 2 7.0
2ZZ3 3 Dmax 6.2
8
4ZZ2 2 7.1
2ZZ4 4 6.1
2ZZ6 6 Dmax 5.0
9
2ZZ9 9 5.1
Note: Cd = deflection amplification factor; CC = chevron brace configuration; SD = single-diagonal brace configuration; ZZ = zigzag brace configuration.

44 ground motion records to evaluate the response modifi- Note that these conclusions may be limited to the archetypes
cation coefficient R and deflection amplification factor Cd designed, the assumptions made in their numerical modeling
used in design. The main conclusions from the study are as and collapse criteria, and the ground motion records used in
follows: the dynamic analyses. The precast concrete buckling-restrained
braced-frame system currently lacks United States–based exper-
• Buckling-restrained braces exert large axial tension imental evaluation, so future work is needed to provide sys-
forces on the beams and columns, which result in large tem-level test data for this novel structure. In addition, this study
reinforcement requirements for these members. The use did not include the design of the brace connections to the beam
of higher-grade reinforcing bars can minimize the sizes and column members. Future experimental testing is needed
of the beam and column members while meeting design to ensure that these connections can be properly designed and
requirements for steel reinforcement ratios. detailed to reach the system performance levels shown in this
study. Future work should also explore the use of buckling-re-
• The large axial forces (both tension and compression) strained braces within prestressed and jointed (that is, nonemu-
that develop in the beams of buckling-restrained precast lative) precast concrete structures. Finally, based on the trends
concrete frames require these members to be designed and shown in the rigorous calculation of the deflection amplification
detailed like columns of special reinforced concrete frames. factor Cd, future work may consider separating the system with
respect to height, to allow for the use of a more favorable (low-
• Current practice for the design of buckling-restrained er) deflection amplification factor Cd value for taller buildings
braces in steel frames is typically based on brace defor- (for example, using a deflection amplification factor Cd of 5 for
mations calculated using a simplified shear building mod- structures with 6 or more stories).
el where the beam and column members of the frame are
assumed to be axially rigid. This assumption can lead to Acknowledgments
unrealistically large brace elongations in precast concrete
braced frames because it ignores the frame deformations This research was conducted with the sponsorship of PCI under
due to cracked beams and columns. As such, buckling-re- a Daniel P. Jenny Research Fellowship. The authors acknowl-
strained braces in precast concrete frames should be edge the support of the PCI Research and Development Council,
designed using the brace deformations calculated directly the PCI Central Region, and the members of the Industry Advi-
from an effective (cracked) linear-elastic drift analysis sory Committee, including Keith Bauer of Buehler Engineering
model. This paper provides recommended effective stiff- Inc., Roger Becker of PCI (retired), Kal Benuska of John A.
ness reduction factors for beams and columns that can be Martin & Associates Inc., Jared Brewe of PCI, S. K. Ghosh
used for design. of S. K. Ghosh Associates, Harry Gleich of Metromont Corp.,
Kevin Kirkley of Tindall Corp., Laura Redmond and Brandon
• The nonlinear numerical modeling approach described Ross of Clemson University, Kim Seeber of Seaboard Services
in this paper provided good comparisons with available of Virginia Inc., and Jeff Viano of Simpson Gumpertz & Heger
experimental shake-table test results of a precast concrete Inc. Any opinions, findings, conclusions, and recommenda-
buckling-restrained braced frame, especially in terms of tions expressed in the paper are those of the authors and do not
the peak displacement response parameters, which are necessarily reflect the views of the individuals and organizations
most important for the FEMA P695 evaluation. These acknowledged herein.
comparisons provided the validation of the models used
in this paper. References
• The numerical analyses showed that typical failure of 1. Watanabe, A., Y. Hitomi, E. Saeki, A. Wada, and M.
the archetype frames was reached due to failure of the Fujimoto. 1988. “Properties of Brace Encased in Buck-
buckling-restrained braces by exceeding maximum brace ling-Restraining Concrete and Steel Tube.” In Proceed-
ductility limits. In comparison, the maximum cumulative ings of the 9th World Conference on Earthquake Engi-
inelastic ductility limit of the braces or the failure of the neering, vol. 4, 719–724. https://www.iitk.ac.in/nicee
beams and columns did not govern the seismic perfor- /wcee/article/9_vol4_719.pdf.
mance of the archetypes.
2. Tremblay, R., G. Degrange, and J. Blouin. 1999. “Seismic
• Based on the nonlinear pushover and dynamic analyses Rehabilitation of a Four-Storey Building with a Stiffened
of the archetype frames in this study, the proposed system Bracing System.” In Proceedings of the 8th Canadian Con-
overstrength factor Ω0 is 2.5, response modification ference on Earthquake Engineering, 549–554. Vancouver,
coefficient R is 8, and deflection amplification factor Cd is BC: Canadian Association for Earthquake Engineering.
8 for precast concrete buckling-restrained frames. These
proposed seismic performance factors were shown to 3. Merritt, S., C. M. Uang, and G. Benzoni. 2003. Subassem-
satisfy the applicable FEMA P695 collapse probability blage Testing of CoreBrace Buckling-Restrained Braces.
limits for a variety of building plans, building heights, Structural Systems Research Project report TR-2003/04.
and brace configurations. La Jolla, CA: University of California San Diego.

