See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/274272551

Seismic Performance of Engineered Masonry Buildings in the 2010 Maule

Earthquake

Article  in  Earthquake Spectra · June 2012

DOI: 10.1193/1.4000040

CITATIONS READS

28 2,104

4 authors:

Maximiliano Astroza María Ofelia Moroni

University of Chile University of Chile
54 PUBLICATIONS   543 CITATIONS    44 PUBLICATIONS   475 CITATIONS   

SEE PROFILE SEE PROFILE

Svetlana Brzev Jennifer Tanner

University of British Columbia - Vancouver University of Wyoming
50 PUBLICATIONS   353 CITATIONS    35 PUBLICATIONS   131 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Masonry View project

Passive seismic protection devices made using shape memory alloys View project

All content following this page was uploaded by Maximiliano Astroza on 21 March 2018.

The user has requested enhancement of the downloaded file.

Seismic Performance of Engineered
Masonry Buildings in the 2010
Maule Earthquake
Maximiliano Astroza,a) Ofelia Moroni,a) M.EERI, Svetlana Brzev,b) M.EERI,
and Jennifer Tanner,c) M.EERI

Engineered masonry, namely reinforced and confined masonry, has been

widely used for housing construction in Chile over the last few decades.
Most one- and two-story single-family masonry dwellings did not experience
any damage due to the 27 February 2010 Maule earthquake, with the exception
of a few dwellings of pre-1970 vintage, which suffered moderate damage. A simi-
lar statement can be made for three- and four-story confined masonry buildings: a
large majority of buildings remained undamaged. However, several reinforced
and partially confined three- and four-story masonry buildings suffered extensive
damage, and two three-story partially confined buildings collapsed. The key
damage patterns and the causes of damage are discussed in the paper. The extent
of damage observed in the field was correlated with calculated vulnerability
indices, and relevant recommendations were made related to the design and
construction practices. [DOI: 10.1193/1.4000040]

INTRODUCTION
A majority of buildings affected by the 27 February 2010 earthquake can be classified as
housing construction. Masonry is the most popular housing construction technology in Chile,
accounting for 38.6% of the national housing stock, according to the 2002 Census (INE
2002). In Regions VI and VII of Chile, masonry construction constitutes more than 50%
of the overall housing stock. In particular, in Maule Region (VII) most severely affected
by the earthquake (total population of 908,097), masonry constitutes 50.4% of the total hous-
ing stock. Adobe and timber construction account for 22% and 21%, respectively, of the
housing stock, while reinforced concrete (RC) construction accounts for 4% of the housing
stock, as shown in Figure 1.
Typical forms of housing include low-rise single-family dwellings (up to two stories
high), and medium-rise apartment buildings (three to five stories high). Older masonry build-
ings in the affected area (built before the 1950), either of unreinforced brick masonry or
adobe construction, were severely damaged in the earthquake. A majority of masonry build-
ings built since the 1930s are of confined masonry construction. Reinforced masonry con-
struction practice started in the 1970s, and it represents a small fraction of the building stock.
Most of confined and reinforced masonry buildings is considered engineered construction,

a)
Department of Civil Eng, University of Chile, Blanco Encalada 2002, Santiago, Chile
b)
Columbia Institute of Technology, Burnaby, British Columbia, Canada
c)
Department of Civil and Architectural Engineering, University of Wyoming, Laramie

S385
Earthquake Spectra, Volume 28, No. S1, pages S385–S406, June 2012; © 2012, Earthquake Engineering Research Institute
S386 ASTROZA ET AL.

Figure 1. Distribution of housing according to the construction material in Region VII of Chile
(INE 2002).

because the construction practice is based on the provisions of building codes and guidelines,
some of which date back to the 1940s. In many cases, designs of engineered masonry build-
ings have been standardized because these buildings were built as social housing. These
building designs are similar and can be used to compare seismic performance of similar
buildings located at different sites.
The scope of this paper is limited to observations related to seismic performance of engi-
neered masonry buildings in the earthquake. For more information on seismic performance of
other types of masonry construction, the reader is referred to Astroza et al. (2012).

TYPICAL BUILDING CONSTRUCTION

Masonry buildings in the earthquake-affected area can be classified as low-rise single-family
dwellings and medium-rise apartment buildings. Single-family dwellings are in the form of
either one-story detached houses or two-story houses in a row (i.e., townhouses). A typical con-
fined masonry townhouse is shown in Figure 2a. Plan dimensions for a typical unit are approxi-
mately 5 m × 6 m, while the clear floor height is on the order of 2.2 m to 2.3 m. The front façade
of the house is usually perforated with openings, while the transverse walls are solid. One- and
two-story dwellings have timber floors and pitched timber roofs with timber gables.
Apartment buildings are usually three- to four-story high, with two to six apartments per
floor (depending on the building size). Typical plan dimensions are as follows: length
ranging from 12 to 30 m and width from 5 m to 8 m. A typical three-story apartment building
is shown in Figure 2b. Most buildings have a regular and symmetric configuration. More than
80% buildings have a rectangular floor plan and regular configuration with insignificant
eccentricity between the center of mass and the center of rigidity. In both directions,
there are seismic force-resisting systems that provide a complete horizontal and vertical
load path. The distance between transverse walls ranges from 7 m to 8 m. These buildings
usually have RC floors and pitched timber roofs with timber trusses supported by RC peri-
meter beams. RC floors are either cast-in-place or precast, with large hollow masonry blocks
SEISMIC PERFORMANCE OF ENGINEERED MASONRY BUILDINGS IN THE 2010 MAULE EARTHQUAKE S387

Figure 2. Typical housing construction in Chile: (a) Two-story townhouses and (b) a three-story
apartment building.

(called bovedillas in Spanish) laid horizontally between precast reinforced concrete beams
(this is known as “Tralix” system).

