Professional Documents
Culture Documents
Paper Mexico 2017
Paper Mexico 2017
net/publication/344460657
CITATIONS READS
0 124
5 authors, including:
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by Sergio Alcocer on 08 October 2020.
Earthquake Spectra
Seismic rehabilitation of 1–24
Ó The Author(s) 2020
concrete buildings after the Article reuse guidelines:
sagepub.com/journals-permissions
Mexico City
Abstract
The 19 September 2017 Puebla–Morelos earthquake provides a unique opportunity
to (a) study the effectiveness of seismic rehabilitation methods implemented in
Mexico City after 1985 and (b) collect large amounts of data on current building
rehabilitation practices. This article presents and discusses seismic rehabilitation
efforts conducted on concrete buildings after the 1985 and 2017 earthquakes.
Building damage and vulnerabilities, code provisions for rehabilitation, and common
repair and retrofit techniques used in these two periods of time are summarized. The
performance of four buildings rehabilitated after 1985 is examined based on results
from field investigations conducted after the 2017 event. These rehabilitations were
effective in improving the performance of these buildings according to observed dam-
ages and estimated shaking intensities. To allow a more systematic assessment of seis-
mic rehabilitations in future earthquakes, an inventory of rehabilitated buildings in
Mexico City is currently under development. Preliminary data obtained from this
inventory effort are presented and discussed.
Keywords
Databases, earthquake damage, repairs, strengthening, reinforced concrete
structures
Date received: 7 August 2020; accepted: 10 August 2020
1
Department of Civil, Architectural and Environmental Engineering, The University of Texas at Austin, Austin, TX, USA
2
Instituto de Ingenierı́a, Universidad Nacional Autónoma de México, Mexico City, Mexico
3
Iktes, Mexico City, Mexico
Corresponding author:
Juan Murcia-Delso, The University of Texas at Austin, 301E Dean Keeton St C1700, Austin, TX 78712, USA.
Email: murcia@utexas.edu
2 Earthquake Spectra 00(0)
Introduction
The 1985 Mexico City earthquake caused thousands of deaths and resulted in extensive
damage and collapse of a large number of concrete buildings. In the months and years fol-
lowing this earthquake, there was an unprecedented effort to repair and retrofit concrete
buildings. Interventions ranged from local upgrades of concrete elements (e.g. jacketing of
columns and beams) to changes in the lateral load resisting system (LLRS) (e.g. steel bra-
cing and addition of shear walls). These retrofit methods were developed based on engi-
neering judgment using strategies that in most cases had not been previously validated
with experimental studies or refined numerical modeling.
The 19 September 2017 earthquake caused 369 deaths, 228 of them in Mexico City, and
damaged thousands of buildings. This earthquake represented the first severe test for many
concrete buildings that were retrofitted after 1985, providing a unique opportunity to col-
lect evidence and study the effectiveness of the applied seismic rehabilitation techniques. In
addition, it is estimated that more than 11,000 single-family dwellings and buildings need
different levels of repair and retrofit after the 2017 event, so that a new wave of rehabilita-
tion efforts is underway.
This article presents and discusses seismic rehabilitation efforts on concrete buildings
conducted in Mexico City after the 1985 and 2017 earthquakes. Common building damage
and rehabilitation methods used after 1985 are summarized, and four case studies of post-
1985 building rehabilitations are presented. The performance of these buildings is qualita-
tively assessed based on observed damage and estimated shaking intensities. Common
building damages observed after the 19 September 2017 earthquake are described along
with new code provisions for seismic rehabilitation. Finally, statistics and examples of
post-2017 rehabilitations are presented in the scope of an ongoing effort to develop a data-
base of rehabilitated buildings in Mexico City.
Figure 1. Elastic design spectra for type B buildings (office and residential buildings) for soil zones II and
III according to 1976 Mexico City Building Code (RDF-76) and 1985 emergency norms (EMG-85).
Federal (DDF), 1988) registered 1433 buildings with structural rehabilitation and 225
buildings were demolished. Seismic rehabilitations were conducted following the 1985
emergency norms and subsequently the 1987 building code.
Figure 2. Location of buildings studied, collapsed, and retrofitted buildings in a soil period map
(courtesy of MA Jaimes) and recording stations.
Figure 3. Acceleration response spectra for 19 September 1985 and 19 September 2017 events (5%
damping ratio): (a) SCT, (b) XP06, (c) CO56, (d) AL01, and (e) CJ04 (motion records obtained from
Centro de Instrumentación y Registro Sı́smico (CIRES) and Unidad de Instrumentación Sı́smica of the
National Autonomous University of Mexico (UNAM)).
