Observed Behaviors of a Long and Deep Excavation

Constructed by Cut-and-Cover Technique in

Shanghai Soft Clay
Yong Tan1 and Bin Wei2
Downloaded from ascelibrary.org by WASHINGTON UNIV IN ST LOUIS on 08/05/13. Copyright ASCE. For personal use only; all rights reserved.

Abstract: A 16–18 m deep excavation for a metro station with in-plane dimensions of approximately 20 m wide by 290 m long was
constructed by using the cut-and-cover technique in soft clay in the Shanghai metropolitan area, in which many high-rise buildings and
utilities exist. To investigate the performance of this deep excavation and the associated effect on the adjacent buildings and utility pipelines,
a long-term comprehensive instrumentation program was conducted during construction. Field observations included deflections of
diaphragm walls, vertical movements at wall tops, ground settlements, and settlements of surrounding buildings and utilities. Analyses
of field data indicated that overexcavation (i.e., removing more soil than designed) and a long construction duration caused the diaphragm
walls to develop substantial deflections. The base and middle floor slabs played dominant roles in suppressing postexcavation wall deflections
and ground settlements. This excavation in general exhibited different behaviors than other excavation projects because of its relatively long
shape. Behaviors of deep excavations in Shanghai soft clay are affected not only by construction duration (i.e., time effects), but also by their
geometries (i.e., space effects). This project serves as a special case study and provides insights into the design and construction of a long and
deep excavation that uses the cut-and-cover technique in soft soils and in a metropolitan environment. DOI: 10.1061/(ASCE)GT.1943-5606
.0000553. © 2012 American Society of Civil Engineers.
CE Database subject headings: Excavation; Soft soils; Clays; Measurement; Time factors; China.
Author keywords: Long and deep excavation; Soft clay; Field measurement; Time effects; Length effects; Postexcavation.

Introduction Hashash et al. 2008; Tan and Li 2011). However, few studies
documented the effects of overexcavation (i.e., removing more
Construction of underground parking garages, deep basements, and soil than designed) on retaining wall deflections and ground
underground transportation systems in urban areas calls for exca- settlements.
vations supported by retaining structures. In a congested urban Two categories of methods commonly used to predict retaining
environment where space is limited, the movements of retaining wall deflections and ground settlements and to evaluate damage
walls and the ground must be controlled within acceptable levels. potentials in adjacent buildings due to excavation are (1) analytical
Otherwise, adjacent structures or facilities may be damaged and studies or numerical analyses (e.g., Finno and Harahap 1991;
loss of lives may occur. Therefore, field monitoring of the perfor- Hashash 1992; Whittle et al. 1993; Hashash and Whittle 1996,
mance of a deep excavation that can provide immediate feedback to 2002; Finno and Calvello 2005; Finno et al. 2005; Son and Cording
engineers is necessary to ensure the safety of the project. Relevant 2005; Hashash et al. 2006; Osouli et al. 2010) and (2) empirical or
field data provide a good case history to calibrate and verify semiempirical methods (e.g., Peck 1969; Mana and Clough 1981;
numerical tools and facilitate a thorough understanding of the gen- Clough and O’Rourke 1990; Ou et al. 1993; Hsieh and Ou 1998;
eral behaviors of deep excavations. Since the publication of Peck Long 2001; Moormann 2004; Roboski and Finno 2006; Kung et al.
(1969), many researchers and engineers have contributed a huge 2007; Hashash et al. 2008; Schuster et al. 2009). Compared with
amount of field data to the knowledge of performance of excava- analytical studies or numerical analyses, empirical and semiempir-
tions and adjacent facilities (e.g., O’Rourke et al. 1976; Boscardin ical methods provide probably less accurate yet more straightfor-
et al. 1978; O’Rourke 1981; Finno et al. 1989; Ou et al. 1998; ward estimations. Thus, empirical and semiempirical methods are
Finno and Bryson 2002; Finno and Roboski 2005; Liu et al. appropriate approaches for the initial estimation of excavation
2005; O’Rourke and McGinn 2006; Blackburn and Finno 2007; performance and to validate numerical analysis results. Because
empirical and semiempirical methods are established from field
1
Assistant Professor, Dept. of Geotechnical Engineering, Tongji Univ., measurements, a large volume of comprehensive and detailed
1239 Siping Rd., Shanghai 200092, P.R. China (corresponding author). project data is usually required.
E-mail: tanyong21th@tongji.edu.cn Before the advent of the three-dimensional (3D) numerical tech-
2
Senior Engineer, China Railway Er-yuan Engineering Group Co., Ltd., nique, excavation was commonly treated as a plane-strain problem
Floor 24, Haili Building, 88 Dapu Rd., Shanghai 200023, P.R. China. in design and the effect of excavation length in the longitudinal
E-mail: bb951245@hotmail.com direction was disregarded. Through field measurements or finite-
Note. This manuscript was submitted on March 13, 2010; approved on
element (FE) analyses, some researchers (Ou and Chiou 1993;
April 12, 2011; published online on April 14, 2011. Discussion period open
until June 1, 2012; separate discussions must be submitted for individual Wong and Patron 1993; Lee et al. 1998; Ou and Shiau 1998;
papers. This paper is part of the Journal of Geotechnical and Geoenvir- Ou et al. 2000; Blackburn and Finno 2007) found that arching
onmental Engineering, Vol. 138, No. 1, January 1, 2012. ©ASCE, ISSN effects cause wall deflections at the excavation corners to be
1090-0241/2012/1-69–88/$25.00. significantly smaller than those at the central sections, where

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / JANUARY 2012 / 69

J. Geotech. Geoenviron. Eng. 2012.138:69-88.

 plane-strain condition can be assumed. Ou and Chiou (1993) sug- for the Shangcheng Road Station of Metro Line 9 is located in the
gested that the small length-to-depth ratio may contribute signifi- Pudong New Area of Shanghai. This area is on the east side of
cantly to the corner effects. After analyzing four metro excavations Huangpu River, facing the old downtown of Shanghai (Puxi) on
in Shanghai soft clay, Wang et al. (2005) concluded that the exca- the west side. Fig. 1 presents the site plan of the project along with
vations showed plane-strain behaviors during excavation, and no the instrumentation layout for field monitoring. The investigated
apparent corner effect was observed. They attributed this to the long excavation lies along Shangcheng Road with South Pudong Road
and narrow rectangular shapes of the excavations. On the basis of in the west. The excavation pit was sword-shaped with approximate
the results of 3D and 2D FE simulations, Finno et al. (2007) pointed in-plane dimensions of 20 by 290 m. Some buildings, including
out that when the excavation length is greater than six times the heavy high-rise concrete structures and one- and two-story brick
excavation depth, 2D simulations yield the same displacements in structures, are situated near the excavation. On the north side of
the center of the retaining walls as the displacements computed by the excavation are the four-story concrete structures of Rushan
Downloaded from ascelibrary.org by WASHINGTON UNIV IN ST LOUIS on 08/05/13. Copyright ASCE. For personal use only; all rights reserved.