CONFINED MASONRY
Chile has a long history of confined masonry construction practice in the world. The use
of confined masonry in Chile started in the 1930s, after the 1928 Talca earthquake (Mw 7.7),
and it was initially used for low-rise single-family housing. The construction of medium-rise
apartment buildings started in the 1970s in the capital Santiago, and in the 1990s in other
urban areas. Good performance of low-rise confined masonry buildings in the 1939 Chillan
earthquake (M 7:8) paved the road for the widespread use of this construction technology in
Chile. Low-rise confined masonry construction maintained a good performance record in
past earthquakes, including the 1985 Llolleo earthquake (Mw 8.0). Very few medium-rise
confined masonry buildings existed in the epicentral zone of the 1985 earthquake, so it
can be considered that three- and four-story confined masonry buildings had not been
exposed to severe ground shaking in Chile prior to the February 2010 earthquake.
Key components of confined masonry buildings are masonry walls and RC confining
elements (tie-columns and tie-beams). Unreinforced masonry wall panels are constructed
first, one story at a time, followed by the cast in-place RC tie-columns, as shown in Figure 3a.
Finally, RC tie-beams are constructed on top of the walls, simultaneously with the floor/roof
slab construction. The wall thickness depends on the type of masonry units used. Most com-
mon masonry units used for confined masonry walls are 140 mm machine-made hollow clay
blocks (tiles), followed by the handmade solid clay bricks. Hollow concrete block units
(usually 150 mm thick) are rarely used in the earthquake-affected area, although they
were found in the walls of one of the collapsed confined masonry buildings in Santa
Cruz. Chilean masonry standard prescribes values for compressive strength of masonry
units ranging from 4.0 MPa for handmade clay bricks to 15.0 MPa for hollow clay tiles
(based on the gross area), while the minimum mortar compressive strength is 10.0 MPa
(corresponding to mortar mix 1:5 cement:sand).
A toothed interface between the walls and the tie-columns is used, as shown in Figure 3b.
Toothing enhances an interaction between masonry walls and RC confining elements
and prevents vertical cracking at the wall ends; this was confirmed during the field visit.
S388 ASTROZA ET AL.

Figure 3. Confined masonry construction: (a) Masonry wall is built first and (b) a toothed wall-
to-tie-column interface.

The construction sequence, a presence of toothing, size, and detailing of RC confining mem-
bers are the key differences between confined masonry and RC frame construction with
masonry infill walls, as shown in Figure 4 (EERI 2011).
RC confining members have an important role in enhancing the overall building stability
and integrity. These members can effectively contain damaged masonry walls in-plane and
out-of-plane, and ensure adequate wall connections to adjacent floors/roofs and foundations.
Based on the field observations, RC tie-columns are provided at 3 m to 3.5 m on-center
spacing, although NCh2123 (INN 2003b) permits a larger spacing up to 6m. Typical
cross-sectional dimensions range from 140 mm to 200 mm (length), while the depth is
equal to the wall thickness. The longitudinal reinforcement in RC tie-columns consists of
four 8 to 10 mm diameter bars, and the ties are typically provided at 150 mm to
200 mm spacing up the tie-column height. It was observed that uniform tie spacing was
provided in the surveyed buildings, although NCh2123 prescribes a closer spacing at the
tie-column ends. The tie size ranges from 6 mm diameter (for reinforcement cages assembled
at the site) to 4.2 mm (for prefabricated reinforcement cages). RC tie-beams have similar
reinforcement and dimensions like tie-columns. In buildings with timber roofs, the width
of tie-beams at the roof level exceeds the wall thickness, for example, 200 mm or
250 mm wide RC tie-beams are supported by 140 mm thick walls. Typical steel grades
used for the RC confining elements are A44-28H (280 MPa yield strength) and A63-
42H (420 MPa yield strength) for the reinforcement cages assembled at the site, while
AT56-50H grade (500 MPa yield strength) is used for the prefabricated reinforcement cages.

REINFORCED MASONRY CONSTRUCTION PRACTICE

Reinforced masonry construction is less common in the earthquake-affected area than
confined masonry. Reinforced masonry construction in Chile started in the 1970s, and
the behavior of medium-rise partially reinforced and hybrid masonry construction during
SEISMIC PERFORMANCE OF ENGINEERED MASONRY BUILDINGS IN THE 2010 MAULE EARTHQUAKE S389

Figure 4. A comparison: (a) Reinforced concrete frame construction with infill and (b) confined
masonry construction.

the 1985 earthquake was very poor, and many buildings experienced structural damage
(Wyllie et al. 1986, Astroza and Delfin 1986, Flores and Campusano 1986). It should be
noted that, at that time of the 1985 earthquake, Chilean codes did not have design provisions
for reinforced masonry buildings.
Reinforced masonry walls are built using hollow concrete or clay blocks, which are typi-
cally 140 mm thick. Vertical reinforcement (usually 8 mm to 12 mm diameter) is installed in
the hollow cores, which are subsequently filled with cement grout (in many cases, mortar was
used instead of grout for this purpose). Horizontal bars (usually 6 mm or 10 mm diameter) are
embedded in mortar bed joints; there are no bond beam blocks such as those used in North
American reinforced masonry construction. In some cases, the horizontal reinforcement is
prefabricated (ladder type) with cold-drawn wire of 4.2 mm diameter. Compressive strength
of masonry units is rather low (8 to 10 MPa), and mortar strength should be minimum
10 MPa (typical mix cement:lime:sand 1:1/2:4).
S390 ASTROZA ET AL.

DESIGN STANDARDS
Since the 1990s, seismic design of buildings in Chile has been regulated by the national
standard NCh433, with the following two editions: NCh433.Of93 (INN 1993) and
NCh433.Of96 (INN 1996). The 1993 edition of the standard prescribed a larger force reduc-
tion (ductility) factor for seismic design of masonry buildings using the modal analysis pro-
cedure, thus resulting in a significant reduction of shear forces (by approximately 30%)
compared to the values obtained by the 1996 edition. In the period from 1972 to 1993,
the base shear coefficient was equal to 0.12 g according to NCh433.Of 1972 (INN
1972). As a result, some Chilean buildings were designed for excessively low seismic
loads. This problem was corrected in the 1996 version of NCh433.
Design and construction of reinforced and confined masonry buildings in Chile has been
regulated by the national standards NCh1928.Of1986 (INN 2003a) and NCh2123.Of1997
(INN 2003b), with the latest editions published in 2003. Before the Chilean standards
were in place, masonry designers followed provisions of the U.S. Uniform Building
Code or the Chilean Ordenanza General de Edificación (1949).
Requirements of the Chilean masonry standards issued in 1980s were considered to be
too strict by some designers. As a result, a concept of “hybrid masonry construction” which
combines reinforced and partially confined masonry walls, emerged in the 1990s. A partially
confined masonry wall has a reinforced concrete column at one end and a vertical reinforcing
bar placed in a masonry block at the other end (usually at the openings). Reinforced masonry
walls typically have one 10-mm diameter vertical reinforcing bar at each end (note that a
minimum 12 mm bar diameter is required by the code) plus 8-mm diameter vertical bars
distributed along the wall section. The spacing between vertical bars usually exceeds the
permitted code limits, and the resulting horizontal steel ratio is below the minimum pre-
scribed by the code (0.06%). Hybrid construction is currently most widely used for one-
and two-story dwellings and medium-rise apartment buildings (up to four stories high).
Design provisions for this type of construction are not covered by the design codes; however
the Chilean Ministry of Housing has issued technical specifications for construction of one-
and two-story dwellings, which are also used for taller buildings.