To qualitatively assess the level of shaking intensity experienced by the buildings, their
fundamental periods have been estimated based on approximate formulas proposed by
Murià-Vila and González (1995). These formulas were obtained empirically based on
dynamic measurements taken on 60 buildings in Mexico City; these expressions account
for the effects of different soil conditions (longer periods are obtained in soft soils). For
soft soil conditions, T = 0:126 N for RC frame structures and T = 0:102 N for RC structures
with frames and walls, where N is the number of stories. Only three measurements corre-
spond to rehabilitated buildings with frames and walls, and their fundamental periods are
lower than the proposed average value: 0:057 N , 0:077 N , and 0:099 N . These differences
(up to 45% less) can be explained by the approximate character of the proposed formula
and the higher stiffness of retrofitted structures. Based on the uncertainty provided by the
approximate formulas and the multiple factors conditioning the actual dynamic properties
Murcia-Delso et al. 7
Figure 4. Building 1: (a) general view, (b) jacketing of beams and columns after 1985 (Aguilar et al.,
1996), (c) flexural cracking of beams in 2017, and (d) pounding with adjacent building in 2017.
Building 1
Building 1 is a 4-story warehouse with a basement located in the corner of a block, as
shown in Figure 4a. It was built in 1959 and its original LLRS consisted of RC frames in
two orthogonal directions. Unreinforced masonry (URM) infill walls were used in the peri-
meter of the building and stairway area. The building is supported on a mat foundation.
Damage after the 1985 earthquake and retrofit intervention. Building 1 was severely damaged
during the 1985 earthquake. Structural damage was mainly concentrated in the second-
floor columns, which experienced concrete spalling as well as buckling and/or fracture of
vertical bars (Aguilar et al., 1996; Jara et al., 1989). Column damage was attributed to
short column effects induced by infill walls, insufficient transverse reinforcement, or lap
splice failures. In addition, URM walls in the building perimeter and stairway were exten-
sively damaged in several floors. No foundation failures were observed, but the structure
was reported to lean 20 cm out of plumb. However, it was unclear if the building inclina-
tion was caused by the earthquake or a differential settlement related to the regional subsi-
dence constantly experienced by the lakebed soil of Mexico City.
8 Earthquake Spectra 00(0)
The building was repaired and strengthened after 1985. Columns were repaired by repla-
cing the segments of bars that had buckled or fractured by new bars that were fillet-welded
to the original bars. The strengthening scheme consisted of RC jacketing of all beams and
columns, as shown in Figure 4b. Concrete jackets with a minimum thickness of 12 cm and
additional longitudinal and transverse reinforcement were provided. Also, URM infills in
the south and west perimeter walls were separated from the structure to reduce torsional
effects.
Damage after the 2017 earthquake. Building 1 presented very minor damage after the 19
September 2017 earthquake. Structural damage was limited to cracking in some beams.
Flexural cracks were observed in several beams running in the two orthogonal directions
of the building in the first, second, and third floors. Most of these cracks are concentrated
near the beam ends, as shown in Figure 4c, and their widths were 0.5 mm or smaller. In
the first floor, the perimeter beam running in the E-W direction presented a 1 mm vertical
crack at the northwest corner of the building. This damage could be related to pounding
with an adjacent building, which was reported to occur in the northwest corner of Building
1 as manifested by the facade damage shown in Figure 4d. A few cracks were also observed
in some infill walls and at the interface between the infills and the frames.
Comparison of building responses during the 1985 and 2017 events. During the 1985 event,
Building 1 suffered major damage, but only minor cracking was observed in the retrofitted
building after the 2017 event. The improvement in the response of the structure can be
explained by the increase in the lateral strength of the jacketed members and reduction of
torsional effects, but the shaking intensity experienced by the building during these two
events needs also to be considered to establish a fair comparison.
The level of shaking intensity has been qualitatively assessed based on the spectral accel-
erations shown in Figure 3. The fundamental period of the building has been estimated in
the range of 0.25–0.75 s considering 650% deviations of the approximate period proposed
by Murià-Vila and González (1995) for RC frames on soft soils. As shown in Figure 3, the
1985 and 2017 response spectra of the SCT station are very similar to each other in the
0.25–0.75 s period range, and their spectral values vary between approximately 0.15 and
0.30 g. The 2017 response spectrum of the XP06 station (Figure 3b), which is closest to the
building (located at approximately 500 m), is very similar to the SCT spectrum and it fluc-
tuates between 0.15 and 0.25 g in this period range. Hence, it can be concluded that the
1985 and 2017 resulted in comparable levels of shaking intensity in the period range of
interest. Nevertheless, the retrofitted structure is expected to present a smaller period than
the original one owing to the higher stiffness of the enlarged frame elements. A decrease in
the period could have also slightly reduced the demands in the building according to the
spectra in Figure 3b. In any case, the spectral accelerations experienced in 2017 were lower
than the seismic coefficient of the elastic design response spectrum which had to be used in
the rehabilitation (between 0.2 and 0.4 g, depending on the period assumed in design).