4. Black, C. J., N. Makris, and I. D. Aiken. 2004. “Compo- 14. Oviedo A., J. A., M. Midorikawa, and T. Asari. 2010.
nent Testing, Seismic Evaluation and Characterization “Earthquake Response of Ten-Story Story-Drift-Con-
of Buckling-Restrained Braces.” Journal of Structural trolled Reinforced Concrete Frames with Hysteretic
Engineering 130 (6): 880–894. https://doi.org/10.1061 Dampers.” Engineering Structures 32: 1735–1746.
/(ASCE)0733-9445(2004)130:6(880). https://doi.org/10.1016/j.engstruct.2010.02.025.
5. Andrews, B. M., L. A. Fahnestock, and J. Song. 2009. 15. Guerrero, H, T. Ji, J. A. Escobar, and A. Teran-Gilmore.
“Ductility Capacity Models for Buckling-Restrained 2018. “Effects of Buckling-Restrained Braces on Rein-
Braces.” Journal of Constructional Steel Research 65: forced Concrete Precast Models Subjected to Shaking
1712–1720. https://doi.org/10.1016/j.jcsr.2009.02.007. Table Excitation.” Engineering Structures 163: 294–310.
https://doi.org/10.1016/j.engstruct.2018.02.055.
6. Clark, P., I. D. Aiken, E. Ko, K. Kasai, and I. Kimura.
1999. “Design Procedures for Buildings Incorporating 16. Viano, J. D., and T. C. Schaeffer. 2017. “Novel Use of Buck-
Hysteretic Damping Devices.” In Proceedings of the 68th ling-Restrained Braces in Precast Concrete Frames.” PCI
Annual Convention of the Structural Engineers Associ- Journal 62 (5): 28–34. https://doi.org/10.15554/pcij62.5-03.
ation of California, 355–371. Sacramento, CA: SEAOC
(Structural Engineers Association of California). 17. Mazzoni, S., F. McKenna, M. H. Scott, and G. L.
Fenves. 2006. OpenSees Command Language Manual.
7. Clark, P., K. Kasai, I.D. Aiken, and I. Kimura. 2000. Berkeley, CA: Pacific Earthquake Engineering Research
“Evaluation of Design Methodologies for Structures Center, University of California, Berkeley. https://
Incorporating Steel Unbonded Braces for Energy Dissi- opensees.berkeley.edu/OpenSees/manuals/usermanual/
pation.” In Proceedings of the 12th World Conference on OpenSeesCommandLanguageManualJune2006.pdf.
Earthquake Engineering. Upper Hutt, New Zealand: New
Zealand Society for Earthquake Engineering. https:// 18. Prakash, V., G. Powell, and S. D. Campbell. 1993. DRAIN-
www.iitk.ac.in/nicee/wcee/article/2240.pdf. 2DX Base Program Description and User Guide; Version
1.10. Report UCB/SEMM-93/17. Berkeley, CA: Structural
8. Sabelli, R., S. Mahin, and C. Chang. 2003. “Seismic Engineering Mechanics and Materials, Department of Civil