DAMAGE OBSERVATIONS
According to the official data, 208,582 housing units in total required repair or replace-
ment after the earthquake (El Mercurio, 2010). Damage observations made in this paper are
based on field surveys of approximately 100 building complexes built in the 1990s located in
the Regions VI and VII, which account for approximately 15,000 housing units or 70% of the
total stock of social apartment buildings (three and four stories high) in the earthquake-
affected area, providing shelter to 90,000 people.
By and large, confined masonry buildings performed very well in the earthquake. Most
one- and two-story single-family confined masonry dwellings did not experience any
damage, with the exception of a few buildings which suffered moderate damage. A large
majority of three- and four-story confined masonry buildings also performed well, however
a few buildings suffered moderate to severe damage. A few medium-rise confined masonry
buildings of older vintages suffered a moderate repairable damage. For example, shear crack-
ing in masonry walls was observed in four-story buildings in Huemul II, Santiago (built in
SEISMIC PERFORMANCE OF ENGINEERED MASONRY BUILDINGS IN THE 2010 MAULE EARTHQUAKE S391

Figure 5. Moderately damaged medium-rise confined masonry buildings: (a) Shear cracks in the
walls at the third-story level (Huemul II, Santiago) and (b) wall damage at the bottom floor level
(Roto Chileno, Santiago).

1947); note that the buildings were damaged in the same manner in the 1985 Llolleo earth-
quake and subsequently repaired (Luego and Mass 2010). Similar damage patterns were
observed in the complex of four-story buildings called Roto Chileno (built in 1960), also
in Santiago, and Costanera Norte (built in 1956) in Talca. The walls in these buildings
were built using handmade solid clay bricks and low-strength cement/lime mortar
(cement:lime:sand mix ratio 1:1:6). In addition, RC tie-columns were provided only at
wall intersections and not at the openings, thus these buildings can be considered as “partially
confined.” Examples of moderately damaged confined masonry buildings are shown in
Figure 5.
For the first time in Chile, two three-story, partially confined masonry buildings col-
lapsed in an earthquake, killing ten people (six adults and four children). The following sec-
tions contain an overview of relevant observations related to the damage patterns in the key
structural elements, causes of building damage and collapse, and vulnerability evaluation for
engineered masonry buildings.

MASONRY WALLS: DAMAGE OBSERVATIONS AND DEFICIENCIES

The most common damage pattern observed in unreinforced masonry walls in confined
masonry buildings was in-plane shear cracking, which is characterized by diagonal
(X-shaped) cracks. In-plane shear cracking occurs when the masonry tensile strength is
exceeded due to the effect of combined shear and gravity loads. Alternatively, stair-stepped
shear cracks develop due to sliding shear at the mortar bed joints due to a loss of bond at the
mortar-to-unit interface. This type of damage was mostly observed at the bottom floor level
of three- and four-story buildings, as shown in Figure 6a. Shear cracking was observed in
walls built using all types of masonry units, including hollow clay tiles, clay bricks, and
hollow concrete blocks. More extensive damage was observed when masonry walls were
not confined at the openings (see Figure 6b); this confirms the findings of a previous Chilean
S392 ASTROZA ET AL.

Figure 6. In-plane shear cracking of masonry walls: (a) Damage at the ground floor level of a
three-story building in Cauquenes and (b) shear cracking of an unconfined masonry pier (note a
RC tie-column located at the center of the pier highlighted with an ellipse).

experimental study (Yañez et al. 2004). Absence of RC tie-columns at openings is believed to

be one of the main causes of severe damage in confined masonry buildings.
A few instances of wall damage due to out-of-plane seismic effects were observed. One
notable example was a three-story building in Cauquenes with the damage concentrated at
upper floors, as shown in Figure 7a. The building had RC floors and a timber truss roof.
Cracks in the third-floor wall panel extended into the RC tie-beam, as shown in Figure 7b
(note that the tie-beam is wider than the wall). The out-of-plane damage was observed in the
transverse direction of the building (note that the building was severely damaged in long-
itudinal direction due to in-plane seismic effects).
Poor quality of masonry materials and construction is believed to have influenced seismic
performance of some surveyed buildings. An example of low-strength concrete blocks and
reinforcement placed in an ungrouted cell is shown in Figure 8a. Concrete block units of low

Figure 7. Out-of-plane damage of confined masonry buildings: (a) Wall damage at the top floor
level and (b) damage in the RC perimeter beam.
SEISMIC PERFORMANCE OF ENGINEERED MASONRY BUILDINGS IN THE 2010 MAULE EARTHQUAKE S393

Figure 8. Substandard quality of masonry construction: (a) Low-strength concrete blocks and
(b) use of handmade bricks and excessive thickness of mortar bed joints.

