SEAOC 2010 CONVENTION PROCEEDINGS

Seismic Retrofit of Unreinforced Masonry Infill Walls Using

Advanced Composite Materials
Benson Shing
Dept. of Structural Engineering, UCSD
San Diego, CA
Ioannis Koutromanos
Dept. of Structural Engineering, UCSD
San Diego, CA
Andreas Stavridis
Dept. of Structural Engineering, UCSD
San Diego, CA
Scott Arnold, PE
Fyfe Co. LLC
San Diego, CA

Abstract
Structural retrofit schemes were developed to strengthen
unreinforced masonry infill walls in reinforced concrete (RC)
structures. A structural proof test program was carried out on
the worlds largest outdoor shake table, located at UCSDs
Englekirk Structural Engineering Center. A two-thirds scale
specimen of a three-story, two-bay, RC frame infilled with
two-whythe brick walls was tested under multiple intensities
of the Gilroy record of the 1989 Loma Prieta earthquake and
the north-south component of the 1940 El Centro earthquake.
Two retrofit schemes were examined on the structure: the
application of an engineered cementitious composite (ECC)
to an undamaged first-story wall and the installation of glass
fiber-reinforced polymer (GFRP) composites to a damaged
second-story wall with horizontal and diagonal cracks along
the mortar joints induced by prior shaking. The fiberwrap not
only fully restored the capacity of the damaged walls but also
provided additional strength and stiffness. The retrofit
maintained its structure while suffering no damage under
150% of the Gilroy earthquake and 190% of the El Centro
earthquake. This paper focuses on the fiberwrap composite
retrofit of the second story walls and discusses the repair

considerations, detailing recommendations, and advantages

of a fiberwrap retrofit scheme when performing a seismic
upgrade.
Introduction
In many countries, reinforced concrete (RC) structures are
often infilled with unreinforced brick walls to serve as
interior and exterior partitions. Such construction can be
found in older buildings in the western United States,
including pre-1930s buildings in California, and newer
buildings in the midwestern and eastern parts of the country.
Although unreinforced masonry infill walls are often treated
as non-structural components, they will positively interact
with the bounding frames when subjected to earthquake
loads.
Many engineers have recognized the benefits of infill walls
in protecting a structure against earthquake loads.
Nevertheless, infill walls were sometimes associated with
catastrophic failures and undesired soft-story mechanisms of
RC structures in a number of severe earthquake events (e.g.,
see EERI 2000). These failures were mostly caused by the

poor quality of the structural systems, inferior infill materials,

and/or the absence of infill walls in critical locations such as
open first stories. The ability to assess the performance and
safety of these RC structures is important for retrofit
decisions. However, such analysis presents a major
challenge. The interaction of an RC frame with masonry
infill can result in a number of possible failure mechanisms
including the cracking and crushing of the infill walls and the
shear failure of the columns. The lateral load resistance of an
infilled frame highly depends on the failure mechanism that
develops. In the US, guidelines for the evaluation of infilled
frames are provided in FEMA reports (FEMA 1998, 2000),
International Existing Building Codes (ICBO 2001), and
ASCE/SEI 41: Seismic Rehabilitation of Existing Buildings
(ACSE/SEI 2007). Nevertheless, they are far from
satisfactory in terms of completeness and reliability.
ASCE/SEI 41 allows the use of nonlinear finite element
models, but does not provide any guidelines for such
analysis. In addition, there are few validated simple retrofit
methods that can be used for these structures.
To address the aforementioned issues, a collaborative
research project has been carried out by a team of researchers
from the University of California at San Diego, Stanford
University, and the University of Colorado at Boulder. The
research objectives are two-fold. One is to develop rational
and reliable analytical methods for assessing the seismic
performance of masonry-infilled RC frames, and the other is
to develop effective seismic retrofit techniques. The retrofit
schemes and analytical tools were validated with quasi-static
tests conducted on small- and large-scale frame specimens
representing
1920-era
construction
in
California.
Additionally, shake-table tests were conducted on two 2/3scale, three-story, two-bay, infilled RC frames. One frame
was tested without retrofit measures. The other was
strengthened with engineered cementitious composite (ECC)
and glass fiber-reinforced polymer (GFRP) composites in the
first and second stories, respectively. This paper summarizes
the large-scale shake-table tests conducted at the NEES
(Network for Earthquake Engineering Simulation) Site at the
University of California at San Diego. Furthermore, the
developed computational models are described, and
numerical analysis results are compared to the experimental
data. Major observations and conclusions are presented.
Prototype Building
The focal point of the experimental study is a prototype
building typical of 1920-era RC construction in California.
The structure is a non-ductile RC frame with unreinforced
masonry infill walls on the exterior. The plan view of the
structure and the elevation view of an external infilled frame,
which is considered in this study, are presented in Figure 1.