Demands on Steel Braced-Frame Buildings with Buck- Engineering, University of California, Berkeley. https://
ling-Restrained Braces.” Engineering Structures 25: nisee.berkeley.edu/elibrary/getpkg?id=DRAIN2DX.
655–666.
19. FEMA (Federal Emergency Management Agency). 2009.
9. AISC (American Institute of Steel Construction) and Quantification of Building Seismic Performance Factors.
SEAOC. 2001. Recommended Provisions for Buck- FEMA P695. Washington, DC: FEMA.
ling-Restrained Braced Frames. Chicago, IL: AISC.
20. ASCE and SEI. 2016. Minimum Design Loads and
10. ASCE (American Society of Civil Engineers) and SEI Associated Criteria for Buildings and Other Structures.
(Structural Engineering Institute). 2005. Minimum Design ASCE/SEI 7-16. Reston, VA: ASCE. https://doi.org/10
Loads for Buildings and Other Structures. ASCE/SEI .1061/9780784414248.
7-05. Reston, VA: ASCE.
21. Kersting, R. A., L. A. Fahnestock, and W. A. Lopez.
11. Ishii, T., T. Mukai, H. Kitamura, T. Shimizu, K. Fujisawa, 2015. Seismic Design of Steel Buckling-Restrained
and Y. Ishida. 2004. “Seismic Retrofit for Existing R/C Braced Frames: A Guide for Practicing Engineers. NEH-
Building Using Energy Dissipative Braces.” In Proceed- RP (National Earthquake Hazards Reduction Program)
ings of the 13th World Conference on Earthquake Engi- seismic design technical brief 11. Gaithersburg, MD:
neering, paper 1209. https://www.iitk.ac.in/nicee/wcee National Institute of Standards and Technology. https://
/article/13_1209.pdf. doi.org/10.6028/NIST.GCR.15-917-34.
12. Mazzolani, F. M., G. Della Corte, and M. D’Aniello. 22. AISC. 2018. Seismic Design Manual. 3rd ed. Chicago,
2009. “Experimental Analysis of Steel Dissipative Brac- IL: AISC.
ing Systems for Seismic Upgrading.” Journal of Civil
Engineering and Management 15 (1): 7–19. https://doi 23. ACI (American Concrete Institute) Committee 318. 2019.
.org/10.3846/1392-3730.2009.15.7-19. Building Code Requirements for Structural Concrete
(ACI 318-19) and Commentary (ACI 318R-19). Farming-
13. Sarno, L. D., and G. Manfredi. 2010. “Seismic Retrofit- ton Hills, MI: ACI.
ting with Buckling Restrained Braces: Application to an
Existing Non-ductile RC Framed Building.” Soil Dynam- 24. ANSI (American National Standards Institute) and AISC.
ics and Earthquake Engineering 30 (11): 1279–1297. 2016. Specification for Structural Steel Buildings. ANSI/
https://doi.org/10.1016/j.soildyn.2010.06.001. AISC 360-16. Chicago, IL: AISC.