Building 2
Building 2 is located in the lakebed zone of Mexico City, as shown in Figure 2. The origi-
nal building had 13 stories. Lateral load resistance was provided by RC shear walls in the
N-S direction, while the E-W direction had two perimeter RC frames composed by wide
U-shape columns connected through deep beams. The floor system is an RC slab
Murcia-Delso et al. 9
Figure 5. Building 2: (a) general view, (b) strengthening of columns (Aguilar et al., 1996), and (c) uplift
of south side after 2017.
supported on truss girders in the N-S direction which are supported on the perimeter
frames. The building sits on a foundation box supported on friction piles. Figure 5a shows
the present view of the building.
Damage after the 1985 earthquake and retrofit intervention. After the 1985 earthquake, the RC
deep beams running in the E-W direction of the building presented moderate to severe
cracking (Aguilar et al., 1996). All cracked structural elements were repaired by injecting
epoxy. In beams presenting severe damage, concrete was removed and a new concrete sec-
tion with new ties was cast. To reduce seismic forces, it was decided to reduce the mass of
the building by removing the four top floors. Moreover, all the U-shaped columns were
strengthened by filling the inner space and defining a rectangular cross section while pro-
viding additional reinforcement, as shown in Figure 5b. The beams in the E-W direction
were externally post-tensioned in the first five floors to increase their shear capacity. Post-
tensioning cables were anchored in new boundary elements of the perimeter shear walls
running in the N-S direction of the building (Figure 5b). The enlarged columns and wall
boundary elements were continued into the foundation box, and it was considered that
the existing foundation beams could transfer additional forces resulting from the strength-
ened columns to the pile caps.
Damage after the 2017 earthquake. No structural damage was observed after the 2017 earth-
quake. However, the building presented a significant inclination in the N-S (short) direc-
tion. An uplift of approximately 10 cm was measured on the south side of the building
with respect to the street level, as shown in Figure 5c. A visual inspection of the foundation
beams did not reveal damage in the foundation box. At the time of the visit, a geotechnical
10 Earthquake Spectra 00(0)
survey was being conducted on the soil next to the building to determine the cause of the
building inclination.
According to the building occupants and current engineering team attending the build-
ing, part of the building uplift had already occurred before the earthquake, and foundation
vertical displacement measurements were being taken since 1999. According to the displa-
cement data, the building emerged gradually from ground level until 2009, and more sig-
nificant foundation tilting was registered afterwards. Measurements taken in April 2019
indicate a differential settlement of 23 cm in the N-E corner with respect to the S-W cor-
ner, and a 49.7 cm relative horizontal displacement at the roof level.
The hypothesis of the engineering team attending the building is that, due to the con-
stant regional subsidence experienced by the lakebed soil of Mexico City, friction piles of
south facade started to interact with the hard rock layer before north facade ones.
Therefore, it is assumed that current building tilting is not directly related to 2017 earth-
quake, and consequently, sidewalk damages presented in Figure 5c could be caused by the
rearrangement of the stiff sidewalk concrete blocks over a previously consolidated soil.
Comparison of building responses during the 1985 and 2017 events. The structural performance
of the building during the 2017 event was improved with respect to 1985, despite tilting
problems of the building.
Shaking intensities at the building site have been compared using the same qualitative
approach used for Building 1. The original building fundamental period has been esti-
mated based on the formulation proposed by Murià-Vila and González (1995) previously
presented in this section. Based on this approximation, the estimated values of the funda-
mental periods were about 1.6 s in the E-W direction and 1.3 s in the N-S direction.
According to recent ambient vibrations measurements performed on building after 2017
earthquake, the current periods in the longitudinal (E-W) and transverse (N-S) directions
are 0.87 and 0.68 s, respectively. The spectral accelerations of the 2017 event for the CO56
station, which is located close to the building site (less than 500 m away), are shown in
Figure 3c. Accelerations between 0.2 and 0.3 g in the 0.5–2 s range are displayed, with a
peak of 0.45 g at around 0.8 s. According to spectral accelerations obtained for the 1985
and 2017 motions in the SCT station, the original structure could have had a significantly
higher demand for a fundamental period over 1.6 s. Therefore, it is likely that the shaking
intensities were slightly smaller during 2017, and the reduction of the mass also contribu-
ted to the decrease in the base shear in the building.
Building 3
Building 3 is located in the lakebed zone of Mexico City, and it is used for both office and
residential purposes (Figure 6). The building was designed and constructed in 1974, pre-
senting a basement, 6 stories, and a small penthouse. The original structural system con-
sisted of RC columns supporting waffle slab floors, with three regular spans between 4.2
and 4.3 m in the transverse direction. The foundation was a foundation box supported on
friction piles.