compressive strength also have corresponding low shear strength. In some instances, exces-
sively thick mortar bed joints (on the order of 30 mm) were observed in brick masonry walls,
see Figure 8b. Handmade clay bricks used for wall construction in some damaged buildings
are characterized by a substantially lower strength compared to machine-made bricks. The
compressive and shear strength of handmade masonry units are typically very low. The qual-
ity of mortar used for masonry construction in Chile has been a recurring problem for many
years. Evidence of this problem is observed in stair-stepped diagonal cracking and the hor-
izontal cracking along the mortar bed joints. Other pieces of evidence are the moisture-related
problems due to water infiltration through wall mortar joints; these problems were reported in
a few damaged buildings before the earthquake, and some sealing products applied on the
exterior wall surface were observed during the survey.
A substandard quality of masonry construction was observed in some severely damaged
buildings. A notable example is the Villa Cordillera building complex in Rancagua (at about
260 km distance from the epicenter), where the maximum observed seismic intensity on the
MSK-64 scale (IMSK; Medvedev and Sponheuer 1969) was moderate (VI to VII; Astroza et al.
2012). For example, approximately 50% of reinforced or confined masonry buildings were
expected to suffer a minor damage in an area with the MSK intensity VII. The complex
consists of four- and five-story reinforced masonry buildings built in 1993 using
140 mm thick hollow concrete blocks. Vertical reinforcement was provided in the hollow
cells, however either partial or total absence of grouting was observed in several instances in
damaged walls; this resulted in corrosion or buckling of vertical reinforcement, as shown in
Figure 9a. Horizontal reinforcing bars (8 mm to 10 mm) were nearly the same size as the
mortar bed joint thickness (typically 10 mm to 12 mm), which will result in bond or anchor-
age failure. In-plane shear cracking was a typical damage pattern observed in the walls at the
ground floor level in these buildings. An example of a wall with horizontal reinforcement
provided at the sill level is shown in Figure 9b. Note a stair-stepped shear cracking in the pier,
which developed in spite of 8 mm horizontal reinforcing bars provided at the pier, mid-
height. The buildings of this complex were declared uninhabitable and were scheduled
for the demolition (Nuñez 2010).
S394 ASTROZA ET AL.

Figure 9. Damage of reinforced masonry concrete block walls: (a) A partially grouted wall
showing buckled vertical reinforcement and (b) horizontal reinforcement embedded in mortar
bed joints was ineffective in enhancing the masonry shear strength.

REINFORCED CONCRETE CONFINING ELEMENTS:

DAMAGE OBSERVATIONS AND DEFICIENCIES
Failure of RC tie-columns was observed in three- and four-story apartment buildings in
which masonry walls suffered severe damage at the ground floor level. The size of tie-
columns (140 mm × 140 mm or 150 mm × 150 mm) in a large majority of surveyed buildings
was smaller than the minimum dimensions prescribed by NCh2123, that is, the length equal
to 200 mm and the width equal to the wall thickness (t). Exposed RC confining elements
(tie-beams and tie-columns) in damaged buildings provided an opportunity to examine detail-
ing of reinforcement in these elements. A few important deficiencies which contributed to a
brittle failure of confined masonry walls are discussed below.
Inadequate confinement (tie size and spacing) at the ends of RC tie-columns was
observed in all damaged buildings. NCh2123 (INN 2003b) and other design guidelines
(EERI 2011) prescribe a closer tie spacing of 100 mm in end regions compared to the middle
portion of a tie-column (200 mm), as shown in Figure 10a. The use of prefabricated rein-
forcement cages with uniform tie spacing along the tie-column height contributes to this
deficiency; in addition, the tie diameter in prefabricated reinforcement cages is 4.2 mm,
as opposed to 6 mm diameter for the ties assembled at the site. Buckling of longitudinal
reinforcement was observed where size or spacing of ties at tie-column ends was inadequate,
as shown in Figure 10b, and the masonry crushing was severe, as shown in Figure 10c. The
buckling took place due to excessive combined compression and shear stresses developed in
RC tie-columns after the masonry was crushed and disintegrated at the wall toe.
Inadequate detailing of reinforcement in tie-column-to-tie-beam joint zones was observed
in all cases where the joints were exposed. An absence of ties in end columns is a deficiency,
and it causes a severe damage in the joint zone, as shown in Figure 10d. A good practice would
be to provide additional ties, as shown in Figure 10e. Another deficiency is related to a lack of
continuity in longitudinal reinforcement through tie-beam intersections with end columns, as
shown in Figure 10f. This is contrary to Chilean detailing provisions for reinforced concrete
structures (ICH 2009), which state that “longitudinal bars should have a 90º hooked anchorage
at intersections to ensure the effectiveness of tie-beams to resist earthquake load,” as shown in
SEISMIC PERFORMANCE OF ENGINEERED MASONRY BUILDINGS IN THE 2010 MAULE EARTHQUAKE S395

Figure 10. Construction and detailing deficiencies in the RC tie-columns: (a) Recommended tie
spacing in tie-columns (EERI 2011); (b) buckling of longitudinal bars; (c) severe crushing at the
wall toe; (d) inadequate reinforcement detailing in the joint region; (e) additional confinement for
vertical reinforcement in the tie-beam and tie-column end joint region (EERI 2011); (f) discon-
tinuous longitudinal reinforcement; (g) recommended detailing practice, and (h) detail showing
“continuity reinforcement” at intersections, length dimensions in cm (plan view).
S396 ASTROZA ET AL.

Figure 11. Poor quality of concrete construction: (a) In-plane shear cracking in the wall extend-
ing into RC tie-columns and (b) excessively large aggregate size.

Figure 10g. It should be noted that reinforcement cages for tie-beams and tie-columns are often
assembled off the building site, therefore additional “continuity reinforcement” should be pro-
vided once the cages are placed in the final position, as shown in Figure 10h.
The quality of concrete construction in RC confining elements was found to be good,
with a few exceptions. Excessively large aggregate size was observed in a few damaged RC
tie-columns; this can result in inadequate consolidation and large voids in these critical ele-
ments. Inconsistent bar sizes was another observed problem. The O’Higgins 859 Building in
Constitución is an illustrative example of concrete construction deficiencies (see Figure 11).