25. ANSI and AISC. 2016. Seismic Provisions for Structural 37. Corley, W. G. 1966. “Rotational Capacity of Reinforced
Steel Buildings. ANSI/AISC 341-16. Chicago, IL: AISC. Concrete Beams.” Journal of the Structural Division 92
(5): 121–146. https://doi.org/10.1061/JSDEAG.0001504.
26. SEAOC. 2013. Examples for Steel-Framed Buildings.
Vol. 4 of 2012 IBC SEAOC Structural/Seismic Design 38. Mattock, A. H. 1967. “Rotational Capacity of Hinging
Manual. Sacramento, CA: SEAOC. Regions in Reinforced Concrete Beams.” Journal of the
Structural Division 93 (2): 519–522.
27. Paulay, T., and M. J. N. Priestley. 1992. Seismic Design of
Reinforced Concrete and Masonry Buildings. New York, 39. Priestley, M. J. N., and R. Park. 1987. “Strength and
NY: John Wiley & Sons. Ductility of Concrete Bridge Columns under Seismic
Loading.” ACI Structural Journal 84 (1): 61–76.
28. Aragon, T. A., Y. C. Kurama, and D. F. Meinheit. 2020.
“Behavior of Ductile Short-Grouted Seismic Reinforcing 40. Berry, M. P., D. E. Lehman, and L. N. Lowes. 2008.
Bar-to-Foundation Connections under Adverse Construc- “Lumped-Plasticity Model for Performance Simulation
tion Conditions.” PCI Journal 65 (4): 33–50. https://doi of Bridge Columns.” ACI Structural Journal 105 (3):
.org/10.15554/pcij65.4-01. 270–279.
29. Brown, J., and S. Kunnath. 2004. “Low-Cycle Fatigue 41. Ameli, M. J., and C. P. Pantelides. 2016. “Seismic Analy-
Failure of Reinforcing Steel Bars.” ACI Materials Journal sis of Precast Concrete Bridge Columns Connected with
101 (6): 457–466. Grouted Splice Sleeve Connectors.” Journal of Structural
Engineering 143 (2). https://doi.org/10.1061/(ASCE)
30. Kunnath, S., Y. Heo, and J. Mohle. 2009. “Nonlinear Uni- ST.1943-541X.0001678.
axial Material Model for Reinforcing Steel Bars.” Journal
of Structural Engineering 135 (4): 335–343. https://doi 42. Saxey, B., Z. Vidmar, C.-H. Li, M. Reynolds, and C.-M.
.org/10.1061/(ASCE)0733-9445(2009)135:4(335). Uang. 2020. “A Predictive Low-Cycle Fatigue Model for
Buckling Restrained Braces.” 17th World Conference on
31. Mander, J., F. Panthaki, and A. Kasalanati. 1994. Earthquake Engineering, Sendai, Japan 2020/21.
“Low-Cycle Fatigue Behavior of Reinforcing Steel.”
Journal of Materials in Civil Engineering 6 (4): 453–468. 43. Bruneau, M., E. Kizilarslan, A. H. Varma, M. Broberg, S.
https://doi.org/10.1061/(ASCE)0899-1561(1994)6:4(453). Shafaei, and J. Seo. 2019. R-Factors for Coupled Com-
posite Plate Shear Walls/Concrete Filled (CC-PSW/CF).
32. Mander, J., M. J. N. Priestley, and R. Park. 1988. “The- AISC report CPF#05-17. https://www.aisc.org/technical
oretical Stress-Strain Model for Confined Concrete.” -resources/research/researchlibrary/r-factors-for-coupled
Journal of Structural Engineering 114 (8): 1804–1826. -composite-plate-shear-walls--concrete-filled-cc-pswcf.
https://doi.org/10.1061/(ASCE)0733-9445(1988)114
:8(1804). 44. Tauberg N. A., K. Kolozvari, and J. W. Wallace. 2019.
Ductile Reinforced Concrete Coupled Walls: FEMA
33. Coleman, J., and E. Spacone. 2001. “Localization Issues P695 Study. Los Angeles, CA: Department of Civil and
in Force-Based Frame Elements.” Journal of Structural Environmental Engineering, University of California, Los
Engineering 127 (11): 1257–1266. https://doi Angeles. https://www.concrete.org/publications
.org/10.1061/(ASCE)0733-9445(2001)127:11(1257). /internationalconcreteabstractsportal.aspx?m=details
&i=51722462.
34. Pugh, J. S., L. N. Lowes, and D. E. Lehman. 2015. “Non-
linear Line-Element Modeling of Flexural Reinforced 45. Vamvatsikos, D., and C. A. Cornell. 2002. “Incremental
Concrete Walls.” Engineering Structures 104: 174–192. Dynamic Analysis.” Earthquake Engineering and Struc-
https://doi.org/10.1016/j.engstruct.2015.08.037. tural Dynamics 31 (3): 491–514. https://doi.org/10.1002
/eqe.141.
35. Vásquez, J. A., J. C. De la Llera, and M. A. Hube. 2016.
“A Regularized Fiber Element Model for Reinforced 46. PEER (Pacific Earthquake Engineering Research Center).
Concrete Shear Walls.” Earthquake Engineering and 2006. PEER NGA Database. Berkeley: PEER, University
Structural Dynamics 45 (13): 2063–2083. https://doi of California, Berkeley. https://ngawest2.berkeley.edu.
.org/10.1002/eqe.2731.
Notation
36. Pozo, J. D., M. A. Hube, and Y. C. Kurama. 2021. “Effect
of Material Regularization in Plastic Hinge Integration Ag = gross area of beam or column section, neglecting
Analysis of Slender Planar RC Walls.” Engineering reinforcement
Structures 239 (15): 112302. https://doi.org/10.1016/j
.engstruct.2021.112302. Asc = steel core (yielding) area of brace