Damage after the 1985 earthquake and retrofit intervention. According to the records presented
in Jara et al. (1989) and Aguilar et al. (1996), the building did not present structural dam-
age, foundation problems, or pounding with adjacent buildings during the 1985
Murcia-Delso et al. 11
Figure 6. Building 3: (a) general view, (b) control piles placed after the original retrofit, (c) light damage
after 2017.
earthquake. Only minor cracking was presented in the perimeter masonry walls running in
the longitudinal direction of the building and in some partitions. The decision to retrofit
this building was preventive in order to upgrade the structure according to the 1985 emer-
gency norms (DDF, 1985).
Retrofitting in the transverse direction—parallel to the street—consisted of X-shaped
steel braces located in the central span of four different frames over the entire height of the
building. Steel jacketing was applied to the columns connected to the steel braces. In the
longitudinal direction, the majority of the perimeter masonry walls were first repaired by
injecting the cracks with a cement-based grout, and later strengthened with infill reinforced
shotcrete walls attached to the original masonry walls. RC boundary elements were also
added to these composite walls.
According to the available information, the after-1985 retrofit did not involve founda-
tion activities. Years later, new control piles were placed in the foundation with the aim of
controlling the building inclination caused by the regional subsidence (Figure 6b).
Damage after the 2017 earthquake. Like in the 1985 earthquake, Building 3 did not exhibit
relevant structural damage after the 2017 event. There were also no observations related to
significant changes in tilting or evidences of interaction with adjacent buildings. Only slight
diagonal cracks were observed on the south perimeter wall of the basement. Light damage
was noticeable at some interfaces between strengthened and original elements in the first 3
stories (starting at street level), mainly corresponding to minor cracking between frame ele-
ments and infill walls (Figure 6c).
12 Earthquake Spectra 00(0)
Comparison of building responses during the 1985 and 2017 events. As indicated above,
Building 3 presented an acceptable performance during both the 1985 and 2017 earth-
quakes. The seismic demands have been estimated based on the spectra shown in Figure
3d. Considering the simplified approach used in previous sections, the building fundamen-
tal period varied from 0.76 s in its original configuration to 0.46 s in the retrofitted
configuration. For the retrofitted structure, the period is estimated based on the average
period-to-number of levels ratio (0:077 N ) obtained by one of the authors from measure-
ments on four buildings retrofitted with steel bracings. Assuming 650% deviations with
respect to this approximate formula, the fundamental period of the retrofitted structure
should be located in the 0.23–0.69 s range, as shown in Figure 3d. According to the 1985
SCT-station spectra, a relatively moderate seismic demand, with a spectral acceleration
between 0.2 and 0.35 g, was expected for the estimated range of fundamental periods of
the original structure. The spectra obtained from the 2017 motion registered in the closest
station—less than 1 km away—(AL01, Figure 3d) present a relatively constant spectral
acceleration demand between 0.2 and 0.4 g for periods ranging approximately from 0.25
to 2 s, which includes the estimated range of periods of the retrofitted structure.
Therefore, the original and retrofitted buildings could have been subjected to similar seis-
mic demands, behaving in this case satisfactorily during both earthquakes.
Building 4
Building 4 consists of two apartment units separated by the stairs and elevator core
(Figure 7). The north block has 7 stories and the south block has 8 stories. The original
structure had waffle slabs supported on RC columns in the first and top levels and a one-
way precast beam-block floor system supported on masonry walls confined by rectangular
RC tie-columns in the rest of levels. The foundation was a grid and slab system supported
on friction piles.
Damage after the 1985 earthquake and retrofit intervention. After the 1985 event, damage was
observed on the masonry walls, especially in the short (E-W) direction of the building.
Confinement elements of these walls were also damaged, some presented cracking and loss
of concrete cover and other completely failed. Horizontal and vertical irregularities in the
building configuration were the main causes for damage (Aguilar et al., 1996).
The damaged masonry walls were repaired with wire mesh and shotcrete. The structure
was strengthened in the short direction of the building by adding four RC frames, as
shown in Figure 7b. To this end, the sections of existing beams and columns were enlarged
and additional reinforcement was provided. New perimeter beams were also added on line
A, between lines 4 and 5, to connect the two apartment blocks and to reduce torsional
effects. In addition, two 15-cm-thick RC walls around the stairs and elevator area were
constructed, as shown in Figure 7b. The foundation was not modified except at the con-
nection of the new elements.
Damage after the 2017 earthquake. The building exhibited distributed flexural cracking of
the columns in the first story, as depicted in Figure 7c. The column in Figure 7c is one of
the enlarged columns on line 2 of the building. The spandrel beam running in the short
(E-W) direction in the first floor also presented vertical and diagonal cracks, as shown in
Figure 7d, although some of these cracks seemed pre-existing. In the upper stories, diago-
nal shear cracks were also observed in masonry infill walls in the E-W direction.
