Download as pdf or txt
Download as pdf or txt
You are on page 1of 14

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

net/publication/276092283

Cracking in Walls of a Building Adjacent to a Deep Excavation

Article  in  Journal of Performance of Constructed Facilities · December 2011


DOI: 10.1061/(ASCE)CF.1943-5509.0000185

CITATIONS READS
38 2,849

2 authors, including:

Lindsey Sebastian Bryson


University of Kentucky
104 PUBLICATIONS   1,134 CITATIONS   

SEE PROFILE

All content following this page was uploaded by Lindsey Sebastian Bryson on 27 May 2021.

The user has requested enhancement of the downloaded file.


Cracking in Walls of a Building Adjacent
to a Deep Excavation
L. Sebastian Bryson, M.ASCE1; and Michael J. Kotheimer, M.ASCE2

Abstract: A major concern for projects involving deep excavations in urban areas is the response of adjacent buildings and utilities to
excavation-related ground movements. Unfortunately, a purely theoretical approach to estimating building response to excavation-related
deformations is not possible due to the variability of the many factors that contribute to the response. Consequently, building response must be
estimated and evaluated primarily based on empirical observations and various structural approximations. The goal of estimating and evalu-
ating building response is to provide limiting criteria that will safeguard the structure against unacceptable damage. Thus, estimating the
extent of the building response and consequently the severity of excavation-related building damage is critical to establishing rational limiting
criteria for excavation support system designs. The most common measure of damage severity is the onset and growth of cracks in interior
walls of adjacent structures. Although several procedures have been suggested for estimating excavation-related crack growth, all of the
procedures have a common aspect in that they require the input of a critical strain, or the strain at the onset of cracking, as a critical input
parameter. This paper presents the results of three-dimensional finite-element analyses of a building adjacent to an excavation. The analyses
were used to evaluate the magnitude of strain that developed in the interior walls in response to the excavation-related ground movements.
This paper describes the procedures used to model and analyze the building. The paper also presents computed building responses at dates
corresponding to observations of cracking and presents discussions of strain levels in infill panel walls where cracking was observed and in
panels where cracking was not observed. The analyses showed that the initial cracking observed in selected infill wall panels could not have
occurred solely in response to excavation-related deformations. Consequently, it was found that the wall panels cracked as a result of a
combination of strains induced in the structure from self-weight settlement and excavation-induced displacements at the supports. These
analyses allowed the writers to suggest critical strain criteria. DOI: 10.1061/(ASCE)CF.1943-5509.0000185. © 2011 American Society of
Civil Engineers.
CE Database subject headings: Excavation; Cracking; Finite element method; Buildings; Ground motion.
Author keywords: Excavation; Adjacent building; Cracks; Three-dimensional; Finite-element model; Infill wall panels; Critical strain.

Introduction damage. These limiting criteria can in turn be incorporated into the
design considerations of the excavation support systems.
A major concern for projects involving deep excavations in urban Estimating the severity of excavation-related building damage is
areas is the impact of excavation-related ground movements on ad- critical to establishing rational limiting criteria for excavation sup-
jacent buildings and utilities. These impacts specifically refer to the port system designs. However, assessment of the degree of building
translation and rotation of the structure in reaction to lateral ground damage can be highly subjective. The most common measure of
movements and surface settlement. These translations and rotations damage severity is the onset and growth of cracks in interior walls
result in direct tensile strains, bending strains, and shear strains in of adjacent structures. Several researchers (Boscardin and Cording
the structural and nonstructural members of the buildings. A purely 1989; Boone 1996; Finno et al. 2005; Son and Cording 2005;
theoretical approach to estimating building response to excavation- Kotheimer and Bryson 2009) have suggested various procedures
related deformations is not possible due to the variability of the for estimating crack growth that results from excavation-induced
many factors that contribute to the response. Consequently, build- ground movements. These procedures tend to require the calcula-
ing response is estimated and evaluated on the basis of empirical tion of shear and tensile strains that are induced in the ground due to
observations and simplified structural approximations. The goal of the excavation activities and the approximation of the correspond-
estimating and evaluating building response is to provide limiting ing shear and tensile strains induced in the adjacent building using
criteria that will safeguard the structure against unacceptable beam-bending approximations. Although all of these procedures
differ in specifics, they all have the common aspect of requiring
1
Assistant Professor, Dept. of Civil Engineering, 161 Raymond Build- the input of a critical strain, or the strain at the onset of cracking,
ing, Univ. of Kentucky, Lexington, KY 40506 (corresponding author). as a critical input parameter.
E-mail: bryson@engr.uky.edu This paper presents the results of a three-dimensional (3D)
2
Associate II, Wiss, Janney, Elstner Associates, Inc., 2751 Prosperity finite-element (FE) analysis of a case study involving the observed
Ave., Suite 450, Fairfax, VA 22031. E-mail: mkotheimer@wje.com responses of a building adjacent to a deep excavation. The subject
Note. This manuscript was submitted on July 8, 2010; approved on of the case study is the response of the Frances Xavier Warde
October 15, 2010; published online on October 15, 2010. Discussion period
open until May 1, 2012; separate discussions must be submitted for
School adjacent to the Chicago Avenue and State Street subway
individual papers. This paper is part of the Journal of Performance of renovation project in Chicago. This case was presented by Finno
Constructed Facilities, Vol. 25, No. 6, December 1, 2011. ©ASCE, ISSN and Bryson (2002). The data presented in this paper are particularly
0887-3828/2011/6-491–503/$25.00. unique in that they include a 3D FE model of a building adjacent to