Asc,min = minimum steel core (yielding) area of brace Pu = factored design axial force of beam or column (pos-
itive for compression and negative for tension)
Ast = total area of longitudinal reinforcement in beam or
column section R = response modification coefficient
BI = numerical coefficient for effective damping Ry = adjustment factor for expected brace yield
strength
BRBC = adjusted brace force in compression
SD1 = design spectral response acceleration parameter at
BRBT = adjusted brace force in tension 1-second period
Cd = deflection amplification factor SDS = design spectral response acceleration parameter at

short periods
Cu = coefficient for upper limit on calculated period
SMT = maximum considered earthquake intensity
d = distance from extreme compression fiber to cen-
^
troid of longitudinal tension reinforcement SCT = median collapse intensity
db = nominal diameter of longitudinal reinforcing bar T = fundamental period
D = average distributed dead load Ta = approximate fundamental period
Ec = initial concrete stiffness VELF = design base shear force
fc' = specified (design) compressive strength of concrete Vmax = maximum base shear strength
fsy = specified (design) yield strength of reinforcing steel Vstory = shear force in story being designed
fymax = highest expected steel core yield strength of brace z = distance from critical section of beam or column to
point of contraflexure
fymin = lowest expected steel core yield strength of brace
α = angle of brace relative to horizontal
Fy = net upward force on beam due to adjusted brace
forces in chevron configuration β = adjustment factor for brace force in compression
g = gravitational acceleration βDR = uncertainty in design requirements
h = depth of beam or column section βI = effective damping due to inherent dissipation of

energy by structure, at or just below effective yield
Ie = seismic importance factor displacement of seismic-force-resisting system
Ig = gross moment of inertia of beam or column section, βMDL = modeling uncertainty

neglecting reinforcement
βRTR = record-to-record variability
L = live load
βTD = test data uncertainty
Lp = plastic hinge length
βTOT = total system collapse uncertainty
Mu = factored design moment of beam or column
δDBE = maximum dynamic roof drift demand at design
nb = number of braces in story being designed earthquake intensity
NQE = brace axial force under equivalent lateral forces δELF = roof drift of effective linear-elastic drift model
under equivalent lateral forces
Nu = factored design brace axial force
δu = ultimate roof drift
Po = nominal compression (uniaxial) strength of beam or
column at zero eccentricity δy,eff = effective yield roof drift

Δby = yield deformation of brace
Δmax = maximum brace deformation
Δinelastic = inelastic brace deformation
θ = nonlinear story drift
θe = effective linear-elastic story drift
μc = maximum cumulative inelastic ductility of brace
μmax = maximum ductility of brace
μT = period-based ductility of braced frame
ρ = redundancy factor
ΣMnb = sum of nominal moment strengths of beams fram-

ing into joint
ΣMnc = sum of nominal moment strengths of columns

framing into joint
φ = capacity reduction factor
ω = strain hardening adjustment factor
Ω0 = system overstrength factor

Shane Oh is a PhD candidate in This paper describes a numerical evaluation of the

the Department of Civil and seismic design of precast concrete buckling-restrained
Environmental Engineering and braced frames based on the Federal Emergency Man-
Earth Sciences at the University of agement Agency’s Quantification of Building Seismic
Notre Dame in Notre Dame, Ind. Performance Factors (FEMA P695) methodology. A
set of 32 archetype braced frames covering a range of
parameters were designed using a procedure consistent
with current U.S. building code requirements. Non-
Yahya C. Kurama, PhD, PE, is a linear numerical models were developed and verified
professor in the Department of against existing experimental data. The results show
Civil and Environmental Engi- that large axial compression and tension forces develop
neering and Earth Sciences at the in both beams and columns, thus requiring these mem-
University of Notre Dame. bers to be designed with large reinforcement ratios or
higher-grade reinforcing bars and beams to be designed
like column members. Recommended beam and col-
umn effective (cracked) linear-elastic axial and flexural
Jon Mohle, SE, is a senior product stiffness reduction factors provide reasonable estimates
and market manager at Clark of story drifts and brace deformations under design
Pacific in Sacramento, Calif. lateral forces. Nonlinear monotonic static pushover
and incremental dynamic analyses of the archetypes
support an overstrength factor of 2.5, response modi-
fication coefficient of 8, and deflection amplification
factor of 8 for the seismic design of this system.
Brandt W. Saxey, SE, is technical https://doi.org/10.15554/pcij66.5-03

director at CoreBrace LLC in Salt
Lake City, Utah. Keywords
Buckling-restrained braced frame, deflection ampli-

fication factor, FEMA P695, incremental dynamic
analysis, monotonic static pushover analysis, precast
concrete frame lateral system, response modification
coefficient, seismic design, seismic performance factor,
system overstrength factor.
Review policy
This paper was reviewed in accordance with the