COLLAPSED BUILDINGS: A CASE STUDY

General features of medium-rise confined masonry buildings in the earthquake-affected
area were described in the previous section. Both collapsed buildings discussed in the paper
were medium-rise apartment buildings built in the last 20 years, but it is not known to what
extent the seismic design provisions of the Chilean design standard NCh433 have been fol-
lowed in the design and construction of these buildings. For medium-rise buildings of this
type, the fundamental period typically ranges from 0.10 sec to 0.16 sec, and the seismic
weight per unit floor area (including dead load plus 25% live load) is on the order of
650-700 kg/m2. The design seismic coefficient ranges from 0.17 g to 0.29 g according to
NCh433 seismic zone (2 or 3) and NCh433 type of soil (II or III).
One of the collapsed buildings was located in Santa Cruz (about 200 km from the epi-
center), where the maximum estimated seismic intensity on the MSK-64 scale (I MSK ) was in
the range from VII to VIII (out of maximum intensity IX) reported in the earthquake-affected
area (Astroza et al. 2012). The closest accelerometer station is Hualañe (50 km to the south),
and the maximum recorded horizontal ground accelerations were 0.389 and 0.461g
(Boroschek et al. 2010). The collapsed building was a part of the complex consisting of
28 identical partially confined masonry buildings, built in 1993, with four apartments per
floor, Figures 12a and b. The southwest part of the complex was adjacent to a farm (left
side in Figure 12a), and the damage distribution was higher in the buildings located in
that zone. The buildings had a rectangular plan. Exterior walls were built using partially
grouted hollow concrete blocks. Interior walls were built using handmade solid clay bricks
SEISMIC PERFORMANCE OF ENGINEERED MASONRY BUILDINGS IN THE 2010 MAULE EARTHQUAKE S397

Figure 12. Collapsed three-story confined masonry building in Santa Cruz: (a) Aerial view of the
building complex; (b) a row of buildings; (c) the collapsed building; (d) stair-stepped cracking
and masonry disintegration; (e) improper detailing of tie-beam-to-tie-column joint; and (f) shear
failure of the tie-column (note absence of ties).

and had excessively thick mortar bed joints (approximately 30 mm). Wall thickness was
150 mm, and the size of RC tie-columns was 150 mm by 150 mm. The collapsed building
lost its ground floor, as shown in Figure 12c.
Typical damage patterns in this building complex included diagonal and horizontal
cracking and crushing of masonry units in the walls, and severe damage of RC tie-columns,
as shown in Figure 12d. The average damage in the building complex was rated as grade 2
and 3 in the longitudinal and transverse direction (Jaramillo 2010), respectively, according to
the MSK damage classification for masonry buildings that was adapted to the Chilean build-
ings by Monge and Astroza (1989), as indicated in Table 1. The presence of stair-stepped
S398 ASTROZA ET AL.

Table 1. Damage categories (Monge and Astroza 1989)

Grade Damage Level Description

0 No damage No damage
1 Slight Fine cracks in plaster, falling of small pieces of plaster
2 Moderate Fine cracks in walls; vertical cracks at wall intersections; horizontal
cracks in chimneys, parapets and gables; falling of fairly large pieces of
plaster; falling of parts of chimneys; sliding of roof tiles
3 Heavy Large and deep diagonal cracks in most walls; large and deep cracks at
wall intersections; some walls lean out-of-plumb; falling of chimneys,
parapets and gable walls; falling of roof tiles
4 Very heavy Partial or total collapse of a wall in the building; collapse of building
partitions
5 Collapse Collapse of two or more walls in the building

cracking in the walls indicated a low shear bond strength along mortar bed joints, as seen in
Figure 12d.
All buildings in the complex were demolished within a few months after the earthquake.
Several factors influenced the seismic performance of the buildings and contributed to the
collapse, including inadequate wall density per unit floor (0.5% in the longitudinal direction,
0.3% in the transverse direction), poor soil conditions, low-strength masonry walls, absence
of grout in the hollow blocks resulting in a very small effective (net) wall area, absence of RC
tie-columns at the openings, inadequate size of RC tie-columns, and inadequate detailing of
reinforcement at tie-beam-to- tie column connection, as shown in Figures 12e and f.
The other collapsed building was located in Constitución, a city located 100 km from the
epicenter and affected both by the earthquake and the subsequent tsunami. The maximum
observed seismic intensity on the MSK scale was IX, that is, significantly higher than intensity
of VII to VIII assigned to Santa Cruz, where the other collapsed building was located (Astroza
et al. 2012). Maximum horizontal ground accelerations recorded at a nearby station located in
the Constitución hospital were 0.55 g and 0.64 g (Boroschek et al. 2010). The collapsed build-
ing was a part of a complex which included three rows of buildings (A, B, and C) built on the
hill, close to a steep slope on the west and north side, as shown in Figure 13a. The building in
the row C located closest to the slope (5 m distance on the west and north side) collapsed, while
the buildings in rows A and B experienced various extent of damage. It appears that several
owner-driven renovations took place in the collapsed building before the earthquake; for exam-
ple, a door opening was made in a load-bearing masonry wall at the second floor level of the
collapsed building shown in Figure 13b (note that the first floor collapsed). In addition, some
construction quality issues were reported before the earthquake (Castro 2010).
The walls were constructed using 140 mm thick hollow clay block (clay tile) units. A few
different sizes of RC tie-columns were observed depending on the location; length ranged
from 140 mm to 200 mm, and depth was equal to the wall thickness. In addition, a few wide
RC columns which were 500 mm to 900 mm long and 140 mm wide, and lightly reinforced
with nominal vertical and horizontal reinforcement, were also found in these buildings. These
SEISMIC PERFORMANCE OF ENGINEERED MASONRY BUILDINGS IN THE 2010 MAULE EARTHQUAKE S399

Figure 13. Collapsed building in Constitución: (a) Aerial view showing building location;
(b) door opening at the second floor level; (c) the collapsed building shifted by approximately
1.5 m away from the plinth; (d) in-plane shear failure of transverse walls; (e) out-of-plane crack-
ing in the longitudinal walls; and (f) absence of ties in the joint region.

wide columns replaced tie-columns at some locations (this is a common construction practice
for medium-rise confined masonry buildings in Chile). It has been reported that a seismic
coefficient of 0.15 g was used for the design.
The collapsed building shifted by approximately 1.5 m in the north direction (toward the
slope), as shown in Figure 13c. Damage in all buildings was more pronounced in the north–
south direction (transverse direction of the building plan). The collapse was initiated by the
shear failure of the transverse walls with severe damage at the ends of RC tie-columns, as
shown in Figure 13d, which was followed by the out-of-plane failure in the longitudinal walls
S400 ASTROZA ET AL.