JOURNAL OF PERFORMANCE OF CONSTRUCTED FACILITIES © ASCE / NOVEMBER/DECEMBER 2011 / 491

Downloaded 19 Jan 2012 to 128.163.155.166. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
an excavation, which is used to evaluate the magnitude of strain that this paper to provide a frame of reference for the modeling assump-
developed in the interior walls in response to the excavation-related tions and simulation choices used for this research.
ground movements. It was believed that a 3D building model The structural response of the Warde School was of particular
was necessary because several researchers (Finno et al. 2002; interest because of its close proximity to the excavation. The base-
Zdravkovic et al. 2005; Finno et al. 2007; Son and Cording 2008) ment wall of the school was located approximately 2.1 m from the
have reported that excavation-related ground movements are inher- centerline of the support wall (i.e., a clear distance between the
ently 3D by nature. Thus, it can be concluded that the imposition support wall and the wall of the school is 1.65 m). A plan view of
of nonuniform movements of the foundation elements (i.e., 3D the Warde School and the Chicago Avenue and State Street subway
ground movements) on a nonsymmetrical building will result in renovation project is shown in Fig. 1.
3D building responses. When viewed in plan, the building has an “L” shape. On the
This paper describes the procedures used to model and ana- north, south, and west elevations of the building, the facade of the
lyze the Warde School. The paper presents computed building building consists of coarse limestone with a concrete masonry unit
performance at dates corresponding to observations of cracking (CMU) backup wall and evenly spaced windows. The east eleva-
and presents discussions of strain levels in infill panel walls where tion of the building consists of a common Chicago brick veneer
with a CMU backup wall.
cracking was observed and in panels where cracking was not
The Warde School is a rigid concrete frame building with a
observed.
combination of gypsum-based and masonry infill walls. Each floor
of the school consists of reinforced concrete pan and joist floor
sections supported at each level by reinforced concrete beams.
Frances Xavier Warde School and Adjacent
The beams are rigidly connected to concrete columns throughout
Excavation
the interior and are connected to masonry bearing walls and
The Frances Xavier Warde School is a three-story elementary columns around the perimeter. The exterior bearing walls rest on
school with a basement and is founded on spread footings. It was reinforced concrete basement walls. The basement walls are 2.8 m
built in the 1960s and is located in downtown Chicago. The tall and 400 mm thick and are supported by an approximately
1.2-m-wide continuous strip footing. The interior columns are
adjacent excavation was a result of the Chicago Avenue and State
supported by reinforced concrete spread footings that are approx-
Street subway renovation project. The renovation project included
imately 760 mm thick and vary in size from 1.5 by 1.5 m to 5.5 by
excavating 12.2 m of soft to medium clay to expose the existing
5.5 m. The average depth of the footings is approximately 3.7 m
subway station and tunnels to allow for capital improvements.
below the ground surface.
The improvements included adding a mezzanine level for office
and vendor spaces and making the station handicap-accessible.
The excavation support system for the renovation project was a Response of the Warde School
secant pile wall supported by both cross-lot struts and tieback
anchors. Details of the Chicago and State excavation and support The excavation-induced building damage within the Warde School
system performance have been presented by Finno and coworkers was primarily in the form of architectural cracking (i.e., nonstruc-
(Finno et al. 2002; Finno and Bryson 2002) and will not be repeated tural) of interior walls and exterior masonry walls. Periodic damage
herein. However, specific information about the case is presented in surveys were performed to monitor the progression of the building

Fig. 1. Plan view of the Chicago and State excavation

492 / JOURNAL OF PERFORMANCE OF CONSTRUCTED FACILITIES © ASCE / NOVEMBER/DECEMBER 2011

Downloaded 19 Jan 2012 to 128.163.155.166. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
damage throughout the duration of the construction activities. horizontal cracks had increased to about 6 and 4 mm, respectively.
A summary of the observed building damage and the correspond- A vertical crack developed in room 103 [Fig. 2(b)] during the ex-
ing construction activity is presented in Table 1. The start of the cavation activities. The vertical crack was approximately 610 mm
project (i.e., day 1) corresponds to the beginning of the installation long and emanated from the top corner of the connection between
of the secant pile wall. However, the secant pile wall installation the beam and the column.
activities are omitted from Table 1 because no damage was ob- Additional first floor damage occurred in the marble façade of
served in the Warde School during this activity. both the north and south walls of the entranceway foyer. The cracks
Building damage was first observed on day 73 when the cross- in the north wall were typically diagonal cracks at the corners of the
lot struts were in place and the excavation had extended into the marble panels and at the corners of fixtures inset into the wall. Most
soft clay layer at 5.5 m below ground surface. The initial cracking of these cracks were inclined at about 45° and radiated upward
was observed in rooms along the west side of the school on all three toward the west. This pattern of cracking indicates that the north
floors. This damage consisted of mostly diagonal hairline cracks wall cracks were caused by shear distortions. The damage along the
approximately 300–500 mm long in non-load-bearing walls. south wall consisted of long horizontal cracks that tended to follow
Occurrences of cracking were greater on the second and third levels the preferential pathways created by grain variations in the marble
of the building than on the first level. Hairline cracks were also panels. The horizontal cracks all emanated from the ends of the
observed in the basement, along the west end of the north founda- panels and radiated inward. The south wall cracks were most likely
tion wall. Upon completion of the excavation along the west side of the result of out-of-plane bending, or possibly out-of-plane shear,
the school on day 116, previously observed cracks widened and because the upper ends of the cracked panels were not flush with
extended on all floor levels. During station renovation and backfill, the lower ends. The initial cracks in the foyer were detected in the
existing cracks generally became larger. Step cracks were observed north and south walls on days 108 and 109, respectively. Cracks in
in the mortar of the south and north segments of the exterior west the north wall tended to develop in the upper half of the wall, while
wall on days 127 and 129, respectively. The final damage survey of cracks in the south wall tended to develop in the lower half of the
the Warde School was performed on day 365. wall. The marble panels for both walls became misaligned in
response to the excavation-related movements. A crack, 3 mm wide
First Floor Level Damage and 102 mm long, developed on day 207 in the marble base near the
floor along the south end of the foyer, well after the excavation
Fig. 2 presents a plan view of the first level and identifies the areas
along State Street had been completed (day 116). This crack could
where damage was observed. The figure shows the locations of
have been the result of some residual strains or some random
cracks on day 73 and day 365. As was previously mentioned, these
construction vibrations.
days reflect the onset of cracking and the postconstruction damage
survey dates, respectively. The damage location plan in the figure Second Floor Level Damage
shows that the majority of the damage on the first floor developed
in east-west trending walls. The first observation of cracking The damage location plan and characteristic cracks for the second
damage on the first level developed in the south wall of the assistant level are presented in Fig. 3. The figure shows that more damage
principal’s secretary’s office [Fig. 2(a)]. The largest crack in this was observed on the second level than on the first level. Further-
wall was a diagonal crack radiating upward and toward the west more, the damage first observed on day 73 occurred almost exclu-
at a 45° angle. Also, a horizontal crack was located near the bottom sively between the west wall of the school and the west wall of the
of the wall. On day 73, the diagonal crack was approximately north-south hallway.
610 mm long and had a maximum width of about 2–3 mm. The The initial damage on the second level typically consisted
horizontal crack was about 405 mm long and had a maximum of horizontal cracks emanating from doorways within the upper
width of 1–2 mm. By day 365, the widths of these diagonal and portions of the walls. Vertical and diagonal cracks developed on