Exterior frame considered

(a) Plan View

(b) Elevation View of Exterior Frame

Figure 1 Prototype Building
Although the building represents older construction, it was
designed using the properties of contemporary construction
materials, which were employed in the experimental
investigation. A working stress design approach was used
with the allowable stresses calculated with factors of safety
used in the 1920s. In accordance with the engineering
practice of that period, the RC members were designed only
for gravity loads that included the self-weight, the weight of
the slabs, toppings, ceilings, masonry walls, parapets, and the
live load. The design live load was 75 psf for the 1 st and 2nd
stories, and 20 psf for the roof. The live load was reduced by
60% for the column design, and 20% and 40% for the design
of the exterior and interior beams, respectively.
Shake-Table Tests & Retrofit Schemes
Shake-table tests were conducted on two 2/3-scale models of
the prototype frame shown in Figure 1b. The dimensions and
reinforcement layout of the concrete frame specimens are
shown in Figure 2, and the setup of the two tests is depicted

in Figure 3. In the tests, the specimens were subjected to a

sequence of ground motions which were scaled versions of
the Gilroy record obtained in the 1989 Loma Prieta
Earthquake and the NS component of the El Centro record
obtained during the 1940 Imperial Valley Earthquake. The
time scale of the records was compressed to satisfy the
similitude requirements. The gravity load was closely
simulated with added mass attached to the frames at each
level, but the seismic mass was smaller than the actual. This
was compensated for by scaling up the acceleration and
compressing the ground motion time.
3'-37 "
8

2#8

2#4+1#3

2#4+1#3
2#3
#2@10.5''

2#8

1'-221"

2#8

4#8

2#4+1#3
2#3
#2@10.5''

2#4+1#3

11"

2#8

2#8

2#8

4#8

1'-10"

3'-37"
8

11"

2#8

2#8

2#4+1#3

2#4+1#3
2#3
#2@10.5''

the beginning, the bottom story had an unretrofitted infill wall

with a door opening in one bay. Upon the development of
mild damage in the second-story infill walls, the bottom-story
wall with the door opening was removed (as shown in Figure
3b), and the specimen was tested again. After the secondstory walls had incurred moderate damage, the formed cracks
were repaired with epoxy injection and then strengthened
with a thin glass fiberwrap composite overlay on one side of
the wall. The fiberwrap composite system consisted of epoxy
and one layer of bi-directional glass fabric with fibers
oriented at 45o with respect to the bed joints of the masonry.
In addition, the masonry around the window opening was
reinforced with one-foot-wide strips of uni-directional glass
fabric to provide trim reinforcement.

7'-4"

19" lap splice

2#8

2#5+1#4

2#8

2#8

2#8

2#8

4#8

1'-221"

2#8

4#8

7'-4"

2#5+1#4
2#4
#2@10.5''

2#5+1#4

2#5+1#4
2#4
#2@10.5''

2#8

2#8

2#5+1#4

2#5+1#4
2#4
#2@10.5''

19" lap splice

2#8

2#8

2#8
2#8

2#8

4#8

2#8

2#8

1'-212"

2#8

4#8

7'-4"

3#5

3#5
2#5
#2@10.5''

3#5
2#5
#2@10.5''

3#5

3#5

0.5" thick construction joint

3#5
2#5
#2@10.5''