Murcia-Delso et al. 13
Figure 7. Building 4: (a) general view, (b) floor plan (after Aguilar et al., 1996), (c) flexural cracking of
columns in 2017, and (d) cracking of spandrel beam in 2017.
Comparison of building responses during the 1985 and 2017 events. Based on the approximate
formula for RC frames in soft soil conditions and the 650% assumption, the fundamental
period of the building (original and retrofitted) is estimated to be between 0.5 and 1.5 s.
Given the enlargement of structural members and addition of frame and wall elements, the
period of the retrofitted structure should be smaller than that of the original one. As shown
in Figure 3a, both the 1985 and 2017 response spectra of the SCT station vary between
approximately 0.20 and 0.40 g in the 0.5–1.5 s period range, but the 1985 spectrum in the
E-W direction presents higher values between 0.7 and 1.2 s. The 2017 response spectrum of
the CJ04 station, the closest to the building (less than 1 km away), also presents spectral
values between 0.20 and 0.40 g. Based on these observations, it is likely that the shaking
intensity in 2017 was comparable to that of 1985.
In terms of performance, the building structure presented less critical damage than in
1985 thanks to the addition of frame elements and shear walls. However, significant crack-
ing was observed in structural and nonstructural elements in the short (E-W) direction of
the building. Hence, the 2017 event introduced significant deformations in the short direc-
tion of the building, which were probably coupled with torsional effects caused by the plan
irregularity of the structure.
students were assigned building survey in specific zones. Their purposes were to gather rel-
evant information to understand building behavior and causes of damage as well as to
help government officials on building tagging. Teams used two different data collection
formats. A simple format was used for external inspections; when internal surveys were
carried out, a more detailed damage collection form was used.
In the first weeks after the earthquake, a total of 1954 buildings had been surveyed; in
835 of them, only an external visual inspection was carried out. When necessary, both
exterior and interior inspections were conducted. In terms of building occupation, almost
three-quarters (73%) was residential and 27% was for commercial use (offices and retail).
Regarding the number of stories, 45% of buildings inspected had between 1 and 3 stories,
48% between 4 and 9 stories, and 7% taller than 9 stories. As a result of the inspection,
60% were green-tagged, 23% red-tagged, and 17% yellow-tagged. A green tag meant that
the building could be used immediately, a red tag meant that entrance to the buildings was
prohibited, and a yellow tag meant that entrance could be allowed for short periods of
time, but an additional review was necessary.
According to the inspection, 684 buildings either collapsed or experienced some level of
structural damage as a result of this earthquake. The distribution of structural damage
intensity is shown in Figure 9. Five levels of damage were used: none, light, moderate,
severe, and collapse. Light damage includes cracks typically smaller than 0.2 mm width in
concrete structural components and 1 mm in URM walls. Moderate damage includes
local damage to structural components, and cracks between 0.2 and 1 mm width in con-
crete members and between 1 and 5 mm in URM walls. Severe damage corresponds to
concrete crushing, exposed and/or buckled steel reinforcement, cracks larger than 1 and
5 mm in concrete and masonry components, respectively, or punching shear cracking
around columns.
Most damaged buildings are located in zones 6, 21, 22, and 37 of Figure 8. These zones
correspond to the edges of the lakebed zone which have a fundamental period of vibration
of around 1 s (see Figure 2). Damaged buildings inspected had an RC structure (flat slab
or frame structures), with or without masonry infills or a masonry load-bearing wall sys-
tem. No steel structures were identified with damage. The types of vulnerability and fea-
tures observed in the 38 collapsed buildings are presented in Table 1. As shown in Table 1,
all but one collapsed building were designed and built prior to the 1985 emergency norms
(DDF, 1985). The main problems causing collapse and damage in structures are further
discussed in the following.
16 Earthquake Spectra 00(0)
Effect of infill walls. Most infill walls were URM or poorly confined masonry elements.
Inclined cracking and, in hollow units, crushing and spalling were typical. In general,
those walls were built against the frame elements but had neither been considered to con-
tribute to lateral stiffness nor to strength during analysis and design. Infill walls were often
responsible for damage and collapse of buildings due to weak/soft story mechanisms, plan
asymmetries, and short columns.
Weak/soft story failure mechanisms. Similar to buildings damaged in 1985, weak/soft story
failure mechanisms were observed, mostly at the ground level. This was the case of con-
crete frames or flat slabs with masonry partitions in the upper floors and bare frames in
the open ground stories. Ground stories are often used as garages and/or for shops. In
these structures, inelastic deformation demands are concentrated on ground story struc-
tural components neither designed with enough strength and stiffness nor detailed for duc-
tile behavior.
Plan and elevation irregularities. Plan and elevation irregularities affected building perfor-
mance. Asymmetries in structural components layout and, in several cases, asymmetries
caused by infill wall layout and/or infill wall failures led to significant torsional demands.