3D simulations. New Village between South Pudong Road and North Nanquan
In this study, the performance of a long and deep excavation Road, a six-story concrete building along North Nanquan Road,
constructed by the cut-and-cover technique in soft clay in the met- two one-story brick structures for gas-pressure control stations,
ropolitan area of Shanghai, China, was investigated by dense arrays and a two-story brick structure. These buildings are all supported
of measurements on diaphragm wall deflections, vertical move- by shallow foundations. On the south side of the excavation pit are
ments at wall top, ground settlements, and settlements of buildings four heavy high-rise concrete structures (i.e., 38-story Xinmei Twin
and utility pipelines in the proximity. This long excavation was Tower Building, 24-story Neiwailian Building, 32-story Liangyou
compared with some cases in Shanghai with different excavation Building, and 29-story Lekai Building). These four high-rise build-
lengths to investigate the potential length effects on wall deflection ings rest on deep foundations, which are basically auger cast-
and ground settlements. in-place (ACIP) piles with diameters of 600–800 mm and lengths
of 40–45 m below the ground surface (BGS). The distances be-
tween the buildings and the excavation pit varied between 10 and
Project Description 40 m (see Fig. 1). All the buildings are relatively new and no more
than 20 years old. Other than the buildings, the excavation is also
As the largest city and the commercial and financial center of surrounded by many utility pipelines, including iron water pipe-
China, Shanghai is located at the estuary of Yangtze River on lines, iron power cable pipelines, and iron communication cable
China’s east coast. The city sits on the flat alluvial plain known pipelines. They are buried at depths of 0.7–1.5 m BGS.
as Yangtze River Delta, which features high groundwater tables To protect the city environment from potentially adverse or even
and thick soft clay in the upper layers. The investigated excavation detrimental effects caused by excavation, local metro construction

Fig. 1. Site plan of the project site and instrumentation layout

70 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / JANUARY 2012

J. Geotech. Geoenviron. Eng. 2012.138:69-88.

 Table 1. Deformation Control Criteria for Deep Metro Excavations in Shanghai (Based on SZ-08-2000)
Protection Allowable ground settlements
grade and lateral wall deflections Classification
1 (1) δvm ≤ 0:10%H e (2) δ hm ≤ 0:14%H e Running metro lines, utility tunnels, water mains, and important buildings or facilities
(e.g., historical buildings and gas pipelines) exist within H e distance from excavations.
2 (1) δvm ≤ 0:20%H e (2) δ hm ≤ 0:30%H e Important pipelines and/or buildings and/or facilities in service exist within H e ∼ 2H e
distance from excavations.
3 (1) δvm ≤ 0:50%H e (2) δ hm ≤ 0:70%H e No important utilities or structures exist within 2H e ∼ 4H e distance from excavations.
Note: δ vm = maximum ground settlements; δ hm = maximum lateral wall deflections; H e = final excavation depth.
Downloaded from ascelibrary.org by WASHINGTON UNIV IN ST LOUIS on 08/05/13. Copyright ASCE. For personal use only; all rights reserved.

standard SZ-08-2000 (Shanghai Municipal Engineering Authority unloading stress levels, the soil samples deformed drastically
2000) was enacted by the Shanghai Municipal Engineering Author- and then failed quickly.
ity (SMEA), in which different protection grades are specified for
metro excavations in Shanghai (see Table 1). According to this
code, protection Grade 2 was designated for construction of this Construction Procedures
deep excavation, i.e., the maximum excavation-induced wall
The investigated excavation consisted of four sections: Section 1 =
deflection, δ hm , should be no more than 54 mm, and the maximum
west shaft, Section 2 = standard segment, Section 3 = standard seg-
ground settlement, δ vm , no more than 36 mm. In addition, the maxi- ment, and Section 4 = east shaft. Tables 2 and 3 summarize the
mum allowable settlement of 60 mm was specified for concrete construction activities at Sections 1–4. The site work began with
structures, 40 mm for brick masonry structures, and 20 mm for the construction of concrete diaphragm walls and compaction-
utility pipelines. These criteria were defined on the basis that grouting the soils within 3 m below the final excavation level; these
settlements should not cause apparently cracking to buildings or activities were conducted between March 24 and May 16, 2007.
severe damage to utilities, which have occurred in excavation In excavation practice within the Shanghai area, the quality of
experiences in Shanghai in the past decades. compaction-grouting is usually difficult to control because of the
techniques and complex soil condition. In most cases, the reinforce-
ment layers below the excavation bases just functioned as curtains
Soil Conditions for cutting off the seepage of waterflow but did not constrain lateral
wall movements. Thereafter, excavation was initiated and strutting
Before the excavation, soil conditions at the site were explored by systems were propped as soils were removed. Excavation started
a series of field exploration programs (e.g., boreholes, standard from Section 4 on May 26, 2007 and ended at Section 3 on April
penetration test (SPT) borings, cone penetration test (CPT) sound- 26, 2008. The strutting systems consisted of steel pipes 609 mm in
ings, and dilatometer (DMT) soundings). In general, the site was diameter (external) and 16 mm in wall thickness, which were hori-
characterized by a thin layer of fill (Layer I) in the upper 2 m BGS, zontally spaced at 2–3 m along the longitudinal direction of the
followed by a layer of firm silty clay (Layer II) to a depth of excavation. All the steel pipe struts were preloaded with approxi-
5 m BGS. The next layer was very soft to soft muddy clay inter- mately 50% of their design strut load before propping. To allow
bedded with a thin layer of sandy silt (Layer III) extending to maximum working space inside the excavation, diagonal strutting
a depth of 17 m BGS, underlain by soft-to-stiff clay and silty was used at the corners. The vertical distance from the centers of
clay (Layer IV) to a depth of 28 m BGS. Beneath Layer IV, the the pipe struts to the nearest excavation surface directly below was
field exploration encountered dense-to-very-dense silty fine sand approximately 0.5 m. Finally, metro-station structures were con-
(Layer V) until the termination depth of 70 m BGS. The ob- structed in the order of base slabs, middle slabs, and roof slabs,
served long-term groundwater table at the site was approximately accompanied by the removal of the propping struts. The Young’s
1.0 m BGS. moduli, E, of the concrete and steel used at this site were 3 ×
The soil properties along the depth were characterized by a 107 kPa and 2 × 108 kPa, respectively.
series of laboratory tests and in situ tests. The laboratory tests in- Fig. 3 presents the typical cross section of the standard segments
cluded oedometer tests (confined compression tests), triaxial tests, (Sections 2 and 3). The excavation depths were 18 m at Sections 1
and unconfined compression tests. The effective cohesion, c0 , and and 4 and 16 m at Sections 2 and 3. The depths of diaphragm walls
friction angle, ϕ0 , were obtained from consolidated undrained were 33.0 m at Section 1, 33.6 m at Sections 2 and 3, and 31.2 m at
(CU) triaxial tests. The coefficient of compressibility, avð0:1–0:2Þ , Section 4. At the two end shafts (Sections 1 and 4), the excavation
and constrained modulus, E0:1–0:2 , were defined by oedometer tests was propped by six levels of steel pipe struts. At Sections 2 and 3,
at stress ranges of 100–200 kPa. The in situ vane shear test results the excavation was originally designed with five levels of steel
showed that the soft soils in the upper 25 m BGS exhibited medium pipes. However, to reduce the project cost and the construction
sensitivity, St , of approximately 2.5–5.0. This implied that the duration, the contractor did not follow the original design but
strength and stiffness of the soft clay would degrade substantially adopted a new excavation scheme in which the pit was propped by
once it was subjected to disturbances arising from construction only four levels of struts. This move caused overexcavation (i.e.,
removing more soil than designed) at each level for the central stan-
activities. The soil profiles along with the measured soil properties
dard segments.
are presented in Fig. 2. Both laboratory and field testing results
indicated that the soft soils in the upper layers feature relatively
high water content, high void ratio, high compressibility, and Instrumentation
low strength. Through a series of triaxial unloading creep tests, Fu
et al. (2009) reported that Shanghai clay exhibits apparent creep To monitor the performance of this excavation and ensure the safety
behavior during the process of unloading. At relatively high of the project, a long-term comprehensive field instrumentation