0.5" thick construction joint

PVC tubes, / =2.5'' @ 2'

1'-478 "

1'-6"

(a) Specimen 1

#3 @6''
1'-712"

2'

2'

2'

2'

2'

2'

2'

2'

2'

2'

2'

2'

2'

2'

1'-721"

Figure 2 Design of Test Frame

To assure that the RC columns had sufficient capacity to
resist the overturning moment in the second specimen, where
some of the infill walls were strengthened, the columns in the
two specimens had a higher quantity of flexural
reinforcement than the original design. Moreover, lap splices
were eliminated at the base of the bottom-story columns. All
the infill walls had two wythes of brick units. The first
specimen (Figure 3a) had solid infill walls in one bay and
infill walls with window openings in the other. Similar to the
first frame, the second specimen (Figure 3b) had solid infill
walls in one bay and window openings in the second and
third stories in the other bay. However, the solid infill in the
bottom story was strengthened with an overlay of ECC,
attached on each side of the wall with stitch dowels spaced at
12-in in each direction. This retrofit technique is described in
Billington et al. (2009). The ECC was sprayed on the wall
and had wire-mesh reinforcement and shear dowels to
connect to the beam and base slab. Each overlay was about 1in. thick near the mid-height of the wall and 2-in. thick near
the top and bottom to provide appropriate cover for the shear
dowels. Specimen 2 was tested under several scenarios. At

(b) Specimen 2
Figure 3 Shake-Table Tests
In general, Specimen 1 performed well with only minor
cracks after the 67% Gilroy motion. After the 100% Gilroy,
the specimen had severe cracks in the bottom-story infill
walls, but remained structurally stable. Shear failure
developed in the middle column during the 120% Gilroy. It
should be mentioned that for the original Gilroy record, the
spectral intensity of the 120% motion corresponds to a
representative Maximum Considered Earthquake (MCE) for
Seismic Design Category D (ASCE/SEI 2005) based on the
initially estimated structural period of 0.1 sec. However, the

actual table motion has a spectral intensity about 50% higher,

while its Peak Ground Acceleration (PGA) is 40% higher
than the MCE. The identified fundamental period of the
structure based on white-noise input varies from about 0.06
sec. initially to 0.19 sec. after the 120% Gilroy (Stavridis
2009). Damage was localized in the bottom story as shown in
Figures 4a and 4b. Severe damage with shear failures in the
exterior columns occurred when the specimen was subjected
to the 250% El Centro as shown in Figures 4c and 4d. The
wall with a window opening almost completely collapsed
because of the lack of an arching mechanism to restrain the
out-of-plane motion of the brick units above the opening.
(c) After 250% El Centro
Figure 4 Damage in Bottom Story of Specimen 1
Before the second-story walls in Specimen 2 were retrofitted
with fiberwrap composites, they had severe cracks after the
100% Gilroy, as shown in Figures 5a and 5b, while the firststory walls remained intact even without the infill in one bay.

(a) After 120% Gilroy

(a) Second Story Before GRFP Retrofit

(b) After 120% Gilroy

(b) Second Story Before GFRP Retrofit

Figure 5 Damage in Specimen 2
(c) After 250% El Centro

The cracks of the second-story walls were epoxy injected in

order to regain the walls stiffness. After all the cracks were
epoxy injected (as shown in Figure 6a), the walls were

prepared by grinding the paint off the walls prior to applying

the fiberwrap composite (as shown in 6b). After the walls
were repaired, they were retrofitted with bi-directional and
uni-directional fiberwrap composite systems, as shown in
Figure 6c.

(a)

fiberwrap strains (after G150, the bottom story failed, and

thus the developed strains in the fiberwrap retrofit for
subsequent motions were smaller). The numbers of the
motions correspond to percentage scaling factor of the
reference motion, e.g., G120 corresponds to the Gilroy
Reference motion, with the amplitude scaled by 120% = 1.20.