Precast/Prestressed Concrete Institute’s peer-review
process.
Reader comments
Please address any reader comments to PCI Journal

editor-in-chief Tom Klemens at tklemens@pci.org or
Precast/Prestressed Concrete Institute, c/o PCI Jour-
nal, 8770 W. Bryn Mawr Ave., Suite 1150, Chicago, IL
60631. J

SPECIFY PCI CERTIFICATION
THERE IS NO EQUIVALENT
Photo courtesy of
USC/Gus Ruelas.
The Precast/Prestressed Concrete Institute’s (PCI) certification is the industry’s

most proven, comprehensive, trusted, and specified certification program. The
PCI Plant Certification program is now 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 demon-
strates 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
w w w.pci.o rg/cer tification

Directories
Board of Directors
Dennis Fink, Chair, Northeast Harry Gleich, Institute Program Phillip Miller, Producer Member
Prestressed Products LLC Director, Technical Activities, Director, PCI Gulf South
J. Matt DeVoss, Vice Chair, Jackson Metromont Corp. Alfred Miller Contracting
Precast Inc. Matt Graf, Producer Member Director, Chris Mosley, Professional Member
Matt Ballain, Secretary-Treasurer, PCI Illinois/Wisconsin, International Director, The Consulting Engineers
Coreslab Structures Concrete Products Inc. Group Inc.
(INDIANAPOLIS) Inc. Shelley Hartnett, Producer Member Patty Peterson, Institute Program
J. Seroky, Immediate Past Chair, Director, PCI Mountain States, Director, Business Performance
High Concrete Group LLC EnCon United Tindall Corp.
Bob Risser, President and CEO, PCI Lloyd Kennedy, Institute Program Jim Renda, Associate Member
Director, Educational Activities, Director, Supplier, Cresset Chemical
Dusty Andrews, Producer Member Finfrock Industries LLC Co.
Director, PCI Washington/Oregon, Brent Koch, Producer Member Cheryl Rishcoff, Professional
Knife River Corp.–Northwest Director, PCI West, Con-Fab Member Director, TRC Worldwide
Dennis Cilley, Associate Member California LLC Engineering Inc.
Director, Erector, American Steel & AJ Krick, Producer Member Director, Lenny Salvo, Producer Member
Precast Erectors Mid-Atlantic, Smith-Midland Corp. Director, Florida Chapter, Coreslab
Todd Culp, Producer Member Matt Mahonski, Producer Member Structures (ORLANDO) Inc.
Director, PCI Midwest, Coreslab Director, PCI Central, High Concrete Bob Sheehan, Associate Member
Structures (OMAHA) Inc. Group Director, Supplier, BASF Corp.
Jim Fabinski, Institute Program Bill Mako, Producer Member Director, Peter Simoneau, Producer Member
Director, Transportation Activities, Georgia/Carolinas PCI, Atlanta Director, PCI Northeast, Dailey
EnCon Colorado Structural Concrete Co. Precast
Greg Force, Institute Program David Malaer, Producer Member Gary Wildung, Institute Program
Director, Research and Director, Texas, Oklahoma, New Director, Quality Activities, FDG Inc.
Development, Tindall Corp. Mexico, Valley Prestress Products Inc. Daniel Eckenrode, Regional Council
Ashley Fortenberry, Associate Member Jane Martin, Institute Program Director, Representative, Nonvoting, PCI Gulf
Director, Erector, Tindall Corp. Marketing, Gate Precast Co. South
Technical Activities Council