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / JANUARY 2012 / 71

J. Geotech. Geoenviron. Eng. 2012.138:69-88.

Downloaded from ascelibrary.org by WASHINGTON UNIV IN ST LOUIS on 08/05/13. Copyright ASCE. For personal use only; all rights reserved.

Fig. 2. Soil profiles and material properties

program was conducted. The observed excavation performance Observed Excavation Behaviors
included the deflections of diaphragm walls, vertical movements
at wall tops, and ground settlements behind the diaphragm walls. Lateral Movements of Diaphragm Walls
Excavation may cause the buildings and utility pipelines nearby to
settle uniformly or non-uniformly. Large settlements, especially Fig. 4 presents the development of diaphragm wall deflections
differential settlements, may induce serious structural damages to during construction. The field data show that at each section,
the measured wall deflections at the north side of the excavation
these buildings and utility pipelines and even cause the loss of lives.
exhibited similar developing patterns as those observed at the south
Therefore, the settlements of buildings and utilities in the proximity
side, which were propped by the same struts. As the excavation
were also surveyed throughout the construction.
proceeded to lower levels, the diaphragm walls began to develop
Fig. 1 shows the instrumentation layout used for monitoring the
deep-seated movements (bulging profiles) toward the excavation
performance of the excavation and the adjacent buildings. The de-
side. At Section 4, the diaphragm walls at B3 and B25 developed
flections of the diaphragm walls were monitored by 26 inclinometer
deflections of approximately 10 mm between Stages 5(a) and 6(a),
tubes (designated as B1 to B26) that were affixed to the steel which contrasted sharply with those measured at B2 and B4. This
reinforcement cages and later covered by concrete. The lengths of discrepancy might be because at B3 and B25 the struts at Level 4
these inclinometer tubes were approximately 17–30 m. To monitor were not promptly or properly propped. The continuous develop-
the ground settlements behind the diaphragm walls, 13 critical sec- ment of deep-seated deflections did not stop even when excavation
tions (designated as JS1-1 to JS1-5 through JS13-1 to JS13-5) were reached the final level. During the subsequent construction of the
selected for instrumentation. The settlement points were 5–25 m base slabs, middle slabs, and roof slabs, the diaphragm walls de-
away from the diaphragm walls and spaced 5 m apart. Ground set- veloped substantial postexcavation deflections toward the excava-
tlements were surveyed by level instruments. To address the safety tion side. The postexcavation wall deflections between Stages 8 and
concerns regarding the adjacent buildings and utility pipelines, their 11(a) were approximately 7.3 mm (50% of the measured final
settlements were also surveyed during and after excavation. deflection, δ hf ) at B2, 10.8 mm (35% δ hf ) at B3, and 7.8 mm
Unfortunately, the inclinometer tubes at locations B1, B6, B9, (28% δ hf ) at B25. According to the construction log in Table 2, the
B16, B22, B24, and B26, and the settlement points at JS2-3, JS5-1, contractor began to remove the struts at Level 6 just 6 days after the
JS5-2, JS6-1, JS6-2, JS10-2, JS11-1, and JS13-5 were damaged cast of the base slabs—much earlier than the concrete slabs could
during construction. Therefore, their measurements were not avail- gain sufficient strength (a curing time of 14 days is required for
able for this study. concrete to reach 70–80% of its design strength). These significant

72 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / JANUARY 2012

J. Geotech. Geoenviron. Eng. 2012.138:69-88.

 Table 2. Main Stages of Construction at Sections 4 and 1
Date (mm/dd/year)
Stage Events Section 4 Section 1
1(a) Construction of diaphragm wall 03/24/2007–05/25/2007 03/24/2007–05/25/2007
1(b) Compaction grouting 05/09/2007–05/16/2007 05/09/2007–05/16/2007
2(a) Excavation to 1.5 m BGS (Level 1) 05/26/2007–06/03/2007 11/11/2007–11/14/2007
2(b) Propping the struts at Level 1 06/03/2007 11/14/2007
3(a) Excavation to 4.7 m BGS (Level 2) 06/04/2007–06/06/2007 11/15/2007–11/17/2007
3(b) Installation of struts at Level 2 06/06/2007 11/17/2007
4(a) Excavation to 7.5 m BGS (Level 3) 06/07/2007–06/08/2007 12/03/2007–12/05/2007
Downloaded from ascelibrary.org by WASHINGTON UNIV IN ST LOUIS on 08/05/13. Copyright ASCE. For personal use only; all rights reserved.

4(b) Installation of struts at Level 3 06/08/2007 12/05/2007

5(a) Excavation to 10.5 m BGS (Level 4) 06/09/2007–06/10/2007 12/06/2007–12/07/2007
5(b) Installation of struts at Level 4 06/10/2007 12/07/2007
6(a) Excavation to 13.3 m BGS (Level 5) 06/17/2007–06/20/2007 12/11/2007–12/16/2007
6(b) Installation of level struts at Level 5 06/20/2007 12/15/2007–12/16/2007
7(a) Excavation to 15.3 m BGS (Level 6) 06/21/2007–06/24/2007 12/30/2007–01/01/2008
7(b) Installation of struts at Level 6 06/24/2007 01/01/2008
8 Excavation to 18.0 m BGS (final level) 06/24/2007–06/26/2007 01/02/2008–01/04/2008
9(a) Construction of base slab 06/28/2007–07/03/2007 01/10/2008–01/16/2008
9(b) Curing of base slab Since 07/04/2007 Since 01/16/2008
9(c) Removal of struts at Level 6 07/09/2007 01/18/2008
10(a) Construction of middle slab 07/15/2007–07/17/2007 02/24/2008–02/27/2008
10(b) Removal of struts at Levels 4 and 5 07/17/2007 02/27/2008
10(c) Curing of middle slab Since 07/18/2007 Since 02/27/2008
11(a) Construction of roof slab 08/03/2007–08/05/2007 03/14/2008–03/16/2008
11(b) Removal of struts at Levels 2 and 3 08/05/2007 03/17/2008
11(c) Curing of roof slab Since 08/05/2007 Since 03/16/2008
11(d) Removal of struts at Level 1 08/19/2007 —