Table 1. Summary of Building Damage Corresponding to Construction Activity


Day Construction activity Building damage
60 to 74 Install cross-lot struts
73 Excavate below first tieback level Interior cracks observed; hairline cracks in infill walls concentrated in second and third
floors; second floor door replaned
78 Cracks in mortar and limestone façade in exterior north wall at the west end of the
school; existing cracks extend, maximum width is 1 mm
79 Install first level tiebacks Cracks observed in first floor wall panels
98 Install second level tiebacks
99 New cracks observed in first floor wall panels; existing cracks widen and extend
108 Cracks observed in marble façade of entranceway foyer; hairline cracks in north
foundation wall in the west corner of the school; existing cracks widen and extend
116 Excavate to final grade Existing cracks widen and extend on all levels
127 and 129 Step cracks observed in the south and north segments of the west exterior wall
142 Cracks observed in floor tiles of cafeteria; second floor door replaned
151 Movements exceeded 32 mm; diagonal cracks observed in first floor wall in cafeteria;
existing cracks widen and extend
207 New crack observed in marble façade in entranceway foyer
225 to 310 Place backfill and remove struts

JOURNAL OF PERFORMANCE OF CONSTRUCTED FACILITIES © ASCE / NOVEMBER/DECEMBER 2011 / 493

Downloaded 19 Jan 2012 to 128.163.155.166. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
Fig. 2. Damage locations and typical cracks on the first level

Fig. 3. Damage locations and typical cracks on the second level

494 / JOURNAL OF PERFORMANCE OF CONSTRUCTED FACILITIES © ASCE / NOVEMBER/DECEMBER 2011

Downloaded 19 Jan 2012 to 128.163.155.166. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
the second level later in the project. In the three northern class- 2 mm wide on day 78. Throughout subsequent construction, most
rooms (room 210, room 208, and room 207), the damage consisted cracks on the third level lengthened, but widened only slightly.
of horizontal cracks in the south walls near the doors, vertical Cracking damage in the east-west trending walls was characterized
cracks in the north walls, and diagonal cracks in the east walls. by large diagonal cracks that often occurred in the middle of the
In the rooms on either side of the library (room 206 and room 202), walls and radiated upward at 45° toward the west. Diagonal cracks
the cracks consisted of vertical cracks in the corners of the east that were smaller in width were also observed emanating from the
walls, horizontal cracks at the beam-wall interface along the tops corners of electrical outlets and in wall cabinets in the east-west
of the north walls [Fig. 3(a)], and vertical cracks along the tops of trending walls. Corner diagonal cracks were observed in the south
the north walls. Some vertical cracks were observed emanating walls of rooms north of the north-south hallway and in the north
from the corners of doors. These cracks were at locations where walls of the rooms south of the north-south hallway. Vertical corner
doorframes had racked. In some instances, the racking of the door- cracks also developed at the interface between the south load-
frames was bad enough to require replaning of the doors. The bearing wall and the infill wall. Vertical cracks on the third level
partition wall in the library apparently shifted to the east, which were typically located in the north-south trending walls on the
caused the floor tiles near the door to bubble up. These tiles had eastern end of the school (the portion from the north-south hallway
to be removed and resized. Also, several horizontal cracks were to the eastern face). In particular, vertical cracks developed at the
observed in the north walls of the east-west hallway. top of the west wall in room 301 and at the top of the east wall of the
north-south hallway, between the boardroom and room 301.
Third Floor Level Damage
Fig. 4 presents the damage location plan for the third level. The
figure shows that the damage was as extensive on the third level Finite-Element Model
as that observed on the second level. Also, similar to the second
level, the damage first observed on day 73 occurred primarily As was previously discussed, excavation-related damage in the
within the central interior of the school. Comparing Figs. 3 and 4, Warde School was characterized via the onset and growth of cracks
it is seen that the distribution of damage on the second and third in interior walls of the building. The development and growth
levels was similar. However, the incidents of first observed damage of cracks were in turn related to the strains induced in the building
on day 73 were less on the third level than on the second level. due to ground movements. However, the excavation-related ground
The damage observed on the third level was primarily diagonal movements behind the support wall were 3D in nature (Finno et al.
cracks. The third level cracks also tended to be wider and longer 2002). Also, the Warde School is nonsymmetrical about both axes
than the cracks observed on the first and second levels. Five days of the excavation. Therefore, it was expected that the movements of
after the first interior cracks were observed, the typical crack widths the building in response to ground movement would be nonuni-
on the third level were approximately 0.75 mm. These widths are form. These nonuniform building movements resulted in a highly
compared to a typical width of 0.2 mm on the first and second indeterminate strain field within the structure. The indeterminacy
levels during this same time. The crack shown in Fig. 4 was was further exacerbated by the fact that the strain field also varied