Chair Harry Gleich Dusty Andrews John Lawler
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 Timothy Salmons
Ex-officio, Code Representatives Wayne Kassian Stephen J. Seguirant
S. K. Ghosh and Stephen V. Skalko Yahya Kurama 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
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
Cher Doherty (312) 583-6781 cdoherty@pci.org Manager, events
Walter Furie (312) 583-6772 wfurie@pci.org Production senior specialist
Christopher Hurst (312) 360-3203 churst@pci.org Membership project manager
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
Becky King (312) 360-3201 bking@pci.org
managing 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
John McConvill (312) 360-3208 jmcconvill@pci.org Controller
Sherrie Nauden (312) 360-3215 snauden@pci.org Continuing education manager
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
Arelys Schaedler (312) 583-6783 aschaedler@pci.org Executive assistant
Neal Sherman (312) 786-0300 nsherman@pci.org 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

Data quality and office administrative
Cindi Ward (312) 360-3214 cward@pci.org
services 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
Innovation
• Structured-light-scanning-based 3-D scanning for
process monitoring and quality control in precast,
prestressed concrete production
Also
• Expected compressive strength in precast,
prestressed concrete design: Review and discussion
of current practice
• Axial load limit considerations for 14 in. prestressed
concrete piles
• Live-load distribution of an adjacent box beam
bridge: Influence of bridge deck
• Meet Kenneth Kruse

Meet Mary Ann Griggas-Smith
Changing platforms
Sarah Fister Gale
M ary Ann Griggas-Smith started her

career at sea, designing offshore oil
rigs and submarines, but once she learned
“It got to the point where I knew that if I ever had a problem, I
could call any one of them and they would be willing to help.”
In 2012, Tindall offered Griggas-Smith a position as the
about precast concrete, everything changed. South Carolina division’s engineering manager. Then in 2014,
Griggas-Smith spent her youth traveling Tindall’s CEO Greg Force asked Griggas-Smith to start an
the world, living in far-flung locales, includ- in-house consulting group.
ing Baghdad, Iraq. After attending 10 differ- She accepted the position and today leads a team of 25 peo-
ent schools before graduating high school, she settled in New ple as Tindall’s director of corporate engineering.
Orleans, La., where she earned her bachelor’s degree in civil Recently, her team has been focused on the design of three
engineering from Tulane University. all–precast concrete residence halls at Western Carolina
Soon after she graduated, she landed a job with Shell work- University in Cullowhee, N.C., using the company’s new
ing on an offshore oil platform, and she says she loved it. She T-SLAB product, a patented precast concrete slab system
spent every other week living offshore and learning everything that uses super lightweight concrete to serve as blocks over
she could about the operations. “It was very hands-on work,” which structural concrete is placed for better load distribu-
she says. tion and span capability. Griggas-Smith’s group helped to
After a few years, she returned to Tulane to get her mas- develop the software to support design of the new product.
ter’s degree in civil engineering while continuing to work full “It’s an exciting innovation, and I think it will get us a lot of
time. She landed a job in Asheville, N.C., with a shipbuilding work across the Southeast.”
company, where she designed submarines and aircraft carriers.
She might have built a career with that company, but a layoff “I did a small precast concrete job, and thought, ‘Wow,
encouraged her to take work as a design consultant. this is really interesting.’”
“I did a small precast concrete job, and thought, ‘Wow, this
is really interesting,’” she says. That was the beginning of her
passion for precast concrete. When not working on committees or designing new
In 1996, she and a colleague started their own firm doing products, Griggas-Smith spends a lot of time mentoring
precast concrete design, working with many big precast con- young engineers at Tindall. She notes that many recent
crete companies, including Metromont, Gate, and Tindall. It graduates come to the company with little or no knowledge
was during that time that Griggas-Smith first discovered PCI. about precast concrete. “I wish the universities covered it
Early on, she was so busy building her firm that she only more,” she says.
dabbled in PCI meetings and events. When Tim Salmons Tindall is trying to close that gap by partnering with
encouraged her to get more involved, she decided to make it a Clemson University in South Carolina, the University of
bigger priority. North Carolina at Charlotte, and North Carolina State
Over the years, Griggas-Smith has contributed to the University in Raleigh to perform precast concrete research
eighth edition of the PCI Design Handbook: Precast and projects and by hiring interns even during the pandemic. “We
Prestressed Concrete and is now a member of the PCI look at it as a summerlong interview process,” she says of the
Handbook Committee and chair of the Industry Handbook internship program.
Subcommittee Chapter 13 for the ninth edition. She is also She encourages other companies to make similar invest-
a member of the Fire Committee, the Total Precast Systems ments in the next generation. “Give them a chance to get
Committee, and the Technical Activities Council. their hands dirty and to spend time on the job sites,” she says.
“It was great to meet all of these people in person,” she says. “That’s how they will develop a passion for the work.”