Epoxy Injection of 2nd Story Wall

(a) Placement of Strain Gages on 2nd Story Wall
8

-4

x 10

GFRPEEM
GFRPEEP

strain

-4

(b) Grinding of 2nd Story Wall

-8
10

12

14
16
time (sec)

18

20

(b) Time History at Strain Gages EEP/EEM at G120

Motion
-4

x 10

GFRPEWM
GFRPEWP

6
4

(c) Second Story with GFRP Retrofit

Figure 6 Retrofit Scheme of Specimen 2 Second Story
Strain gages were placed at second story wall locations as
shown in Figure 7a to measure the strains of the second story
unreinforced masonry brick infill wall strengthened with
fiberwrap composites during the seismic testing. The time
histories were recorded to correspond to the motions G120
and G150 (as shown in Figures 7b, 7c, 7d, 7e, 7f, 7g and 7h),
which were the most demanding in terms of developed

strain

2
0
-2
-4
-6
-8
10

12

14

16

18

20

time (sec)

(c) Time History at Strain Gages EWP/EWM at G120

Motion

-4

-4

x 10

x 10

GFRPWEM
GFRPWEP

GFRPW WM
GFRPW WP

strain

strain

1
0

-1
-1
-2
-3
10

12

14
16
time (sec)

18

-2
10

20

(d) Time History at Strain Gages WWP/WWM at G120

Motion

14
16
time (sec)

18

20

(g) Time History at Strain Gages WEP/WEM at G150

Motion

-4

12

-4

x 10

4
GFRPWEM
GFRPWEP

x 10

GFRPW WM
GFRPW WP

strain

strain

1
0

0
-1

-1

-2
-3

-2
10

12

14
16
time (sec)

18

20

(e) Time History at Strain Gages WEP/WEM at G120

Motion
-3

x 10

strain

-1

12

14
16
time (sec)

18

12

14
16
time (sec)

18

20

(h) Time History at Strain Gages WWP/WWM at G150

Motion
Figure 7 Strain Gage Placement and Time Histories on
2nd Story Wall

GFRPEW M
GFRPEW P

-2
10

-4
10

20

(f) Time History at Strain Gages EWP/EWM at G150

Motion

After the repair and retrofit of the second-story walls, no

further damage was observed for subsequent ground motions.
However, the infill walls in the third story eventually
developed cracks. No cracks were observed in the ECC, but
the shear dowels connecting the ECC to the beam were
damaged as evidenced by mild concrete cracking and spalling
near the dowel locations. Noticeable sliding occurred
between the ECC and the beam. While all three bottom-story
columns in Specimen 1 had shear failures, Specimen 2 had
only one severe shear crack developed in the beam-to-column
joint right above the exterior column adjacent to the infill in
the first story after the 190% El Centro. The crack first
developed during the 150% Gilroy.
Damage developed in the second-story walls of Specimen 2
at a story-shear much lower than expected. The effect of the
fiberwrap strengthening (including crack injection) of the

second story of Specimen 2 is shown in Figure 8. The repair

and retrofit enhanced both the strength and stiffness. No
additional damage was observed in the second story infill
walls after the retrofit. The drift time history for the second
story of the specimen under the Gilroy motion scaled to a
level that is 40% higher than the Maximum Considered
Earthquake (MCE) is shown in Figure 8. In that test,
Specimen 2 had one infill wall in the bottom story already
removed. Figure 9 shows that Specimen 2 had a larger drift
in the second story than Specimen 1. Specimen 1 exhibited a
distinct weak first story which isolated the second story from
any severe seismic force. As a result, no major cracks were
developed in the second and third stories of the specimen.
300

shear force (kips)

200

Figure 9 2nd Story Drift Time History with Gilroy Record

Scaled to 40% Higher than MCE

100
0

Advantages of Fiberwrap Composites

Gilroy 67%
Gilroy 83%
Gilroy 91%
Gilroy 100%

-100
-200
-300
-0.3

-0.2

-0.1
0.0
0.1
interstory drift (in)

0.2

0.3

(a) Before Retrofit with Fiberwrap

300

shear force (kips)

200
100
0

Gilroy 67%
Gilroy 83%
Gilroy 91%
Gilroy 100%
Gilroy 120%
Gilroy 150%

-100
-200
-300
-0.3

-0.2

-0.1
0.0
0.1
interstory drift (in)