Table 3. Main Stages of Construction at Sections 2 and 3

Date (mm/dd/year)
Stage Events Section 2 Section 3
1(a) Construction of diaphragm wall 03/24/2007–05/25/2007 03/24/2007–05/25/2007
1(b) Compaction grouting 05/09/2007–05/16/2007 05/09/2007–05/16/2007
2(a) Excavation to 2.0 m BGS (Level 1) 12/13/2007–12/15/2007 03/20/2008–03/22/2008
2(b) Propping the struts at Level 1 12/15/2007 03/22/2008
3(a) Excavation to 5.5 m BGS (Level 2) 12/16/2007–12/19/2007 03/23/2008–03/24/2008
3(b) Installation of struts at Level 2 12/19/2007 03/25/2008
4(a) Excavation to 9.0 m BGS (Level 3) 12/21/2007–12/23/2007 03/28/2008–03/31/2008
4(b) Installation of struts at Level 3 12/23/2007 03/31/2008
5(a) Excavation to 12.0 m BGS (Level 4) 12/29/2007–12/31/2007 04/01/2008–04/05/2008
5(b) Installation of struts at Level 4 12/31/2007 04/04/2008–04/05/2008
6 Excavation to 16.0 m BGS (final level) 01/14/2008–01/17/2008 04/20/2008–04/26/2008
7(a) Construction of base slab 01/28/2008–01/30/2008 05/02/2008–05/04/2008
7(b) Removal of struts at Level 4 01/29/2008–01/30/2008 05/04/2008
7(c) Curing of base slab Since 01/30/2008 Since 05/04/2008
8(a) Construction of middle slabs 02/08/2008–02/10/2008 06/19/2008–06/22/2008
8(b) Removal of struts at Level 3 02/11/2008 06/18/2008
8(c) Curing of middle slabs Since 02/10/2008 Since 06/22/2008
9(a) Construction of roof slabs 02/18/2008–02/20/2008 06/30/2008–07/03/2008
9(b) Removal of struts at Level 2 02/21/2008 06/29/2008
9(c) Curing of roof slabs Since 02/20/2008 Since 07/03/2008

postexcavation wall deflections may result from insufficient con- the roof slabs were constructed. Further examination showed that
crete curing time. As the base slabs gained strength over time most of the postexcavation wall deflections occurred before the
and the middle floor slabs were cast, the diaphragm wall move- completion of the middle slabs. Thereafter, the diaphragm walls
ments decreased gradually and almost stopped completely after developed very limited additional wall deflections. The diaphragm

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / JANUARY 2012 / 73

J. Geotech. Geoenviron. Eng. 2012.138:69-88.

 walls at B14 and B15 of Section 1 experienced deflection patterns
similar to those observed at Section 4. The final readings at
Section 1 were taken at the completion of the middle slabs. The
measured postexcavation deflections between Stages 8 and 10(a)
were 14.0 mm (41.8% δ hf ) at B14 and 13.3 mm (42.5% δ hf ) at B15.
During the construction at Section 2, the diaphragm walls
developed deep-seated deflections similar to those observed at
Sections 4 and 1. However, the diaphragm walls experienced un-
expected large deflections between Stages 5(a) and 6, which were
approximately 10.1–24.7 mm (45% to 77.2% of the measured
deflection at completion of excavation, δ he ). To interpret such ex-
Downloaded from ascelibrary.org by WASHINGTON UNIV IN ST LOUIS on 08/05/13. Copyright ASCE. For personal use only; all rights reserved.

ceptional wall deflections, the magnitudes of the theoretical axial

forces of the struts and the apparent earth pressure (AEP) envelopes
in the design and construction at the central standard segments are
plotted in Fig. 5 for comparison. The axial forces of the struts were
calculated by using the beam-on-elastic-foundation method (Liu
and Hou 1997). The AEP envelopes were converted from the com-
puted strut forces by using the method of Terzaghi and Peck (1967)
for propped excavations in soft-to-medium clay. Fig. 5 shows that
the calculated axial forces of the struts in case of overexcavation
(construction) were on average greater than the strut forces in
the design. For the final excavation to the base level, the computed
strut force at Level 4 in overexcavation was more than twice the
strut force at Level 5 in the design. Fu et al. (2009) point out that
Shanghai soft clay exhibits strong creep behavior and tends to de-
form drastically and fail quickly when subjected to high unloading
stress levels. As a result, the significant diaphragm wall deflections
between Stages 5(a) and 6 can be justified by both the high unload-
ing stress level caused by overexcavation and relatively long con-
struction duration (from December 31, 2007, to January 17, 2008).
Unfortunately, the final measurements at Section 2 were taken at
the completion of the base slabs. Therefore, the development of the
Fig. 3. Cross section of the excavation at the central standard segments
long-term postexcavation wall deflection remained unknown for

Fig. 4. Lateral deflections of wall at the north and the south sides of the excavation

74 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / JANUARY 2012

J. Geotech. Geoenviron. Eng. 2012.138:69-88.

 Design Over-excavation
which provided a good opportunity to examine the development of
138 kN/m 0 0 long-term postexcavation wall deflections in the case of of overex-
Design 183 kN/m
Over-excava. 2 cavation. Similar to the postexcavation wall deflections observed
2
at Sections 4 and 1, the diaphragm walls at B8 and B21 experi-
382 kN/m
4 4 620 kN/m enced substantial postexcavation deflections, which were around
13.7 mm (36.6% δ hf ) at B8 and 17.5 mm (42.2% δ hf ) at B21. Most
514 kN/m 6 6 of the postexcavation wall deflections at B8 and B21 took place

Depth (m)
Depth (m)
between the construction of the base slabs and the construction of
8 8 531 kN/m the middle slabs; afterwards, the diaphragm walls hardly developed
814 kN/m additional movements.
10 10 The deflection rate of a diaphragm wall is another critical index
1064 kN/m
Downloaded from ascelibrary.org by WASHINGTON UNIV IN ST LOUIS on 08/05/13. Copyright ASCE. For personal use only; all rights reserved.