HAMILTON FORM
INDUSTRY INNOVATORS
“Hamilton Form guided our company with a high level

of expertise as we collaborated on the form design.
The forms are innovative and efficient, which will
allow our company to manufacture a high-quality
product with minimal effort. Their knowledge and
teamwork throughout this process has been a key
in this project’s success.”
Ron Sparks, VP/GM, Columbia Precast Products

OUR INNOVATIVE CUSTOM FORMS,
BRING YOUR PROJECTS TO REALITY.
Hamilton Form is known for high-quality, hardworking custom forms and
equipment, like the forms built for Columbia Precast to produce sound
walls for the Washington State DOT I-90 corridor improvement project.
The forms cast 8-foot sound walls with integral monolithic pilasters and
were designed to accept custom form liners. Internal headers are used
to adjust wall heights from 31-35 feet. Hydraulically-actuated side forms
make stripping and set-up efficient and easy.
When you need innovative formwork and custom equipment solutions,

call on Hamilton Form to deliver: 817 590-2111 or
sales@hamiltonform.com
Hamilton Form Company, Ltd.

7009 Midway Road, Fort Worth, Texas 76118
www.hamiltonform.com

Parking Structures: Precast/Prestressed Concrete Institute

Uploaded by

Document Information

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

Parking Structures: Precast/Prestressed Concrete Institute

Uploaded by

Copyright:

Available Formats

SEPTEMBER–OCTOBER 2021

VOLUME 66, NUMBER 5

Precast/Prestressed Concrete Institute

The PCI Safety Culture

See how it works:

Your Connection Connection

23 Influence of Hollow-Core Wall Panels on the Cyclic Behavior

Parsa Monfaredi, Mehdi Nazarpour, and Abdoreza S. Moghadam

Seismic Design and Analysis of Precast Concrete 54

PCI Journal | September–October 2021

From PCI Headquarters 9

PCI Staff Directory 86

JOURNAL ADVISORY COMMITTEE Millard J. Barney Pinar Okumus Coming Ahead 87

EDITORIAL DESIGN & PRODUCTION

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

PCI Journal | September–October 2021

PCI Journal | September–October 2021 5

PCI Foundation at 20: Making

PCI Journal | September–October 2021 7

New PCI white paper Fifth edition of PCI quality

A new PCI white paper lays out the institute’s view

Brown added as new

P CI welcomes Trina Brown, as the new

PCI Journal | September–October 2021 9

Universities Locations Programs Professors

Chico State University

University of Delaware Newark, Del. • Engineering Jovan Tatar

• Engineering George Morcous

The transition of certified architectural plants to the new

T he transition to the new architectural certification cate-

10 PCI Journal | September–October 2021

Professors Seminar attracts

T he 2021 PCI Foundation Professors Seminar was

PCI releases new certificate

P CI-certified plants will notice some changes to their plant

PCI Journal | September–October 2021 11

efforts to attract talent to the company. He is

R evisions to PCI Policy 29 on the Erector Certification

PCI personnel training and certification schools

12 PCI Journal | September–October 2021

PCI Journal | September–October 2021 13

Metromont promotes Aultman mentor for young engineers. She is heav-

M etromont Corp. has promoted Suzanne Aultman from

United Steel American Progress Group Corp.

PCI Journal | September–October 2021 15

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

16 PCI Journal | September–October 2021

T he Advanced Building Construction (ABC)

Event details are subject to change.

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

PCI Journal | September–October 2021 17

KANSAS CITY CONVENTION CENTER | KANSAS CITY, MISSOURI

Pavilion parking structure

A new five-level underground precast concrete parking struc-

project. Shockey Precast, based in Winchester, Va., was select-

PCI Journal | September–October 2021 21

22 PCI Journal | September–October 2021

Mohamed K. Nafadi, Gregory Lucier, Tugce Sevil Yaman,

PCI Journal | September–October 2021 23

24 PCI Journal | September–October 2021

PCI Journal | September–October 2021 25

26 PCI Journal | September–October 2021

Figure 3. Typical views of panels during fabrication.