Fig. 4. Damage locations and typical cracks on the third level

JOURNAL OF PERFORMANCE OF CONSTRUCTED FACILITIES © ASCE / NOVEMBER/DECEMBER 2011 / 495

Downloaded 19 Jan 2012 to 128.163.155.166. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
in magnitude from floor level to floor level. The resulting complex- element to displace with that shell element. The columns were
ity of the building movements and the subsequent strain field modeled as frame elements connected at the corners of two or more
necessitated the use of a 3D FE model of the Warde School to shell elements and formed moment-transferring connections with
realistically evaluate the excavation-related damage and the corre- the floor slab and beams. The floor, roof, and wall sections were
sponding strain levels. modeled with four-node quadrilateral shell elements, which com-
A 3D FE model of the Warde School was created using the 3D bined membrane and plate-bending behavior (i.e., in-plane transla-
FE analysis and design package SAP2000 version 7.42, developed tional and rotation components and out-of-plane two-way
by Computers and Structures, Inc. (CSI), Berkeley, California translational and rotation components, respectively).
(SAP2000 2000). The model of the Warde School was developed
from structural data presented by Bryson (2002). These data in- Wall Systems
cluded floor system details, beam and column schedules, and de-
The exterior walls of the school are reinforced concrete load-bearing
tails for the load-bearing walls. No details for the interior walls
walls. These walls are similar to columns in that, in addition to
or the overall building assemblage (i.e., structural and nonstructural
their own self-weight, they also carry vertical loads distributed
member connections, floor and wall penetrations, and the actual
by beams. The wall elements require that a connection, or node,
location of the interior walls) were included. However, Bryson
must exist at the point where a beam frames into the wall. The
(2002) provided room layouts that were taken from nonstructural
connection requirement was achieved by constructing the finite-
drawings.
element mesh for the walls in such a way that the beams
Material Properties connected to a corner point of one or more of the shell elements
making up the wall. This connection allowed the floor loads to be
Linear elastic behavior was used to characterize the structural transferred from the beams to the bearing walls. It should be
components of the model of the Warde School. Also, all materials noted that loads are only transferred to walls at the corner points
used in the model were assumed to be isotropic. Thus, material of the shell elements and therefore the loads are not uniformly
parameters required for the model included modulus of elasticity, distributed across the top of the shell elements making up the ex-
Poisson’s ratio, shear modulus, and the coefficient of thermal terior walls. The connections between the beams and the bearing
expansion. Mass density of the materials and material strengths walls were modeled as moment-transferring rigid connections be-
were also specified. cause the school building was built primarily using cast-in-place
Three material types were used for the Warde School model: concrete construction. The thickness of the exterior load-bearing
concrete, gypsum, and steel. Table 2 shows the material properties walls was input as 419 mm and the material properties of these
used for each material in the model. Concrete was used primarily walls correspond to the concrete material properties listed in
for all of the load-bearing structural components including beams, Table 2. Although mild reinforcement was included in the con-
columns, floor systems, and load-bearing walls. Gypsum was used struction of the walls, only the modulus of elasticity and the
for the non-load-bearing elements including interior infill and Poisson’s ratio for concrete were used to calculate internal
partition walls. Steel was used only for reinforcement in the load- stresses.
bearing structural components. No design specifications concern- The non-load-bearing walls of the school included both infill
ing material properties were available. Therefore, values for the walls and partition walls. Infill walls are defined as wall panels that
material properties were obtained from the recommendations given are connected to columns or load-bearing walls along their ends
by Hibbeler (1997) and Dowding (1996). and are connected to beams along their top and bottom. Partition
walls are connected to other walls or may be connected to columns
Beams and Columns or load-bearing walls on only one end. Also, in some cases the
The beams and columns of the Warde School are reinforced con- partition walls do not extend from a floor beam to a ceiling beam.
crete members with rectangular cross sections. Structural plans of Fig. 5 shows how the typical infill wall panel is connected to the
the building showed that the interior columns have spiral reinforce- beams and columns and how the physical model is represented as
ment and the exterior columns have square tied reinforcement. The an object-based model in SAP2000.
beams are reinforced for both flexure and shear with continuous The physical model [Fig. 5(a)] shows that the uniform floor slab
reinforcement across connections. The beam and column frame load is transferred to the beams. The beams, in turn, transfer load to
elements were represented in the finite-element model using the columns and the columns transfer load to the foundation. Floor
two-node beam elements. The beam elements were connected at
the midsurface of the shell elements with rigid constraints. This
condition caused the beam nodes connected to a particular shell

Table 2. Material Properties Used in the Building Model


Material
Properties Steel Concrete Gypsum
Mass density (kg=m3 ) 7,860 2,320 1,800
Modulus of elasticity (GPa) 200 25 12.5
Poisson ratio 0.3 0.2 0.2
Thermal expansion coefficient (106 =°C) 11.7 10 7.6
Shear modulus (GPa) 77 10.3 5.2
Steel yield strength (MPa) 250
Steel ultimate strength (MPa) 400
Concrete compressive strength (MPa) 28 Fig. 5. Wall system: (a) physical model; (b) object-based model

496 / JOURNAL OF PERFORMANCE OF CONSTRUCTED FACILITIES © ASCE / NOVEMBER/DECEMBER 2011

Downloaded 19 Jan 2012 to 128.163.155.166. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
slabs 1 and 2 (FLR 1 and FLR 2) represent the floor of a particular 6.1 m and 9.1 m and two distribution ribs are included near the
level, and floor slabs 3 and 4 (FLR 3 and FLR 4) represent the floor 1=3 and 2=3 points for spans between 9.1 m and 12.2 m.
of the level above. The object-based model [Fig. 5(b)] shows that In structural analysis and design, a distinction is typically made
the wall panel is represented by two shell elements (referred to as between ribs and beams. Beams are members that connect directly
area objects in the figure), connected to the beams and columns to columns or load-bearing walls and transfer floor loads to these
(referred to as line objects in the figure) at the nodes (referred columns and load-bearing walls. Ribs are interconnected between
to as point objects in the figure). The outside nodes of the shell beams, stiffening the floor system and distributing floor loads to the
elements [point object (Pt) 1, Pt 3, Pt 6, Pt 4) correspond to the beams. Both beams and ribs were modeled with beam elements.
beam and column connections. The inside nodes (Pt 2, Pt 5) of Two rib sizes were used throughout the model of the Warde School.
the shell elements correspond to the expansion joints between The ribs modeled in the northern portion of the school, which
FLR 1 and 2 and FLR 3 and 4. The end reactions of the beams included the first floor cafeteria and the areas corresponding to the
and the tributary loads from the floor slabs are transferred to the cafeteria on all levels, had the following dimensions:
columns and the corners of the walls as vertical point loads. Internal • Dimensions: 356 mm ðdepthÞ × 178 mm ðwidthÞ
stresses in infill wall panels, as a result of their connectivity, are the • Spacing: 686 mm center-of-rib to center-of-rib
result of vertical loading at the connection points and deformation All other areas of the school used ribs with the following
of the structural frame. Partition walls do not experience vertical dimensions:
loading and are not connected at all four corners of the elements • Dimensions: 254 mm ðdepthÞ × 127 mm ðwidthÞ
within a structural frame. Therefore, the internal stresses in parti- • Spacing: 635 mm center-of-rib to center-of-rib
tion walls result from deformations of the frame at the connection The concrete floor slab was 114 mm thick for both rib systems.
points only. The floor loads used in the model included the self-weight of both
Detailed structural plans of the infill and partition walls were not the slab and beam concrete elements, and some assumed dead
available, therefore all interior walls were assumed to have the loads. Reinforced concrete was assumed to weigh 23:6 kN=m3 .
material properties corresponding to gypsum in Table 2. Further, The assumed dead loads were actually design floor loads taken
all interior walls were assumed to have a thickness of 25 mm. from the structural plans. The calculation of the self-weight of the
structural members and the distribution of the loads was performed
Floor and Roof Systems internally by the software. The assumed dead loads used were as
follows: 4.8 kPa for corridors, stairs, the cafeteria, and storage
The floor system of the Warde School is composed of concrete rooms; 4.5 kPa for the kitchen; 2.9 kPa for the library; 1.9 kPa
floor slabs poured monolithically with the supporting joists. This for classrooms; and 1.2 kPa for the roof.
system is described as a one-way pan-and-joist floor system in To more accurately reflect actual concrete construction tech-
which the term “one-way” refers to flexural reinforcement oriented niques, rigid diaphragm assignments were made corresponding
in only one principal direction. Fig. 6 shows a typical floor system to the joist schedules found in the structural plans. A floor section
of the Warde School with the joist sections included. These joists, modeled as a rigid diaphragm translates within its own plane
or ribs, span in the direction toward the shortest dimension of the (global x;y plane) and rotates about an axis perpendicular to its
floor section. The floor systems include distribution ribs that are own plane (global z-axis) as a rigid body. Designating a floor sec-
oriented normal to the direction of the span. Load distribution ribs, tion as a rigid diaphragm has no effect on the out-of-plane behavior
as the name implies, are provided to equalize the deflection and of the floor section; it only affects the in-plane behavior. The rigid
ensure that the effect of concentrated loads is minimized. One diaphragm assignment provides a diaphragm constraint to all of
distribution rib is included near the center for spans between the corner points (nodes) of a particular floor section and to any
additional point objects (nodes) that are enclosed within the boun-
daries of that particular floor section. To approximate the presence
of expansion joints between floor sections, individual floor slab
systems were not rigidly connected.
Vertical and Horizontal Openings
The vertical and horizontal openings throughout the building were
also addressed in the FE model. Vertical openings in exterior walls
due to doors and windows were assumed to influence building
stiffness. Interior infill walls were non-load-bearing and conse-
quently vertical openings within these walls were assumed to have
a negligible influence on the overall stiffness of the building. There-
fore, the exterior walls of the model included vertical openings
while the interior walls did not. Building models typically include
horizontal openings to model stairwells. Although stairwells often
provide a structure with additional shear resistance, stairwells were
not included for simplification of the model. The uncertainty
associated with the design and construction details of the stairwells
outweighed any gain obtained from including the stairwells in the
model. Fig. 7 presents a 3D model of the Warde School showing
the vertical openings for the exterior windows and doors.
Foundation Systems
The foundation system of the Warde School was idealized as point
Fig. 6. Typical roofing and flooring system
objects (nodes) with six degrees of freedom, which included three