for project safety. If the diaphragm wall develops large deflections

503 kN/m 12 12
within a short time, it may incur adverse or even disastrous results
to a project. Therefore, the maximum wall deflections, δh , are plot-
14 14
ted in a 120-day time frame for each section in Fig. 6. At Section 4,
16 16 the wall deflections during excavation almost increased linearly
0 50 100 150 200 250 300 with time. The deflection rates showed an apparent decrease after
Apparent earth pressure (AEP) the base slabs and middle slabs were constructed. The early
envelopes (kN/m2)
removal of the struts at Level 6 did not affect the deflection rate
substantially, which indicates the previous finding concerning
Fig. 5. Comparison of the theoretical strut forces and lateral active
significant postexcavation wall deflections was not correct. More
earth pressures in the case of design and construction
likely, a long construction duration and creep behavior and sensi-
tivity of Shanghai soft clay accounted for the significant postexca-
vation wall deflections—especially the time interval between when
this section. The observed postexcavation deflections in 2 weeks the excavation was completed and when the middle slabs were
[between Stages 6 and 7(a)] were approximately 2.0–6.5 mm. completed. The big jumps between Stages 5(a) and 6(a) at B3
The measured wall deflections at Section 3 duplicated the sim- and B25 were exceptional and may be attributable to the aforemen-
ilar development patterns as those measured at Section 2. Because tioned strutting problem. At Section 1, the diaphragm walls main-
of overexcavation and the relatively long construction duration tained constant deflection rates until the base slabs were cast and
(from April 5–26, 2008), the diaphragm walls developed signifi- then nearly stopped deflecting. This indicates that most of the post-
cantly large deflections between Stages 5(a) and 6, which were ap- excavation wall deflections occurred before the base slabs were
proximately 8.5–17.5 mm (35.1–56.3% δ he ). At B8 and B21, the completed and the soil-structure system almost reached the stable
final measurements were taken at the completion of the roof slabs, state after the completion of the base slabs. The excavation-induced

50 50
Remove struts

Middle slabs

constructed

Middle slabs
Roof slabs
constructed

Section 4 Section 1
constructed
45

Remove struts

constructed
45
Base slabs
constructed
Base slabs

at level 6
to final level

40
to final level

40

at level 6
Excavate

Excavate

35 35
30 30
δ h (mm)

25 25
Excavation

B2
to level 2

20 20
15 B3
15
10 B4 10 B14
5 B25 5 B15
0 0
6-13-2007

6-23-2007

7-13-2007

7-23-2007

8-12-2007

8-22-2007

9-11-2007

9-21-2007
6-3-2007

7-3-2007

8-2-2007

9-1-2007

11-14-2007

11-24-2007

12-14-2007

12-24-2007
12-4-2007

1-13-2008

1-23-2008

2-12-2008

2-22-2008
1-3-2008

2-2-2008

3-3-2008

50 50
45 Section 2 B10 45 Section 3
constructed
Roof slabs

40 B11 40
35 B12 35
30
δ h (mm)

30
Middle slabs

B13
constructed

25 25
B16-1 B7
20 20
constructed
Base slabs

B17
to final level

B8
to final level
constructed

15
Base slabs

15
Excavate

Excavate

10 B18 B20
10
5 B19 B21
5
0
0
12-15-2007

12-25-2007

1-14-2008

1-24-2008

2-13-2008

2-23-2008

3-14-2008

3-24-2008

4-13-2008
1-4-2008

2-3-2008

3-4-2008

4-3-2008

3-22-2008

4-11-2008

4-21-2008

5-11-2008

5-21-2008

5-31-2008

6-10-2008

6-20-2008

6-30-2008

7-10-2008

7-20-2008
4-1-2008

5-1-2008

Date (mm/dd/year) Date (mm/dd/year)

Fig. 6. Development of the measured maximum wall deflections with time during construction

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / JANUARY 2012 / 75

J. Geotech. Geoenviron. Eng. 2012.138:69-88.

 deflections at Section 1 were almost twice those of B2 and B4 at should be cast as soon as possible; and (3) overexcavation should
Section 4, although their deflection rates were similar. This should be treated with caution, especially at a great depth.
be primarily caused by the relatively longer construction time at
Section 1 and apparent creep behavior of Shanghai clay. It can Maximum Wall Deflections Attributable to Excavation
be verified by the following three facts: (1) similar to B2 and Fig. 7 shows the relationship between the measured maximum wall
B4 at Section 4, B14 and B15 of Section 1 were propped by diago- deflections at each excavation level, δ h , and the excavation depth,
nal struts; (2) B2, B4, B14, and B15 developed similar deflection H. For Sections 1 and 4, which were excavated according to the
rates over time during the excavation; and (3) as shown in Fig. 6, design, δ h lay between δ h ¼ 0:02%H and δh ¼ 0:08%H when H
it took almost 65 days before the casting of the rigid base slabs was less than 12 m, whereas δh fell between δ h ¼ 0:02%H and
at Section 1 (B14 and B15), whereas it took only approximately δ h ¼ 0:13%H when H was more than 12 m. For the overexcavated
30 days for Section 4 (B2, B4, B3, and B25). At Sections 2 Sections 2 and 3, δ h still lay between δ h ¼ 0:02%H and δ h ¼
Downloaded from ascelibrary.org by WASHINGTON UNIV IN ST LOUIS on 08/05/13. Copyright ASCE. For personal use only; all rights reserved.

and 3, wall deflection rates during excavation were much larger 0:08%H when H was less than 10 m; it then fell between δ h ¼
than those at Sections 1 and 4. As noted previously, overexcavation 0:08%H and δ h ¼ 0:13%H when H was equal to 12 m. Until H
resulted in greater unloading stress levels and caused the soft clay reached 12 m, δ h at the overexcavated sections followed the same
to creep. However, because the contractor managed to finish exca- developing pattern as that at the normally excavated sections. How-
vation as soon as possible, the magnitudes of the wall deflections ever, δ h at the overexcavated sections showed an abrupt increase
during excavation at Sections 2 and 3 were still comparable to those when H reached 16 m. Except for one scattering point, most of the
at Section 1 and those of B3 and B25 at Section 4. The deflection measured δ h lay between δ h ¼ 0:13%H and δh ¼ 0:23%H when
rates were greatly reduced after the base slabs and middle slabs H ¼ 16 m. This phenomenon indicates that in this project, overex-
were poured. Similar to the observations at Sections 1 and 4, the cavation in the upper 12 m BGS hardly imposed substantial effects
diaphragm walls at Section 3 did not developed apparent deflec- on δ h , while it affected δ h significantly when H was deeper than
tions following the casting of the middle slabs. The preceding 12 m BGS. Therefore, overexcavation at greater depths should
analyses indicate that to reduce diaphragm wall deflections induced be avoided.
by a cut-and-cover excavation in sensitive and creep Shanghai soft On the basis of case histories worldwide, some empirical rela-
clay, (1) excavation should be finished within a short time; (2) base tionships between δ h and H have been developed. The comparisons
slabs should be constructed right after excavation and middle slabs in Fig. 7 indicated that the measured δ h at this site was significantly

Fig. 7. Relationship between the maximum wall deflections and excavation depths

76 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / JANUARY 2012

J. Geotech. Geoenviron. Eng. 2012.138:69-88.

 H (m) excavations in medium-stiff Boston clay. The measured maximum
0 2 4 6 8 10 12 14 16 18 20 wall deflections were less than δ hm ¼ 0:3%H e specified for protec-
25 25
tion Grade 2 (see Table 1). The relatively smaller δ h at this site
should partially result from three factors: (1) the use of rigid con-
20 H 20 crete diaphragm walls; (2) the short duration of wall exposure; and
=
Hm
Sections 1 and 4
(3) the rapid cast of rigid concrete floor slabs after completion of
15 15 B2 excavation.