In corner buildings, failure was attributed to either large torsional demands or lack of
enough lateral stiffness and strength. The latter was observed in neighborhoods where
building height was very similar, and building separation is very small, and in many cases,
nonexistent. Along the block, similar-height buildings exhibited in-phase lateral displace-
ments so that they deform as a single unit. This displacement pattern allows an interior
building to brace or to be braced by adjacent constructions, while weak corner buildings
suffer damage when displaced out of the block.
Flat slab construction. Like in 1985, failures in flat (waffle) slab buildings were observed at
or near slab–column connections. Their inherent lateral flexibility and reduction in capac-
ity to transfer moment and shear at large deformations led to damage and total or partial
collapses. Although punching failure modes were not observed at this time, severe concrete
damage at the connection was noted.
Weak concrete columns. Inclined cracking due to low shear capacity was observed in col-
umns. Column crushing and spalling as well as steel buckling were also noted. This was
particularly the case where tie spacing was larger than 20 cm and tie hooks had a 90° bend.
Short columns effects caused by wall parapets built against concrete columns were also
Murcia-Delso et al. 17
observed. The increase in the lateral stiffness caused by the short column effect combined
with insufficient transverse reinforcement prompted shear damage in those components.
Lack of or poor maintenance. Damaged buildings, with distinctly different damage intensity,
shared a common feature: lack of or poor maintenance of the structure. Concrete cracking,
due to poor concrete mix design, inadequate casting and/or curing, small reinforcement
cover, and corrosion of longitudinal and transverse steel reinforcement were observed. It
may be argued that such pre-event damage further reduced the capability of structural
components to resist the earthquake demands.
(a) In the concrete TN, for new flat slab construction, the allowable inelastic drift was
reduced to 0.5% and the seismic response factor Q (that accounts for inelastic
behavior) was set equal to 1.0. The conservatism adopted is aimed at eradicating
the use of flat slabs as the sole LLRS. As it was mentioned above, performance of
flat slabs as LLRS was poor. To comply with the 2017 TN requirement, some type
of bracing system is needed to control lateral displacements and attain a maximum
drift of 0.5% For the MCBC, an allowable inelastic drift is obtained from an elas-
tic analysis (with reduced forces for inelastic behavior) multiplied by the over-
strength and ductility factors. Both factors depend on the type of structure’s
material/system and redundancy and on the type of structure’s material/system
and fundamental period, respectively. Allowable inelastic drift and Q-factor for
new flat slabs are also required to be used for evaluating the seismic safety of dam-
aged flat slab buildings.
18 Earthquake Spectra 00(0)
(b) In the masonry TN, a new chapter for the design, construction, and detailing of
masonry partitions, not being part of the LLRS, was included. As indicated
before, infill frames, with poorly designed and constructed masonry walls, exhib-
ited moderate-to-severe wall damage that led buildings to be vacated. The new
chapter complements that for infill walls which are part of the LLRS, to aid the
structural designers with specific requirements for both options.
(c) In the seismic design TN, the new online system for determining uniform hazard
spectra was verified against recorded ground motions. Although few adjustments
were made in specific areas of the city, for most records, the design spectra were
equal to or larger than calculated spectra from ground motions.
As a result of several revisions of the damaged building surveys and inventories, the
final number of damaged structures is made of 11,880 single-family masonry housing units
and 250 RC buildings. As of May 2020, of the 11,880 dwellings, 9050 are under rehabilita-
tion and 2830 are either being rebuilt in the same plot or being relocated. Rehabilitation
and reconstruction costs will be covered with public funds and support from private foun-
dations. Due to distinctly different damage intensity, the type and extent of rehabilitation
of the 9050 houses are variable. Jacketing of masonry walls with welded wire meshes
(WWMs) covered with cement mortar is the preferred rehabilitation technique. Typically,
WWM with 10- and 6-gage wires, spaced at 150 mm, and connected to the wall through
nails, 6 mm diameter wire or fasteners installed through powder-actuated tools, are used.
The design, construction, and detailing of this technique was codified in the 2004 version
of the masonry TN (DDF, 2004) and is used in the rehabilitation program.
For the evaluation and rehabilitation of concrete damaged buildings, MCACSS devel-
oped an ad hoc regulation (GCM, 2017a). These rehabilitation norms (RN) include mini-
mum requirements which are complementary to those in the TN. For the evaluation of
structural safety, RN requires the determination of the compressive strength and modulus
of elasticity of concrete and the yield strength of reinforcement from tests on samples
taken from the building. As an alternative, conservative values may be used depending on
the age of construction. Three different intensity damage levels (light, moderate, and
severe), consistent with those used for damage classification in Figure 9, are established.