JOURNAL OF PERFORMANCE OF CONSTRUCTED FACILITIES © ASCE / NOVEMBER/DECEMBER 2011 / 497

Downloaded 19 Jan 2012 to 128.163.155.166. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
excavation-induced displacements for each day that data were
collected.

Self-Weight Settlement
The self-weight response of the Warde School was modeled in the
FE model using linear springs at support locations. The springs
idealized the force-deformation characteristics of the soil beneath
the supports. No settlement data were available for the Warde
School before the start of the excavation activities. Consequently,
a modulus of subgrade reaction values was used to compute the
spring constants. The modulus of the subgrade reaction, ks is
the linear approximation of the relationship between the applied
bearing pressure beneath a loaded area, q, and settlement, δ. It is
given as k s ¼ q=δ. The actual relationship between bearing pres-
sure and settlement is nonlinear and is a function of the shape, size,
and rigidity of the loaded area as well as the constitutive response of
the underlying soil. However, the careful use of selected subgrade
moduli to define spring constants is often used in practice, particu-
larly in cases where the main purpose of the analysis is to define the
behavior of a structural system.
For the Warde School, the spring constants beneath the interior
spread footings were calculated by multiplying the individual foot-
ing areas by the modulus of subgrade reaction. Similarly, the spring
constants beneath the continuous wall footings were obtained by
multiplying the tributary footing areas by the subgrade modulus.
Because of the stratigraphy at the Warde School location, different
modulus values were assigned to footings based on size and depth
of influence. A larger modulus was used for the exterior wall foot-
ings to reflect the fact that these footings had smaller bearing areas
and only stressed the upper levels of the stratigraphy. The larger
Fig. 7. Three-dimensional model of the Warde School interior spread footings stressed the lower levels of the stratigraphy
and tended to produce larger settlements. Therefore, interior spread
footings were assigned much smaller modulus values. A trial-and-
translational and three rotational degrees of freedom along the error process was employed to achieve a maximum settlement of
global x, y, and z-axes. Excavation-induced movements of the about 25 mm and settlement distribution that appeared reasonable
Warde School were modeled using fixed base supports (i.e., all for the school. This value was chosen because a settlement of
six degrees of freedom restrained) for the load-bearing walls 25 mm is commonly specified as a limiting value for commercial
and 3D pinned base supports (i.e., the three translational degrees structures. From that process, the subgrade modulus values of K s ¼
8:6 MN=m3 for the exterior wall footings and K s ¼ 1:6 MN=m3
of freedom restrained) for the spread footings. In addition, springs
for the interior spread footings were selected for the self-weight
were used to model the preexisting soil conditions for the school.
settlement analysis.
Combinations of rotational and translational springs were used at
various stages in the overall simulations.
Evaluation of the Cracking Reponses
Numerical Models of the Ground Movements The cracking responses of the Warde School were quantified by
evaluating the strain levels in select infill walls of the school.
Excavation-Induced Movements The strain levels corresponding to the onset of cracking are termed
critical strains. Critical strains represent the strain at which cracking
The response of the Warde School to the excavation-induced move-
becomes visibly evident. Burland and Wroth (1974) reported that
ments was isolated and evaluated by applying “restraint displace- the critical tensile strain for brickwork and other masonry infill
ment loads” to the support joints in the FE model. These loads were frames using the maximum bending and direct tension strains
simply the excavation-induced ground displacements and rotations ranged from 0.5 mε to 1 mε with the recommended value being
that were measured during the excavation activities. However, in 0.75 mε. Mainstone (1974) gave the maximum principal tensile
several cases there were no physical measurements of ground strains at the onset of cracking (i.e., critical strains) for full-scale
movements corresponding to the locations of the support joints. frames with infills as 0.15 mε to 0.2 mε. However, these early
Therefore, 3D contour plots were developed from the data recorded researchers developed critical strain criteria solely from the obser-
for the excavation, and displacements at the support locations were vations of buildings settling in response to self-weight only.
taken from these contour plots. Bryson (2002) presented the dis- Building response to excavation-related ground movements
placement contours for the major stages of the excavation, and differs from the building response to ground movements caused by
these data were used for the analysis herein. It is important to note self-weight settlement. Ground movements caused by self-weight
that the restraint displacements were the only loads applied to the settlement are typically centered within the structure itself. An
excavation model. The self-weight of the structure was not ap- appropriate approximation using a beam-bending analogy would
plied in this part of the analysis. The model was subjected to the be that of a simply supported beam with a point load at the midpoint

498 / JOURNAL OF PERFORMANCE OF CONSTRUCTED FACILITIES © ASCE / NOVEMBER/DECEMBER 2011

Downloaded 19 Jan 2012 to 128.163.155.166. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
or a uniform distributed load. The primary strains induced in the
N
structure from these types of deformations would be bending or

(2nd Floor)
diagonal strains. Excavation-related ground movements tend to

Room 210
012345
Scale (m)

E-W Hallway
gradually impinge upon portions of an adjacent building at a time.