Hm (m)
Hm (m)

B3 On the basis of the field data collected from eight case histories,
B4
B14 Ou et al. (1993) found that the ratio of H m =H is approximately
10 10 B15 equal to 1.0 for excavations in Taipei soft clay, where H m = location
-6 B25
H where the maximum wall deflection occurs; and H = excavation
=
Downloaded from ascelibrary.org by WASHINGTON UNIV IN ST LOUIS on 08/05/13. Copyright ASCE. For personal use only; all rights reserved.

Hm depth. Fig. 8 shows the relationship between H m and H observed

5 5
at this project site. The ratios of H m =H fell between H m ¼ H
6 m
and H m ¼ H at Sections 1 and 4, indicating almost all the maxi-
0 0
0 2 4 6 8 10 12 14 16 18 20 mum wall deflections occurred within 6 m above the excavation
H (m) surfaces. For the overexcavated sections, H m =H fell between H m ¼
H
6 m and H m ¼ H þ 3 m, indicating some of the maximum
H (m) wall deflections occurred above the excavation surfaces and some
0 2 4 6 8 10 12 14 16 18 20
25 25 below. This discrepancy verified that if excavation proceeded as
designed, the lateral passive earth pressures provided by the soils
Sections 2 and 3 on the excavation side should be sufficient to constrain the develop-
B7
20
=
H 20
B8 ment of wall movements below the excavation surface and δ h
Hm B10 tended to occur above the excavation surface. However, if the pit
3
+ B11
15 =
H 15 was overexcavated, the lateral active earth pressures against the
Hm (m)
Hm (m)

B12
Hm B13 deeper portions of diaphragm walls would increase significantly
B16-1 (refer to Fig. 5). As a result, the chance that δ h occurred below
10 10 B17
-6 B18 the excavation surface would be increased.
H B19
=
Hm B20 Three-Dimensional Effects on the Maximum Wall
5 5
B21
Deflections
0 0 The dense arrangement of inclinometer casings at this project site
0 2 4 6 8 10 12 14 16 18 20 enabled the investigation on the effects of 3D geometry shape on
H (m) wall deflections. Fig. 9 presents the relationships between the nor-
malized maximum wall deflections, δ hm =H e , and the distance ratio,
Fig. 8. Relationship between the location of maximum wall deflections
d h =L, where δ hm = measured maximum wall deflection at the com-
and excavation depths
pletion of excavation to the final level; H e = final excavation depth;
d h = horizontal distance between an inclinometer casing and the
smaller than those reported by Peck (1969) for excavations in soft leftmost point of the excavation; and L = total length of the exca-
clay, by Clough and O’Rourke (1990) for excavations in stiff clay, vation along the longitudinal direction. To further examine the
and by Kung et al. (2007) for excavations in soft-to-medium clay, geometry effects on wall deflections in Shanghai soft clay, the
although the measured δh was close to Hashash et al. (2008) for reported data from four metro-station excavations (Yishan Road,

Distance ratio, dh/L

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
0.0
0.1 North side wall (this study)
0.2 South side wall (this study)
0.3 South Pudong Road (side 1)
δ hm/He (%)

0.4 South Pudong Road (side 2)

0.5 Pudian Road (side 1)
0.6 Pudian Road (side 2)
0.7 Yanchang Road
0.8 Yishan Road (side 1)
0.9 Yishan Road (side 2)
1.0

This study: L = 290 m, D = 31.2 ~ 33.6 m, W = 20 m, He = 16 ~ 18 m, L/W = 14.5, D/He = 1.7 ~ 2.1, L/He = 16.1 ~ 18.1;
Yishan Road: L = 335 m, D = 28 m, W = 17.3 m, He = 15.5 m, L/W = 19.4, D/He = 1.8, L/He = 21.6;
South Pudong Road: L = 169 m, D = 38 m, W = 22.8 m, He = 20.6 m, L/W = 7.4, D/He = 1.8, L/He = 8.2;
Pudian Road: L = 194 m, D = 26.5 m, W = 20.4 m, He = 16.5 m, L/W = 9.5, D/He = 1.6, L/He = 11.8;
Yanchang Road: L = 226 m, D = 27 m, W = 18.1 m, He = 15.2 m, L/W = 12.5, D/He = 1.8, L/He = 14.9;
L - Length of excavation in the longitudial direction;
D - Depth of diaphragm wall;
W - Width of central standard segment in the traverse direction;
He - Final excavation depth.

Fig. 9. Relationship between the normalized maximum wall deflection and the distance ratio

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / JANUARY 2012 / 77

J. Geotech. Geoenviron. Eng. 2012.138:69-88.

 South Pudong Road, Pudian Road, and Yanchang Road) summa- Factor of Safety Effects on the Maximum Wall
rized by Wang et al. (2005) were included in Fig. 9 for comparison. Deflections
For this sword-shaped excavation, the overall δ hm =H e along the Fig. 10 plots the relationship between the normalized maximum
longitudinal direction did not show substantial variations. The mea- wall deflections, δ hm =H e , and the factors of safety (FOS) against
sured δ hm =H e values close to the two ends were slightly smaller basal heave, in which δ hm = measured maximum wall deflection
than those measured in the middle; however, this cannot be solely at end of excavation; H e = final excavation depth; FOS was calcu-
attributed to the strengthening effects of corners because Sections 2 lated by using the method of Terzaghi (1943); and the two boun-
and 3 were overexcavated. dary curves were proposed by Mana and Clough (1981). The
By comparing the measurements from this project with those ground settlement data from the six metro excavations (Wang et al.
reported by Wang et al. (2005), the relationships between 2005) and 92 excavations for high-rise building basements (Xu
δ hm =H e and d h =L from this study were noted as similar to those 2007) in Shanghai were also included for comparison. Except
Downloaded from ascelibrary.org by WASHINGTON UNIV IN ST LOUIS on 08/05/13. Copyright ASCE. For personal use only; all rights reserved.