(horizontal cracks due to out-of-plane seismic effects were observed in the longitudinal wall
of buildings of row B, Figure 13e). It is believed that the building location and geotechnical
effects were the key factors contributing to the collapse; note that an identical building com-
plex in another part of the city remained undamaged (Castro 2010). Additional factors con-
tributing to the collapse were the excessively low wall density per unit floor (0.7% in the
transverse direction), the small size of RC tie-columns (150 mm × 140 mm), the crushing of
the hollow clay tiles due to high percentage of perforations thus resulting in low masonry
compressive strength, an inadequate amount of transverse reinforcement at the ends of RC
tie-columns, and the absence of ties in the joint region (see Figure 13f).

WALL DENSITY AND VULNERABILITY ASSESSMENT

Wall density index is a ratio of cross-sectional areas for all walls in one orthogonal direc-
tion and the total floor area of the building. The required wall density for a particular building
depends on the seismic hazard and type of soil at the building site, plus masonry shear
strength and the number of stories. Wall density can be used as a design tool to ensure seismic
safety of new buildings (EERI 2011), or as an indicator of seismic vulnerability for existing
buildings (Meli 1994, Gallegos and Rios 1986).
Meli (1994) proposed an index (Wall Density Index), which uses a wall density ratio per
unit floor, d n , and it takes into account the wall height/length aspect ratio by reducing con-
tribution of walls with the higher aspect ratio. This index was used in the past to estimate the
damage to masonry structures during the 1985 Mexico earthquake and the 1985 Chile earth-
quake (Moroni et al. 2004).
Gallegos and Rios (1986) developed another index (Gallegos Index), which considered
characteristics of a building, such as stiffness distribution, plan shape, type of soil, building
condition, among others. Each of these parameters can be qualified as good, acceptable, or
bad for a specific building design, and a numerical value can be assigned accordingly. For
example, a building with an interior wall that is much stiffer than the exterior walls would be
assigned a bad rating for the stiffness distribution parameter. Gallegos Index (I g ) can be
determined by combining the values assigned to various parameters. A building can be con-
sidered safe in terms of the expected seismic performance if I g is 0.45 or higher. When a
building is characterized by the index of 0.3 and below, it is considered to be unsafe and the
design needs to be substantially modified.
A representative number of masonry buildings surveyed after the 2010 earthquake were
evaluated using both the Wall Density Index and the Gallegos Index. The effect of the type of
masonry unit on masonry shear strength was taken into account by reducing the wall area by a
factor of 0.4 or 0.5 when hollow concrete or handmade clay bricks were used, respectively.
Also, wide RC columns (if present) were not considered in the wall density calculations, as
these elements become effective only after the crushing of masonry has taken place.
Figure 14 illustrates the effectiveness of these two indices in correlating the building
damage level and the MSK earthquake intensity estimated at the building location. A cor-
relation of the Wall Density Index values shown on the vertical axis and the maximum build-
ing damage for a building complex on the horizontal axis is shown in Figure 14a. (Note that
only the building direction resulting in the lower value has been considered.) It can be seen
from the figure that buildings characterized by similar d n values suffered damage of different
SEISMIC PERFORMANCE OF ENGINEERED MASONRY BUILDINGS IN THE 2010 MAULE EARTHQUAKE S401

Figure 14. Vulnerability assessment of surveyed buildings: (a) Wall Density Index (Meli Index)
and (b) Gallegos Index.

extent/grade (ranging from 1 to 5), depending on their location and soil conditions (which are
related to the MSK intensity). Descriptions of damage grades are presented in Table 1. It can
be concluded that a wall density per unit floor index (d n ) of less than 0.75% results in an
unsatisfactory seismic performance, that is, damage is more extensive (and the grade is
higher) when MSK intensity is greater than VII. A building with d n value higher than
0.9% in both directions is considered to be safe even at the highest earthquake intensities.
However, a high wall density index in one direction only may not be sufficient to ensure
seismic safety of a building. For example, d n values for the collapsed confined masonry
building in Constitución were 1.4% and 0.7% in the longitudinal and transverse direction,
respectively. Note that the building collapsed after experiencing extensive damage in the
transverse direction and losing its bottom floor. Figure 14b shows values of the Gallegos
Index (I g ) on the vertical axis corresponding to different damage grades and earthquake
intensities. The “safe” limit value of 0.45 proposed by Gallegos and Rios (1986) appears
to be unsafe for masonry buildings. It can be seen from the figure that most of the surveyed
buildings were characterized by I g value of 0.4 or higher, and yet some of these buildings
suffered extensive damage (Grade 3 or higher according to Table 1).

GEOTECHNICAL ISSUES
Geotechnical effects have contributed to the collapse of a partially confined masonry
apartment building in Constitución (Figure 13) located in proximity of a steep slope, as dis-
cussed earlier in the paper. A localized influence of the unrestrained slope boundary and
localized variations in sub-surface strata might have generated localized variations of hor-
izontal (and possibly vertical) ground accelerations at the building site. Although Article
4.2.1 in NCh433.Of96 states that the soil parameters used to determine base shear are
given for sites located far away from geomorphologic or topographic singularities, this
was not considered in the design of the collapsed building.
Severe damage due to geotechnical issues was observed in a complex of several three-
story confined masonry buildings in Constitución called Francisco Mesa Seco. The complex
was built in 1992 and has 750 apartments in total. The buildings were built on a moderate
slope, in the proximity of a creek. All buildings, except for those located closest to the creek,
remained undamaged and were inhabited at the time of the visit, Figure 15a. The buildings
S402 ASTROZA ET AL.

Figure 15. Structural damage due to soil settlement in the Francisco Mesa Seco building com-
plex in Constitución: (a) A view of the buildings showing the direction of the slope; (b) damage in
exterior walls; (c) floor on grade settlement inside the building; and (d) explanation of the
observed damage pattern.

located closest to the creek were abandoned due to severe damage, caused by soil settlement
or downward movement of soil toward the creek. The damage was in the form of cracking in
the walls and floors at the ground floor level, as shown in Figures 15b and c. Note that
diagonal cracks in the walls developed only in one direction; this corresponds to the down-
ward soil movement toward the slope, and can be explained by the free-body diagram shown
in Figure 15d.