(3rd Floor)
Thus, an appropriate approximation using a beam-bending analogy
would be that of a short cantilever beam. These types of deforma- Key
tions would primarily induce direct tensile strains and shear strains Column Locations

Chicago Avenue
in structures. The implication of the aforementioned discussion is
that the critical strain criteria for self-weight settlement are most Main Office
(1st Floor)
likely different from the critical strain criteria associated with
excavations.
To develop critical strain criteria for the Warde School, the
Library North Library South
occurrences of cracking were evaluated with respect to the (2nd Floor) (2nd Floor)
performance of the building in response to: (1) excavation-
Room 206 Assnt Principal Faculty Room
induced-movements-only, (2) self-weight settlement only, and (3) a (2nd Floor) Offc (1st Floor) (3rd Floor)
combination of self-weight settlement and excavation-induced
movements. This approach to the analyses allows the critical strain
criteria to be evaluated in terms of self-weight settlement and
excavation-induced movements separately. This approach is also State Street
an investigation into the assertion made by Finno and Bryson
Fig. 8. Locations of selected interior walls evaluated during analysis
(2002) that damage in the Warde School was observed to cause
cracking in reinforced concrete buildings at strain levels below
those reported in the literature. Finno and Bryson (2002) suggested Table 3. Summary of Observed Performance for Selected Infill Walls
that this was because preexisting strains induced in the building
due to self-weight settlement contributed to the final strains in Room Floor Observed cracking
the building at the time excavation-related damage was observed. Assistant principal’s office 1st On day 73
The basis of their assertion is that ground distortions caused by the Room 206 2nd On day 73
excavation-induced movements result in strains that, if imposed on
Library-south 2nd On day 73
an undeformed structure, may not cause damage. However, when
Faculty room 3rd On day 73
these strains are imposed on a structure that has already deformed
in response to the applied building loads, these excavation-induced Room 210 2nd Between day 73 and day 225
strains may result in damage at apparent strain levels smaller than Main office 1st No cracking observed
applicable to self-weight loading only. Thus, the reported damage Library-north 2nd No cracking observed
for the Warde School was most likely caused by a combination of East-west hallway 3rd No cracking observed
the two sources of movements.

0:3 mε and an upper-limit strain range of 0:4 mε. These


Strains in Infill Wall Panels limits were developed using the critical principal strain criteria
presented by Mainstone (1974) and Boone et al. (1999). The
This study evaluated the strain levels corresponding to the onset appropriateness of using these criteria for this application can be
of cracking in several selected interior infill walls. The locations shown from the plane strain transformation of component strains
of the selected infill walls are shown in relation to the building plan (Hibbeler 1997) given as
in Fig. 8.
The particular walls presented in Fig. 8 highlight specific con- sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
   2ffi
ditions relating to the response of the Warde School. For example, ε þ εz εz  εx 2 γ
εmax;min ¼ x  þ zx ð1Þ
cracking was first observed on day 73 of the excavation in the 2 2 2
assistant principal’s office (first floor), room 206 (second floor),
and faculty room (third floor). For room 210, cracking was not
where εx represents the component strain in the horizontal direc-
observed on day 73, but was observed sometime between the in-
tion, εz represents the component strain in the vertical direction,
stallation of the second level tiebacks (day 98) and placement of the
and γzx represents the in-plane component shear strain. For condi-
backfill (day 225). The library-south wall is a north-south trending
tions in which the horizontal and vertical component strains are an
wall (i.e., parallel to the major axis of the excavation) in which
order of magnitude less than the component shear strain, the critical
cracking was observed on day 73. The library-north wall is a north-
strain in terms of a component shear strain is approximately two
south trending wall in which no cracking was observed. The E-W
times the principal strain. For example, Mainstone (1974) and
hallway is an east-west trending wall (i.e., perpendicular to the
Boone et al. (1999) have reported the critical principal strain ranges
major axis of the excavation) in which no cracking was observed.
as being between 0:15 mε and 0:2 mε. Two times these critical
The main office wall is located on the first floor, approximately at
strain levels yields critical component shear levels between
the midpoint of the Warde School, and no cracking was observed.
0:3 mε and 0:4 mε. It is noted that these critical component
Table 3 presents a summary of the observed performance for the
shear strain levels appear to be about one order of magnitude
selected infill walls chosen for this study.
smaller than the critical maximum shear strain levels typically re-
Critical Strains ported in the literature (Boone 1996). It must be remembered that
although the cracking strains in the infill walls are being evaluated
The cracking responses of the selected infill walls are evaluated using the component shear strains, the critical strain criteria being
against two ranges of critical strain: a lower-limit strain range of used are essentially the maximum principal strains.

JOURNAL OF PERFORMANCE OF CONSTRUCTED FACILITIES © ASCE / NOVEMBER/DECEMBER 2011 / 499

Downloaded 19 Jan 2012 to 128.163.155.166. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
Strains from Excavation-Induced Movements office infill wall panel was incorrectly predicted to crack at the
beginning of the excavation. The room 210 and library-north infill
Fig. 9 shows the component shear strains obtained for several of the
selected interior walls. The excavation-induced-movements-only wall panels were predicted not to crack. Although this prediction is
FE model of the Warde School produced corresponding component correct for the library-north infill wall panel, it does not agree with
strains in which the shear strain levels were approximately one the observed response for the room 210 data. The east-west hall
order of magnitude greater than the horizontal and vertical compo- was observed not to crack during the excavation activities.
nent strain levels. The minimal values of the horizontal and the Although the upper-limit critical shear strain range matches this
vertical component strains within the infill wall panels would tend observation, the lower-limit critical shear strain ranges predicts
to indicate that the shear strains were the governing indication of the wall would crack on approximately day 190. Thus, the shear
damage. This finding agrees with conclusions drawn by several strain levels highlighted in Fig. 10 show that strains corresponding
other researchers (Boscardin and Cording 1989; Boone 1996; Son to the excavation-induced movements only do not produce crack
and Cording 2005; Osman and Bolton 2006), which state that shear response predictions that adequately match the observed responses.
deformation and the resulting shear strain is likely to be the The implication of this finding tends to agree with the assertion of
predominant mode of excavation-related deformation. Finno and Bryson (2002) that strains corresponding to self-weight
The data presented in Fig. 9 show that when only excavation- settlement must be taken into account when evaluating the overall
induced movements were considered, the assistant principal’s cracking observed in the Warde School.