0.2

0.3

(b) After Retrofit with Fiberwrap

Figure 8 Comparison of 2nd Story Shear Force-vs.-Story Drift
Hysteresis Curves Before and After Retrofit of Specimen 2

The use of fiberwrap composites for structural strengthening

started with seismic retrofits but has grown over the years.
There are many material advantages of fiberwrap composites
which include high strength-to-weight and stiffness-to-weight
ratios, corrosion resistance, material flexibility and relative
ease of installation. The material can be designed to
rehabilitate concrete, masonry, and steel structures and its
flexibility allows it to mold to any shape or existing structure
with minimal modifications. Fiberwrap repairs are
economical and eco-friendly because the lightweight material
eliminates the need for heavy, costly equipment, since the
necessary supplies can be carried on a truck. The application
usually requires no more than a single crew for section repair,
and automated equipment expedites the process. All factors
considered, fiberwrap rehabilitation provides a rapid and
cost-effective solution for seismic retrofits. The fiberwrap
method has had a solid performance record in many structural
settings and its acceptance is gradually increasing as longterm durability tests are made available. This rapid, costfriendly solution offers value, proven performance and
reliability for the engineering community.
There are more traditional options available in addition to the
fiberwrap composites used for the seismic retrofit of the
unreinforced brick infill walls on the second story like
shotcrete and steel that should be considered in order to be
sure the most cost-effective solution is provided. Had
traditional solutions such as concrete or steel been used for
this test, more weight would have been added to the test
frame thus increasing the loading and causing more damage
to the first story of the test structure. The major benefit of
using fiberwrap for this test was that the seismic retrofit was

effectively provided to the unreinforced brick infill walls on

the second story without adding any weight to the test frame.
The installation was completed in only two days and the
system was cured and ready to take load after only 48 hours.
The reduced impact to tenants and relative speed of the
installation can often prove to make the fiberwrap option
much more cost effective. Fiberwrap composites have been
thoroughly tested over 20 years for long-term durability and
if properly designed the fiberwrap can last the life of the
structure. Fiberwrap composites have many advantages and
can be a feasible strengthening option for seismic retrofit
applications due to its past extensive testing and successful
applications over the years.
Detailing Recommendations
Like any structural repair, all fiberwrap advantages depend
heavily on the proper installation and design of the composite
system. Although installation is straightforward, it is critical
that design and application measures are followed to ensure
system performance and retrofit lifespan. In addition,
applicators must be trained and have written certification
from fiberwrap manufacturers before contracting such
rehabilitation projects. Good detailing and workmanship is a
very important part of the design and field application
processes of the fiberwrap composite systems. No matter the
size of the application, each one that has been tested or
completed to date using fiberwrap is unique and cannot be
designed and installed purely on the basis of past
applications. Even with proper fiberwrap design, the skills of
the contractor installing the material can prove to be a weak
link if they have not been properly trained and certified.
Therefore, the fiberwrap system must be properly designed
and installed by specialty contractors in order to provide a
cost-effective and efficient long-term retrofit solution.
Conclusions
The research presented herein was part of a larger project
aiming to the development of reliable analytical models and
effective retrofit techniques for masonry-infilled, non-ductile,
RC frames. Sub-assemblages of a prototype structure were
tested to validate the proposed analytical tools and retrofit
methods. These included two 2/3-scale, three-story, two-bay,
infilled frame specimens, which were tested on a shake table.
The experimental results have shown that unreinforced
masonry infill walls can contribute significantly to the
seismic resistance of a concrete frame. However, the failure
behavior of such structures could be brittle, leading to a
sudden collapse associated with the development of a weakstory mechanism. It has also been concluded that an infill
wall with an opening could be more vulnerable to collapse
because of the lack of an arching action. The effectiveness of
the retrofit measures using the ECC and fiberwrap