CONCLUSIONS AND RECOMMENDATIONS

The 27 February 2010 Maule earthquake exposed approximately 50% of the population
of Chile and a major portion of the building stock in the country to significant ground shak-
ing. It is estimated that only about 1% of the total building stock was damaged by the earth-
quake. This is considered to be a very good performance record for all building types, subject
to an earthquake of significant magnitude (Mw 8.8), duration, and high ground accelerations.
A large majority of engineered masonry buildings performed well in the earthquake. Low-
rise confined masonry buildings either remained undamaged or suffered a minor damage.
The same statement can be made for medium-rise buildings, with an exception of a few
severely damaged buildings and two collapsed buildings.
SEISMIC PERFORMANCE OF ENGINEERED MASONRY BUILDINGS IN THE 2010 MAULE EARTHQUAKE S403

Lessons learned from detailed studies on the damaged and collapsed buildings and obser-
vations related to seismic performance of engineered masonry buildings, and particularly
confined masonry construction, are valuable for evaluating current design provisions and
improving construction practices for confined masonry buildings in Chile and other coun-
tries. Key lessons and recommendations related to the performance of engineered masonry
construction in the 2010 Maule earthquake can be identified as follows:
1. Wall density is an important parameter that can be used to assess seismic
vulnerability of confined masonry buildings. The results of a seismic vulnerabil-
ity study on confined masonry buildings affected by the earthquake indicate that a
wall density per unit floor index (d n ) of less than 0.75% results in unsatisfactory
seismic performance when MSK intensity is greater than VII. A building with
d n value of more than 0.9% in both directions is considered to be safe even at
the highest earthquake intensities.
2. Closer spacing of ties in end zones of RC tie-columns is critical for maintaining
their stability and strength. Inadequate confinement (tie size and spacing) at the
ends of RC tie-columns was observed in all damaged buildings. NCh2123 (INN
2003b) and other design guidelines (EERI 2011) prescribe a closer tie spacing
of 100 mm for the end regions compared to the middle portion of a tie-column
(200 mm). Adequate bar size and tie spacing in the end zones of RC tie-columns
is critical for preventing shear failure and buckling of longitudinal reinforcement in
localized areas of tie-columns where masonry is completely disintegrated.
3. RC confining elements, especially tie-columns, must have adequate size to
prevent a loss of compression resistance in medium-rise confined masonry
buildings. The size of tie-columns (140 mm × 140 mm or 150 mm × 150 mm)
in a large majority of surveyed medium-rise buildings (both damaged and unda-
maged) was smaller than the dimensions prescribed by NCh2123 (minimum tie-
column size: 200 mm × t, where t denotes the wall thickness). Results of recent
experimental studies indicate that the masonry walls confined with RC tie-columns
with excessively small cross-sectional dimensions suffer a severe deterioration of
ductility and energy dissipation capacity (San Bartolome et al. 2010a). In order to
prevent the brittle failure of confined masonry wall panels and maintain integrity of
RC tie-columns after masonry walls suffer a severe damage, tie-column dimensions
must meet the minimum NCh2123 requirements.
4. Presence of RC tie-columns at door and window openings is critical for ensuring
confining effect in the masonry walls. Without the confinement provided by RC tie-
columns, it is not possible to develop compressive struts in masonry wall panels; this
is the key mechanism for lateral load resistance in confined masonry walls. NCh2123
(INN 2003b) and other design guidelines require jamb reinforcement at openings
(EERI 2011).
5. Ungrouted masonry units with excessive perforations should not be permitted
for use in confined masonry construction in areas with high seismic hazard.
Severe crushing of masonry occurred due to the large percentage of voids and the
small face shell and web thickness in hollow clay tiles. This brittle failure of masonry
units in critical wall regions (e.g., toes) should be prevented. To ensure satisfactory
seismic performance of walls in confined masonry buildings, the ratio of net to gross
S404 ASTROZA ET AL.

unit cross-sectional area should be greater than 70% (San Bartolome et al. 2010b),
minimum face shell and web thickness should be prescribed, and the use of masonry
units with horizontal perforations should not be permitted by the code.
6. Poor quality of masonry materials and construction has negatively affected
seismic performance of confined masonry buildings. In severely damaged build-
ings, it was observed that substandard quality of concrete blocks was used for con-
struction. In other cases, use of low-strength handmade bricks with excessively thick
mortar bed joints was observed. The quality of mortar used for masonry construction
in Chile has been a recurrent problem for many years.
7. Reinforced masonry construction must be inspected to ensure proper grouting.
Vertical reinforcement was provided in the hollow cores, however either partial or
total absence of grouting was observed in several instances. Reinforced cores in
concrete block walls must be grouted, in order to avoid the crushing of masonry
and buckling of the vertical reinforcement. Reinforced masonry construction
must be inspected at the grouting stage. The option “Construction without technical
inspection” must be eliminated from the NCh1928 code for this type of construction
(INN 2003a).
8. Geotechnical issues have caused severe damage in a few confined masonry
buildings. One of the collapsed confined masonry buildings was located close to
a very steep slope. A localized influence of the unrestrained slope boundary and loca-
lized variations in sub-surface strata might have generated localized variations of hor-
izontal and vertical ground accelerations at the building site. However, adequate soil
tiebacks and building location would have prevented the collapse. In another Con-
stitución building complex (Francisco Mesa), severe structural damage due to soil
settlement was observed in buildings located in the proximity of a creek. In both
cases, otherwise similar buildings within the same building complex experienced
minor to moderate damage. Regulations related to the construction of buildings in
proximity of slopes in area of high seismic hazard should be enforced in Chile.

ACKNOWLEDGMENTS
The field trip was financially supported by the Universidad de Chile, the Earthquake
Engineering Research Institute, and the Masonry Society. The authors would like to
acknowledge significant contribution of the former graduate students at Universidad de
Chile: Carolina Jaramillo, Felipe Castro, and Manuel Nuñez.