0.0015

0.0010

CRACKING OBSERVED

0.0005
0.4mε
Shear Strain

0.3mε

NO CRACKING OBSERVED
0.0000

-0.3mε
-0.4mε
-0.0005
Asst Principal's Office - Cracking on Day 73
CRACKING OBSERVED
Room 210 - Cracking After Day 73
-0.0010
Library North - No Cracking
E-W Hallway - No Cracking
-0.0015
0 50 100 150 200 250 300 350 400
Days

Fig. 9. Excavation-induced shear strains only

0.0020

0.0015

0.0010
Component Shear Strain

CRACKING OBSERVED

0.0005
0.4mε
0.3mε

0.0000 NO CRACKING OBSERVED

-0.3mε
-0.4mε
-0.0005
Room 206 - Cracking on Day 73 CRACKING OBSERVED
Faculty Room - Cracking on Day 73
-0.0010 Asst Principal's Office - Cracking on Day 73
Library South - Cracking on Day 73
Room 210 - Cracking After Day 73
-0.0015
0 50 100 150 200 250 300 350 400
Days

Fig. 10. Walls where cracking was observed

500 / JOURNAL OF PERFORMANCE OF CONSTRUCTED FACILITIES © ASCE / NOVEMBER/DECEMBER 2011

Downloaded 19 Jan 2012 to 128.163.155.166. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
Strains from Self-Weight Settlement and observed results is that reinforced concrete frame buildings
with rigid strip foundation systems tend to be stiffer near the ground
The self-weight settlement component strain levels of the eight floors. The reason for this increased stiffness is a combination of
infill wall panels chosen to represent the response of the Warde the rigidity of the strip foundation and the constraint against lateral
School are presented in Fig. 11. The initial visual inspection of moments offered from the depth of embedment. Unfortunately, this
the Warde School indicated that only minimal to no hairline cracks
increased stiffness was not quantified during the development of
were observed prior to the excavation process. Based on the critical
the FE model. Therefore, its effects were not captured in the results.
strain criteria presented by Boone (1996) for buildings settling
The data presented in Fig. 10 show the 0:3 mε critical strain
under their self-weight, the component strain levels in Fig. 11 com-
range predicted that the room 206 infill wall panel would crack on
pare favorably to this initial observed building response. Fig. 11
day 80. The observed response for the room 206 wall was cracking
also shows that the component shear strain levels are approximately
sometime around day 73. Thus, the predicted response well
one order of magnitude greater than the horizontal and vertical
matched the observed response.
component strain levels. This observation indicates that shearing
The library-south infill wall panel was the only north-south
is the controlling mode of deformation within the Warde School
trending wall (i.e., parallel to the primary axis of the excavation)
under both excavation-only and self-weight settlement conditions.
observed to crack around day 73. This wall is located on the second
floor of the Warde School and was farthest away from the midpoint
Strains from Combined Excavation-Induced
of the excavation than the other walls. The predicted response of
Movements and Self-Weight Settlement
this wall was somewhat different than the walls previously dis-
The component strains due to the combined effect of both self- cussed. Whereas the previous walls experienced rather abrupt in-
weight and excavation-induced deformations were calculated using creases in the component shear strains at the time of cracking, the
the method of superposition. The combined component shear library-south wall experienced a gradual increase throughout the
strains in the selected wall panels were then compared to the critical excavation activities. The 0:3 mε critical strain range incorrectly
strain ranges. Fig. 10 presents the results for walls in which crack- predicted the infill wall would crack on day 15. However, near day
ing was observed. 73 the strain levels exceeded 0.35 mε and remained approximately
The data presented in Fig. 10 show the 0:3 mε critical strain at that level until near the end of the excavation.
range incorrectly predicted the observed response of the faculty The room 210 infill wall panel is unique among the walls evalu-
room infill wall panel. Under this critical strain range, the faculty ated for this study in that cracking was observed sometime between
room wall would be cracked prior to the excavation-related activ- the installation of the second level tiebacks (day 98) and place-
ities. However, the faculty room infill wall panel was observed to ment of the backfill (day 225). Fig. 10 shows the 0:3 mε critical
crack on day 73. The 0:4 mε critical strain range much better strain range predicted cracking on day 130. The 0:4 mε criti-
predicted the observed responses of the faculty room wall. Data in cal strain range predicted cracking on day 175. Fig. 10 also shows
Fig. 10 show the strain levels corresponding to this infill wall panel a gradual increase in component shear strain throughout the exca-
remain within the 0:4 mε critical strain range before exceeding vation activities, similar to what was observed for the library-south
0.4 mε on approximately day 74. infill wall panel. Thus, although both the 0:3 mε and 0:4 mε
Both the 0:3 mε and 0:4 mε critical strain ranges predicted strain ranges yield acceptable predictions, an average critical strain
that the assistant principal’s office infill wall panel would crack dur- range of 0:35 mε may be more appropriate.
ing excavation activities. However, the 0:3 mε range predicted Fig. 12 presents the results for walls in which cracking was not
the wall would crack on day 87 and the 0:4 mε range pre- observed. The data presented in Fig. 12 show the 0:3 mε critical
dicted the wall would crack on day 92. The assistant principal’s shear strain range correctly predicted the observed response of the
office infill wall panel was observed to crack on day 73. It is noted east-west hallway (i.e., no cracking). However, this critical shear
that the assistant principal’s office infill wall was located on the first strain range incorrectly predicted cracking in both the library-north
floor. A possible explanation for the discrepancy between predicted infill wall panel and the main office infill wall panel on day 185.