composites, which can increase both the stiffness and

strength of an infilled frame, has been experimentally
verified. A finite element modeling strategy combining the
discrete and smeared crack approaches and a cyclic cohesive
interface model have been successfully developed to capture
the nonlinear load-displacement response and the
complicated failure mechanisms of infilled frames
characterized by the shear failure of the columns and the
fracture of the masonry infill.
Acknowledgments:
The study presented in this paper was supported by the US
National Science Foundation Grant No. 0530709 awarded
under the George E. Brown, Jr. Network for Earthquake
Engineering Simulation (NEES) program. The authors would
like to thank Prof. Kaspar Willam and graduate student Ben
Blackard at the University of Colorado, Boulder, and Prof.
Sarah Billington and graduate student Marios Kyriakides for
their contributions in this project. Input of the Professional
Advisory Panel (PAP) is also gratefully acknowledged. The
panel members are David Breiholz, John Kariotis, Gregory
Kingsley, Joe Maffei, Ron Mayes, Paul Murray, and Michael
Valley. The fiberwrap retrofit of the masonry walls was
performed by Fyfe Co. The authors are also grateful to the
dedicated laboratory staff at UCSDs Englekirk Center for
setting up and conducting the shake-table tests. However, the
opinions expressed in this paper are those of the authors and
do not necessarily represent those of the NSF, PAP, or other
collaborators.
References:
ASCE/SEI (2005), Minimum Design Loads for Buildings
and Other Structures (ASCE/SEI 7-05), ASCE/SEI,
Reston, VA.
ASCE/SEI (2007), Seismic Rehabilitation of Existing
Buildings, ASCE/SEI Standard 41-06, ASCE/SEI,
Reston, VA.
Billington, S.L., Kyriakides, M.A., Blackard, B., Willam, K.,
Stavridis, A. and Shing, P.B. (2009), Evaluation of a
Sprayable, Ductile Cement-Based Composite for the
Seismic Retrofit of Unreinforced Masonry Infills,
Proceedings, ATC & SEI Conference on Improving the
Seismic Performance of Existing Buildings and Other
Structures, Applied Technology Council and Structural
Engineering Institute, San Francisco, CA.
Blackard, B., Willam, K., and Mettupulayam, S. (2009),
Experimental Observations of Masonry Infilled RC
Frames with Openings, ACI Special Publication SP-265,
American Concrete Institute, Farmington Hills, MI.
EERI (2000), 1999 Kocaeli, Turkey Earthquake
Reconnaissance Report, Earthquake Spectra, Vol. 16,
Issue S1.
FEMA (1998), Evaluation of Earthquake Damaged Concrete
and Masonry Wall Buildings: Basic Procedures Manual,

FEMA 306, Federal Emergency Management Agency,

Washington, D.C.
FEMA (2000), Prestandard and Commentary for the Seismic
Rehabilitation of Buildings. FEMA 356, Federal
Emergency Management Agency, Washington, D.C.
ICBO (2001), Guidelines for Seismic Retrofit of Existing
Buildings, International Conference of Building
Officials, Whittier, CA.
Lotfi, H. R. and Shing, P.B. (1994), An Interface Model
Applied to Fracture of Masonry Structures, Journal of
Structural Engineering, ASCE, 120(1): 63-80.
Lotfi, H. R. and Shing, P. B. (1991), An Appraisal of
Smeared Crack Models for Masonry Shear Wall
Analysis, Journal of Computers and Structures, ASCE,
Vol. 120(1): 63-80.
Stavridis, A. and Shing, P.B. (2010), Finite Element
Modeling of Nonlinear Behavior of Masonry-Infilled RC
Frames, Journal of Structural Engineering (in press).
Stavridis, A. (2009). Analytical and Experimental Study of
Seismic Performance of Reinforced Concrete Frames
Infilled with Masonry Walls, Doctoral Dissertation,
Department of Structural Engineering, University of
California, San Diego, La Jolla, California, USA.
Willam, K., Blackard, B. and C. Citto, 2008, Performance
Evaluation of Masonry Infills subjected to Axial and Inplane Shear: Meso- and Macro Experiments and
Simulations, Proceedings, 5th European Congress on
Computational Methods in Applied Sciences and
Engineering, Venice, Italy.