Evaluation acceptance criteria in terms of inelastic lateral drift and structure damage level
were developed. When a structure exhibits light damage and the calculated inelastic drift
levels (computed with the new seismic design TN) are smaller than 0.6%, the RN repair
methods may be applied. If the computed drift is larger than 0.6%, regardless of damage
level, rehabilitation is required. The effects of damage intensity to strength and stiffness of
members and of building inclination were included in RN. For structures with moderate-
to-severe damage levels, a seismic modification factor Q = 2 should be used.
RN includes requirements for the design and construction of different rehabilitation
techniques. With the exception of flexure-dominated members, strength factors were
reduced by 25% from those for new construction to account for member damage and
uncertainties. Requirements for jacketing of concrete members, through RC jackets, steel
plates, and fiber reinforced polymer (FRP), are included. Addition of new concrete walls
or steel braces is a technique that is abundantly detailed, especially for weak/soft story
building applications. In these cases, RN requires the ground story to resist calculated
elastic forces (i.e. Q = 1.0) without exceeding a 0.6% lateral drift. In the case of building
inclination, RN requires the correction of out-of-plumbness for inclinations larger than
1% of the total building height above ground.
Murcia-Delso et al. 19
The final section of RN is devoted to quality control and inspection during construc-
tion. Emphasis on the coordination and agreement around specifications, procedures,
methods, strategies, and materials for all parties involved is made in this section.
This database is developed using information provided by the Mexico City Government
through its Institute for Construction Safety (ICS). ICS estimates that the total number of
RC buildings to be rehabilitated is 300; as of May 2020, 285 have been approved for con-
struction. As of May 2020, 208 buildings have been introduced in the database, and field
information has been collected in 50 of them. Main features of the 208 buildings currently
in the database are
Structural damage: concentrated mostly in ground story columns and walls; dam-
age intensity distribution is severe in 27% of the buildings; moderate in 36% of the
buildings, and light in 12% of the buildings; in 25% of the cases, damage is not well
classified;
Inclination: 40% of buildings under rehabilitation;
Nonstructural damage: infill walls, 67%; finishing, 73%; facade, 58%; and stair-
cases, 37%;
Buildings rehabilitated before the 2017 earthquake: 34. According to building occu-
pants and/or neighbors, 20 additional structures had been rehabilitated in previous
years; however, no official record has been found.
Eighteen different rehabilitation techniques have been identified in the 206 buildings
studied to date; their frequency of use is shown in Figure 10. Two buildings were removed
from the analysis, as they were demolished for economic reasons.
Consistent with the prevalence and intensity of damage in masonry walls, wall jacketing
through WWMs covered with mortar has been used in almost 49% of the cases (Figure
11).
Addition of new concrete walls is also widely used (46%) to improve the lateral capac-
ity of buildings (Figure 12). New walls often require strengthening of the foundation, com-
prising addition of new piles and/or enlargement of existing grade beams and slabs. The
foundation was intervened in 31% of the buildings. Frame jacketing is also a widely used
technique aimed at increasing lateral strength and stiffness while maintaining the frame
LLRS. Concrete jacketing has been used in 35% of the buildings, whereas steel jacketing
in 34%.
Aimed at reducing lateral drifts, steel braces have been added to existing structures in
30% of the cases. This rehabilitation strategy has been typically used in structures
Figure 10. Rehabilitation techniques of RC structures used in Mexico City after the 2017 earthquake,
as of May 2020 (based on data from 206 buildings).
Murcia-Delso et al. 21
Conclusion
The 19 September 2017 Puebla–Morelos earthquake provides a unique opportunity to
study the performance of buildings rehabilitated after 1985 and to collect valuable data on
new rehabilitation efforts. This article has summarized the rehabilitation techniques and
provisions used after the 1985 and 2017 earthquakes in Mexico City and the observed per-
formance of a limited number of rehabilitated buildings during the 2017 event. The follow-
ing conclusions are reached:
Figure 13. Steel braces added to increase the lateral strength and stiffness.
Modifications of the lateral force load resisting systems, which mainly involved
addition of RC walls or steel braces, were also employed to increase the lateral
strength and stiffness of the building structures. As of May 2020, similar rehabilita-
tion strategies and techniques have been used in buildings damaged after the 2017
earthquake. The most common types of intervention reported to date have been the
jacketing of existing masonry walls using WWM covered with mortar and the addi-
tion of new concrete walls.
Four case studies of concrete buildings rehabilitated after 1985, which were previ-
ously studied by Jara et al. (1987) and Aguilar et al. (1996), have been presented.
The buildings are located in the soft soil zone of Mexico City, and all but one were
severely damaged in 1985. The buildings were repaired and/or retrofitted using a
variety of techniques (e.g. concrete jacketing, steel bracing, new RC walls, and
removal of floors). The rehabilitated structures exhibited limited or no damage after
the 2017 earthquake. The shaking intensities experienced by these buildings in 2017
were presumably comparable to those of 1985 based on a qualitative examination
of the acceleration response spectra near the building sites. Hence, it is concluded
that the rehabilitations were effective in improving the performance of these build-
ings. Nevertheless, there is evidence of building failures during the 2017 earthquake
despite previous interventions. More in-depth and extensive investigations are
needed to arrive to more general conclusions on the effectiveness of rehabilitation
methods implemented after 1985, but the limited availability and quality of the
information on past interventions remains a major challenge for conducting this
type of studies.