observed from the other four metro excavations, but the magnitude for the Yishan Road Station, all the metro excavations and build-
of δ hm =H e was smaller. Usually, the excavation-induced wall de- ing basement excavations were constructed by the cut-and-cover
flections are affected by factors such as the geometric shape of a method with multipropped concrete diaphragm walls. However, the
pit, wall embedment ratio, type of retaining wall, type of propping aspect ratios of length over width, AR, for the metro excavations
system, construction duration and length of unpropped wall expo- were much larger than those of the basement excavations. For the
sure, construction method, soil condition, and construction quality. metro excavations, AR ranged from 7.4–19.4, much larger than
To make the cases comparable, these factors were checked one 1.0–2.0 for those basement excavations. Fig. 10 shows that most
by one. The detailed geometric information and wall embedment building excavation data fell within the boundaries proposed by
ratio, D=H e , of each case are listed in Fig. 9, and they are reason- Mana and Clough (1981), although most of the metro excavation
data were near or below the lower boundary. This comparison dem-
ably close. All of the five metro excavations were retained by
onstrates the previous finding, i.e., a long and narrow multipropped
600–800 mm thick stiff concrete diaphragm walls propped by
excavation supported by rigid diaphragm walls is strongly capable
609-mm-diameter steel pipes. It is common practice in the excava-
of suppressing wall deflections during excavation.
tions of Shanghai metro to remove soils by 2–3 m with propping
systems, hence the wall exposure lengths were very similar in all Effects of System Stiffness on the Maximum Wall
cases. For the construction method, only Yishan Road was con- Deflections
structed by the top-down method, and the rest were constructed
Fig. 11 plots the normalized maximum wall deflections, δ hm =H e ,
by the cut-and-cover technique. Among the five cases, two are
and the normalized retaining system stiffness, EI=ðγw h4 Þ at this
located in the Puxi area (Yishan Road and Yanchang Road), where
site, in which EI=ðγw h4 Þ was defined by Clough et al. (1989).
the soil conditions are different from that of Pudong area. Puxi The six metro-station excavations reported by Wang et al. (2005)
features soft clay in upper 40–45 m BGS followed by dense sand and the 92 building excavations of Xu (2007) in Shanghai were also
layers, and Pudong is characterized by soft clay in upper 25–30 m included in this figure for comparison. Similar to Fig. 10, most
BGS, underlain by dense sand layers. However, when comparing of the observed δ hm =H e from the building excavations fell within
the cases of similar soil conditions, the measured δ hm =H e at Yishan the ranges proposed by Clough et al. (1989), although most of the
Road was much smaller than that of the Yanchang Road, and the observed δ hm =H e from the metro excavations was below the curve
measured δ hm =H e at Shangcheng Road (this case) was smaller of corresponding FOS of Clough et al. (1989) because of the afore-
than those of the South Pudong and Pudian roads. The excavation mentioned excavation geometry effects. As reported in literature
lengths of Yishan Road and Shangcheng Road (this case) were (e.g., Long 2001; Moormann 2004), an increase in system stiffness
exceptionally long. On the basis of the aforementioned consider- only imposes limited effects on reducing the normalized wall
ations, the significantly smaller δhm =H e measured in this study deflection.
may result from its long and narrow geometry. This can be justified
in that a long and narrow multipropped excavation supported by Vertical Movements at Wall Top
a concrete diaphragm wall is strongly capable of suppressing wall Substantial differential settlements of diaphragm walls could lead
deflections along the longitudinal direction during excavation. to the wracking of the entire supporting system. Therefore, it is

3.0
Section 1 (this study)
Sections 2 and 3 (this study)
2.5
Section 4 (this study)
Yishan Road Station
2.0 South Xizang Road Station
δ hm/He (%)

South Pudong Road Station

1.5 Pudongdadao Station
Pudian Road Station
1.0 Yanchang Road Station
Xu (2007)
Mana and Clough (1981)
0.5

0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
FOS

Fig. 10. Normalized maximum wall deflection versus FOS against basal heave

78 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / JANUARY 2012

J. Geotech. Geoenviron. Eng. 2012.138:69-88.

 2.0
Section 1 (this study)
1.8

FOS
Sections 2 and 3 (this study)

FOS
Section 4 (this study)

FO
1.6

= 1.
= 1.
Yishan Road Station

S=

0
1
1.4 South Xizang Road Station

1 .4
South Pudong Road Station

δ hm/He (%)
1.2
Pudongdadao Station
1.0 Pudian Road Station
FO Yanchang Road Station
0.8 S=
2.0 Xu (2007)
0.6
FO
0.4 S=
3.0
Downloaded from ascelibrary.org by WASHINGTON UNIV IN ST LOUIS on 08/05/13. Copyright ASCE. For personal use only; all rights reserved.

0.2
0.0
10 100 1000 10000
System stiffness, EI/γ wh4

Fig. 11. Normalized maximum wall deflection versus normalized system stiffness

22
20 Section 1 B14
Vertical movement at wall top

18

Stage 10(a)
16 B15

Stage 11(a)
14
Stage 9(a)
Stage 7(a)

12 B16
Stage 6(a)
Stage 4(a)
Stage 5(a)

10
Stage 8
(mm)

Stage 2(a)

8
Stage 3(a)

6
4
2
0
-2 Date (mm/dd/year)
-4
12-31-2007
11-1-2007

12-1-2007

1-30-2008

2-29-2008

3-30-2008

4-29-2008

5-29-2008

6-28-2008

7-28-2008
22
20 Section 2
Vertical movement at wall top

B10
18
Stage 6
Stage 5(a)

16 Stage 8(a) B11

Stage 4(a)
Stage 2(a)

14 Stages 9(a)
B12
12
(mm)

10 B13
8 B16-1
6
4 B17
2 Stage 7(a) B18
0
-2 Date (mm/dd/year) B19
Stage 3(a)
-4
12-31-2007
11-1-2007

12-1-2007

1-30-2008

2-29-2008

3-30-2008

4-29-2008

5-29-2008

6-28-2008

7-28-2008

22
20 B7 Section 3
Vertical movement at wall top

Stage 6

18
16 B8
14 B9
Stage 5(a)

12
Stage 3(a)
Stage 4(a)

B20
(mm)

10
Stage 2(a)

8 B21
Stage 8(a)

6
Stage 7(a)

4 B22
2
0
-2 Date (mm/dd/year)
-4
12-31-2007
11-1-2007

12-1-2007

1-30-2008

2-29-2008

3-30-2008

4-29-2008

5-29-2008

6-28-2008

7-28-2008

Fig. 12. Development of the vertical movements at wall top during construction

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / JANUARY 2012 / 79

J. Geotech. Geoenviron. Eng. 2012.138:69-88.

 necessary to monitor the wall settlements during excavation to en- Ground Settlements
sure the safety of the project. Fig. 12 presents the development of
Fig. 13 presents the development of ground settlements along JS1-1
vertical wall movements during construction. Beyond expectation,
to JS1-5 and JS2-1 to JS2-2 during the excavation of Section 4.
the diaphragm walls experienced significant heaves instead of set-
The ground settlements almost increased linearly with time during
tlements during the excavation. Two factors may contribute to such the excavation. Following the casting of the base slabs, the settle-
upward wall movements: (1) the elastic and plastic rebound of the ments stabilized quickly with time, which verified the previous
basal soils inside the excavation (i.e., basal heave) because of stress interpretation—rapid cast of rigid base slabs helps stabilize the
relief (soil removal), and (2) soil movements around the diaphragm retaining structures and surrounding soils. The measured maximum
walls. However, the magnitudes of the wall heaves at Sections 2 ground settlements were approximately 7–14 mm along JS1-1 to
and 3 were generally larger than those at Section 1 because over- JS1-5, and 8–20 mm at JS2-1 to JS2-2.
excavation increased the magnitudes of unloading and induced Fig. 14 summarizes the development of ground settlements
Downloaded from ascelibrary.org by WASHINGTON UNIV IN ST LOUIS on 08/05/13. Copyright ASCE. For personal use only; all rights reserved.