Fig. 11. Component strains due to self-weight settlement

JOURNAL OF PERFORMANCE OF CONSTRUCTED FACILITIES © ASCE / NOVEMBER/DECEMBER 2011 / 501

Downloaded 19 Jan 2012 to 128.163.155.166. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
0.0010
Library North - No Cracking
E-W Hallway - No Cracking
Main Office - No Cracking
CRACKING OBSERVED
0.0005
0.4mε

Component Shear Strain


0.3mε

NO CRACKING OBSERVED
0.0000

-0.3mε
-0.4mε
-0.0005
CRACKING OBSERVED

-0.0010
0 50 100 150 200 250 300 350 400
Days

Fig. 12. Walls where cracking was not observed

Table 4. Summary of Wall Panel Response for Combined Shear Strains excavation-related ground movements. Specifically, this study
Day cracking as predicted evaluated critical strain criteria using the results of 3D FE analyses
of the Frances Xavier Warde School, adjacent to the Chicago
0:3 mε 0:4 mε
Avenue and State Street excavation. From the analyses of the
Critical strain Critical strain Day cracking
Infill wall panel range range was observed
Warde School, the following can be concluded:
1. The FE model of the Warde School produced infill wall panel
Faculty room ≈0 ≈74 ≈73 strain levels that were consistent with the field observations of
Assistant principal’s ≈87 ≈92 ≈73 cracking.
office 2. Evaluation of maximum principal tensile strains showed that
Room 206 ≈80 ≈85 ≈73 the initial cracking observed in some of the wall panels se-
Library-south ≈15 No cracking ≈73 lected for this study could not have occurred solely in response
Room 210 ≈130 ≈175 73 to 225 to excavation-related deformations. Consequently, strain levels
of the infill wall panels within a building adjacent to a deep
East-west hallway No cracking No cracking No cracking
excavation must include preexisting strains induced in the
Main office ≈185 No cracking No cracking
building during the initial self-weight settlement. The magni-
Library-north ≈185 No cracking No cracking tudes of these preexisting strains depend on the initial settle-
ment distribution.
The 0:4 mε critical shear strain range correctly predicted that 3. The modulus of the subgrade reaction self-weight settlement
none of the three wall panels would crack during the excavation. approximation method adequately produced infill wall panel
However, given the responses predicted for the walls that cracked, strain levels representative of the preexcavation building
the 0:35 mε strain range may be more appropriate for this study. condition.
None of the three wall panels were predicted to crack using the 4. Computed component shear strains arising from the combined
0:35 mε strain range. self-weight and excavation-induced support displacements
Table 4 presents the summary of the wall panel responses for the were an order of magnitude greater than the component ver-
combined shear strain components. The table shows that although tical and horizontal strains. Hence, a condition approximating
the 0:4 mε critical strain range tended to perform better for the pure shear was induced in the wall panels. If pure shear con-
selected wall panels than the 0:3 mε critical strain range, a critical ditions are assumed, the critical strain criteria are equal to two
strain range of 0:35 mε appears to be more indicative of the times the maximum principal strains given by Boone (1996)
observed building response overall. and Mainstone (1974).
5. The results of the FE analyses indicated a critical strain range
of 0:35 mε appears to be more reflective of the observed
Conclusions building response overall.

The purpose of this research was to develop critical strain criteria


based on observed damage of a structure adjacent to a deep exca- Acknowledgments
vation. It is intended that these criteria be used to develop rational
design methodologies for excavation support systems that will pre- The material presented in this paper is based upon work supported
vent excavation-related damage to adjacent buildings. This study by the National Science Foundation under Grant No. CMS
developed critical strain criteria by evaluating the magnitude of 06-50911 under Program Director Dr. R. Fragaszy. This support
strain that developed in the interior walls in response to the was greatly appreciated.

502 / JOURNAL OF PERFORMANCE OF CONSTRUCTED FACILITIES © ASCE / NOVEMBER/DECEMBER 2011

Downloaded 19 Jan 2012 to 128.163.155.166. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
References Finno, R. J., Voss, F. T., Rossow, E., and Blackburn, J. T. (2005). “Evalu-
ating damage potential in building affected by excavations.” J. Geotech.
Boone, S. J. (1996). “Ground-movement-related building damage.” Geoenviron. 131(10), 1199–1210.
J. Geotech. Eng. 122(11), 886–896. Hibbeler, R. C. (1997). Mechanics of materials, 3rd Ed., Prentice Hall,
Boone, S. J., Westland, J., and Nusink, R. (1999). “Comparative evaluation Upper Saddle River, NJ, 518–520.
of building responses to and adjacent braced excavation.” Canadian Kotheimer, M. J., and Bryson, L. S. (2009). “Damage approximation
Geotech. J. 36(2), 210–223. method for excavation-induced damage to adjacent buildings.”
Boscardin, M. D., and Cording, E. J. (1989). “Building response to International Foundations Congress and Equipment Expo (IFCEE09),
excavation-induced settlement.” J. Geotech. Eng. 115(1), 1–21. ASCE, Reston, VA, 8.
Bryson, L. S. (2002). “Performance of a stiff excavation support system in Mainstone, R. J. (1974). “Discussion on Session No. 5.” Proc., Conf. on
soft clay and the response of an adjacent building.” Ph.D. thesis, Settlement of Structures, Pentech, London, 771.
Northwestern Univ., Evanston, IL.
Osman, A. S., and Bolton, M. D. (2006). “Ground movement predictions
Burland, J. B., and Wroth, C. P. (1974). “Settlement of buildings and as-
for braced excavations in undrained clay.” J. Geotech. Geoenviron. Eng.
sociated damage.” Proc., Conf. on Settlement of Structures, Pentech,
132(4), 465–477.
London, 611–654.
Dowding, C. H. (1996). Construction vibrations, Prentice Hall, Upper SAP2000. (2000). “SAP2000 three dimensional static and dynamic finite
Saddle River, NJ. element analysis and design of structures.” Analysis reference, version
Finno, R. J., Blackburn, J. T., and Roboski, J. F. (2007). “Three- 7.4, Computer and Structures, Inc., Berkeley, CA, 421.
dimensional effects for supported excavations in clay.” J. Geotech. Son, M., and Cording, E. J. (2005). “Estimation of building damage due to
Geoenviron. Eng., 133(1), 30–36. excavation-induced ground movements.” J. Geotech. Geoenviron. Eng.
Finno, R. J., and Bryson, L. S. (2002). “Response of building adjacent to 131(2), 162–177.
stiff excavation support system in soft clay.” J. Perform. Constr. Facil. Son, M., and Cording, E. J. (2008). “Numerical model tests of building
16(1), 10–20. response to excavation-induced ground movements.” Can. Geotech.
Finno, R. J., Bryson, L. S., and Calvello, M. (2002). “Performance of a stiff J. 45(11), 1611–1621.
support system in soft clay.” J. Geotech. Geoenviron. Eng. 128(8), Zdravkovic, L., Potts, D. M., and St John, H. D. (2005). “Modeling of a 3D
660–671. excavation in finite element analysis.” Geotechnique, 55(7), 497–513.

JOURNAL OF PERFORMANCE OF CONSTRUCTED FACILITIES © ASCE / NOVEMBER/DECEMBER 2011 / 503

View publication stats Downloaded 19 Jan 2012 to 128.163.155.166. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org

You might also like