A database of RC buildings rehabilitated in Mexico City after the 2017 earthquake
is being developed to inform research activities and development of future guide-
lines and provisions. Among other uses, the collected data will allow a more
Murcia-Delso et al. 23
Acknowledgments
The authors are grateful to James Jirsa, Professor Emeritus at the University of Texas at Austin, for
providing information about the building rehabilitations conducted after 1985. Similarly, the authors
would like to thank Renato Berrón and his team at the Mexico City’s Institute for Construction
Safety for the support, information, and encouragement for developing this project. Santiago Garcia
and Francisco Garcia are also acknowledged for the provided information. The authors would also
like to acknowledge students Jorge Luis Abarca, Rubén Bautista, Germán Bogoya, Vı́ctor Cruz,
Bernardo Moctezuma, Diana Ramı́rez, Gianella Valencia, and César Viramontes for their dedicated
work and contagious enthusiasm.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/
or publication of this article: This work was financially supported by CONTEX, The University of
Texas System and CONACYT. The authors would also like to gratefully acknowledge the support
of the Mexican Nacional Council for Science and Technology through FORDECYT 297246 project.
The first author also acknowledges the support from the University of Texas at Austin through a
VPR Special Research Grant for the reconnaissance trip.
ORCID iD
David Murià-Vila https://orcid.org/0000-0003-3597-617X
References
Aguilar J, Breña SF, Del Valle E, Iglesias J, Picado M, Jara M and Jirsa JO (1996) Rehabilitation of
existing reinforced concrete buildings in Mexico City: Case studies. PMFSEL 96-3, The University
of Texas at Austin, Austin, TX.
Alcocer SM and Castaño VM (2008) Evolution of codes for structural design in Mexico. Structural
Survey 26: 17–28.
Departamento del Distrito Federal (DDF) (1976) Reglamento de Construcciones para el Distrito
Federal. Mexico City: Departamento del Distrito Federal.
Departamento del Distrito Federal (DDF) (1985) Normas de Emergencia al Reglamento de
Construcciones del Distrito Federal. Mexico City: Departamento del Distrito Federal.
Departamento del Distrito Federal (DDF) (1987) Reglamento de Construcciones para el Distrito
Federal. Mexico City: Departamento del Distrito Federal.
Departamento del Distrito Federal (DDF) (1988) Los sismos de 1985. Control de edificaciones.
Mexico City: Departamento del Distrito Federal.
Departamento del Distrito Federal (DDF) (2004) Reglamento de Construcciones para el Distrito
Federal. Mexico City: Departamento del Distrito Federal.
Fundación ICA (1988) Experiencias derivadas de los sismos de septiembre de 1985. Mexico City:
Limusa Editorial.
24 Earthquake Spectra 00(0)
Gobierno de la Ciudad de México (GCM) (2017a) Normas para la Rehabilitación Sı´smica de Edificios
de Concreto Dañados por el Sismo del 19 de Septiembre de 2017. Mexico City, Mexico: Gaceta
Oficial de la Ciudad de México, No. 211 bis, 4 December.
Gobierno de la Ciudadde México (GCM) (2017b) Normas Técnicas Complementarias. Mexico City,
Mexico: Gaceta Oficial de la Ciudad de México, No. 220 bis, 15 December.
Jara M, Hernández C, Garcia R and Robles F (1987) Casos Tı´picos de Reparación y Refuerzo de
Estructuras de Concreto. Mexico City: Universidad Autónoma Metropolitana.
Jara M, Hernández C, Garcia R and Robles F (1989) The Mexico earthquake of September 19, 1985.
Typical cases of repair and strengthening of concrete buildings. Earthquake Spectra 5: 175–193.
Murià-Vila D and González R (1995) Propiedades dinámicas de edificios de la ciudad de México.
Revista de Ingenierı´a Sı´smica 5: 25–45.
Rosenblueth E and Meli R (1986) The 1985 earthquake: Causes and effects in Mexico City. Concrete
International 8: 23–34.
Sociedad Mexicana de Ingenierı́a Estructural (SMIE) (2019) Resumen Preliminar de Daños de los
Inmuebles Inspeccionados por las Brigadas del CICM del Sismo del 19/09/2017. Available at:
http://www.smie.org.mx/archivos/eventos/mantengase-informado/2017-noviembre-resumen-
preliminar-danos-inmuebles-inspeccionados-brigadas-cicm.pdf (accessed on 2 September 2020)