REFERENCES
Astroza, M., and Delfin, F., 1986. Capitulo 5: Construcciones de Albañilería, in El Sismo del 3 de
Marzo 1985, Chile, Facultad de Ciencias Físicas y Matemáticas, Universidad de Chile,
Santiago, Chile, 93–110 (in Spanish).
Astroza, M., Ruiz, S., and Astroza, R., 2012. Damage assessment and seismic intensity analysis
of the 2010 (Mw 8.8) Maule earthquake, Earthquake Spectra 28, this issue.
Boroschek, R., Soto, P., and Leon, R., 2010. Registros del Terremoto del Maule Mw =8.8 de 27 de
Febrero de 2010, Informe RENADIC 10/05 Rev. 1, Departamento de Ingeniería Civil,
Facultad de Ciencias Físicas y Matemáticas, Universidad de Chile (in Spanish).
SEISMIC PERFORMANCE OF ENGINEERED MASONRY BUILDINGS IN THE 2010 MAULE EARTHQUAKE S405

Castro, F., 2010. Estudio de los Efectos del Terremoto del 27 de Febrero del 2010 en las Vivien-
das Sociales de Constitución, Civil Engineering Thesis, Universidad de Chile (in Spanish).
Earthquake Engineering Research Institute (EERI), 2011. Seismic Design Guide for Low-Rise
Confined Masonry Buildings, Earthquake Engineering Research Institute, Oakland, CA,
available at http://www.confinedmasonry.org.
El Mercurio, 2010. Noviembre 13, Cuerpo C, p. C10.
Flores, R., and Campusano, R., 1986. Daños de estructuras de albañilería en viviendas económ-
icas a raiz del sismo del 3 de Marzo de 1985, 4as. Jornadas Chilenas de Sismología e Inge-
niería Antisísmica and International Seminar on the Chilean March 3, 1985 Earthquake,
Tomo 2, Viña del Mar, Chile, Abril, C43–C62 (in Spanish).
Gallegos, H., and Rios, R., 1986. Indice de calidad estructural sismo resistente, 4as. Jornadas
Chilenas de Sismología e Ingeniería Antisísmica and International Seminar on the Chilean
March 3, 1985 Earthquake, Tomo 2, Viña del Mar, Chile, Abril, I37–I61 (in Spanish).
Instituto del Cemento y del Hormigón de Chile (ICH), 2009. Manual de Detallamiento para
Elementos de Hormigón Armado. Instituto del Cemento y del Hormigón de Chile (in Spanish).
Instituto Nacional de Estadísticas (INE), 2002. XVII Censo Nacional de Población y VI de
Vivienda, Instituto Nacional de Estadísticas, Santiago, Chile, available at http://www.ine.cl
(in Spanish).
Instituto Nacional de Normalización (INN), 1972. NCh433.Of1972: Cáculo antisismico de
edificios, Norma Chilena Oficial, Instituto Nacional de Normalización, Santiago, Chile (in
Spanish).
Instituto Nacional de Normalización (INN), 1993. NCH433.Of1993: Diseño sísmico de edificios,
Norma Chilena Oficial, Instituto Nacional de Normalización, Santiago, Chile (in Spanish).
Instituto Nacional de Normalización (INN), 1996. NCH433.Of1996: Diseño sísmico de edificios,
Norma Chilena Oficial, Instituto Nacional de Normalización, Santiago, Chile (in Spanish).
Instituto Nacional de Normalización (INN), 2003a. NCh1928.Of1993. Modificada en 2003:
Albañilería armada-Requisitos para el diseño y cálculo, Norma Chilena Oficial, Instituto
Nacional de Normalización, Santiago, Chile (in Spanish).
Instituto Nacional de Normalización (INN), 2003b. NCh2123.Of1997. Modificada en 2003:
Albañilería Confinada-Requisitos para el diseño y cálculo, Norma Chilena Oficial, Instituto
Nacional de Normalización, Santiago, Chile (in Spanish).
Jaramillo, C., 2010. Estudio de los efectos del terremoto del 27 de Febrero de 2010 en las vivien-
das sociales de la VI Región, Civil Engineering Thesis, Universidad de Chile (In Spanish).
Luego, O., and Mass, E., 2010. La “X” del Desalojo en Población Huemul II, available at
http://www.docomomo.cl (in Spanish).
Medvedev, S. H., and Sponheuer, W., 1969. Scale of seismic intensity, IV World Conference on
Earthquake Engineering, Santiago, Chile, 143–153.
Meli, R., 1994. Structural design of masonry buildings: the Mexican practice, in Masonry in the
Americas, ACI Publication SP-147, American Concrete Institute, Detroit, pp. 239–262.
Monge, J., and Astroza, M., 1989. Metodología para determinar el grado de intensidad a partir de
los daños, 5as. Jornadas Chilenas de Sismología e Ingeniería Antisísmica, Santiago, Chile,
483–492 (in Spanish).
Moroni, M. O., Astroza, M., and Acevedo, C., 2004. Performance and seismic vulnerability of
masonry housing types in Chile, Journal of Performance of Constructed Facilities, ASCE, 18,
173–179.
S406 ASTROZA ET AL.

Nuñez, M., 2010. Análisis de los Daños Provocados por el Terremoto del 27 de Febrero de 2010 a
los Edificios de Villa Cordillera, Comuna de Rancagua, Civil Engineering Thesis, Universidad
de Chile (in Spanish).
Ordenanza General de Edificación, 1949. Reglamento sobre Viviendas Económicas, Ediciones
Gutemberg, Santiago, Chile.
San Bartolome, A., Bernardo, J., and Peña, M., 2010a. Efectos del peralte de las columnas en el
comportamiento sísmico de los muros de albañilería confinada, in Congreso Chileno de
Sismología e Ingeniería Antisísmica, X Jornadas, Santiago, Chile (in Spanish).
San Bartolome, A., Pareds, J., and Caycho, D., 2010b. Control de la trituración de los ladrillos
huecos en muros de albañilería confinada sujetos a carga lateral cíclica, Investigación en
Albañilería, Pontificia Universidad Católica del Perú, available at http://blog.pucp.edu.pe/
media/688/20090114-Trituracion%20de%20ladrillos%20huecos.pdf (in Spanish).
Wyllie, L. A., Bolt, B., Durkin, M. E., Gates, J. H., McCornick, D., Smith, P. D., Abrahamson,
N., Castro, G., Escalante, L., Luft, R., Olson, R. S., and Vallenas, J., 1986. The Chile
earthquake of March 3, 1985, Section 4: Performance of structures, Earthquake Spectra 2,
293–373.
Yañez, F., Astroza, M., Holmberg, A., and Ogaz, O., 2004. Behavior of confined masonry shear
walls with large openings, in Proceedings of the 13th World Conference on Earthquake
Engineering, Vancouver, Canada, Paper 3438.
(Received 11 March 2011; accepted 9 February 2012)

View publication stats