additional rebound of basal soils. At Section 3, exceptionally large along JS6-3 to JS6-5, JS7-1 to JS7-5, and JS8-1 to JS8-5 during
wall heaves occurred with a magnitude twice the wall heaves the construction of Section 1. During excavation, the measured
at Section 2 and three times the wall heaves at Section 1. Other ground settlements at these sections showed development patterns
than overexcavation, the excavation geometry and construction similar to those observed at JS1-1 to JS1-5 and JS2-1 to JS2-2.
sequence may contribute to this. When Section 3 was excavated, As the cast base slabs gained strength over time, the ground settle-
a wide-ranged stress field had been induced because of the extraor- ments started to decrease and completely stopped on February 9,
dinary length of the excavation. Therefore, soils were mobilized at 2008. Fig. 14 also shows that although excavation of Section 2 was
a large scale and tried to move toward those boundaries with weak executed around the same time, it did not impose any apparent
restraints (i.e., Section 3 being excavated). Following the cast of influences on the settling rates of the ground around Section 1.
Fig. 15 presents the development of ground settlements at JS5-3
concrete base slabs after the excavation, the wall heaves continued
to JS5-5 and JS9-1 to JS9-5 during the construction of Sections 2
to increase with time until the construction of the middle slabs. The
and 3. Contrary to the observations at JS5-3 to JS5-5 and those in
maximum wall heaves were approximately 5–7 mm at Section 1,
Figs. 13 and 14, even after the completion of the roof slabs, the
6–12 mm at Section 2, and 10–21 mm at Section 3. The diaphragm ground settlements at JS9-1 to JS9-5 did not stop until the start
walls at Section 3 developed a maximum differential heave of excavation at Section 3. This exception may be attributable
of 11 mm. to the combined effects of ancillary construction activities, traffic
The postconstruction measurements (i.e., after the completion of nearby, or other unknown factors. The ground almost stabilized at
the roof slabs) at Section 2 indicated that the diaphragm walls the time of the excavation of Section 3.
stopped heaving and settled down before they completely stabi- Fig. 16 presents the development of ground settlements at JS4-1
lized. This could be ascribed to the soil-structure interaction that to JS4-5 and JS10-1 to JS10-5 during the construction of Sections 2
tended to stabilize as excavation terminated and major structures and 3. Because these two surveyed sections were close to the boun-
were completed. dary of Sections 2 and 3, the measured ground settlements were
6-11-2007

6-21-2007

7-11-2007

7-21-2007

7-31-2007

8-10-2007

8-20-2007

8-30-2007
6-1-2007

7-1-2007

9-9-2007

0.0
Date (mm/dd/year)
JS1-1
-5.0
Settlement (mm)

JS1-2
-10.0 JS1-3
JS1-4
-15.0
Excavation of section 4 JS1-5
Construction of floors
-20.0

-25.0
6-11-2007

6-21-2007

7-11-2007

7-21-2007

7-31-2007

8-10-2007

8-20-2007

8-30-2007
6-1-2007

7-1-2007

9-9-2007

0.0
Date (mm/dd/year)
-5.0 JS2-1
Settlement (mm)

JS2-2
-10.0

Construction of floors
-15.0

-20.0
Excavation of section 4
-25.0

Fig. 13. Development of the ground settlements at JS1-1 to JS1-5 and JS2-1 to JS2-2

80 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / JANUARY 2012

J. Geotech. Geoenviron. Eng. 2012.138:69-88.

 11-11-2007

11-21-2007

12-11-2007

12-21-2007

12-31-2007
11-1-2007

12-1-2007

1-10-2008

1-20-2008

1-30-2008

2-19-2008

2-29-2008

3-10-2008

3-20-2008

3-30-2008
2-9-2008
0.0
Date (mm/dd/year)
-5.0 Excavation at

Settlement (mm)
section 2
-10.0
Construction of floor slabs at section 2
-15.0
-20.0 JS6-3
-25.0 JS6-4
Excavation at
-30.0 JS6-5 section 1
Downloaded from ascelibrary.org by WASHINGTON UNIV IN ST LOUIS on 08/05/13. Copyright ASCE. For personal use only; all rights reserved.

Construction of floor slabs at section 1

-35.0

11-11-2007

11-21-2007

12-11-2007

12-21-2007

12-31-2007
11-1-2007

12-1-2007

1-10-2008

1-20-2008

1-30-2008

2-19-2008

2-29-2008

3-10-2008

3-20-2008

3-30-2008
2-9-2008
0.0
Date (mm/dd/year)
-5.0 Excavation at
Settlement (mm)

section 2
-10.0 Construction of floor slabs at section 2
JS7-1
-15.0 JS7-2
-20.0 JS7-3 Excavation at
section 1
-25.0 JS7-4
-30.0 JS7-5 Construction of floor slabs at section 1
-35.0
11-11-2007

11-21-2007

12-11-2007

12-21-2007

12-31-2007
11-1-2007

12-1-2007

1-10-2008

1-20-2008

1-30-2008

2-19-2008

2-29-2008

3-10-2008

3-20-2008

3-30-2008
2-9-2008
0.0
Date (mm/dd/year)
-5.0 Construction of floor slabs at section 2
Settlement (mm)

-10.0
JS8-1
-15.0
JS8-2 Excavation at
-20.0 section 2
JS8-3
-25.0 JS8-4
Excavation at Construction of floor slabs at section 1
-30.0 JS8-5 section 1
-35.0

Fig. 14. Development of the ground settlements at JS6-3 to JS6-5, JS7-1 to JS7-5, and JS8-1 to JS8-5

greatly affected by excavation at Section 2. At JS4-1 to JS4-5, the Section 4. Therefore, the actual ground settlements caused by ex-
ground settlements stopped following the construction of the con- cavation at these two sections remained unknown. The magnitudes
crete floor slabs at Section 3. However, the ground at JS10-1 to of the ground settlements at JS12-1 to JS12-5 and JS13-1 to JS13-4
JS10-5 continued to settle with time. As the cast floor slabs gained located at Section 4 were only half of those observed at JS3-1 to
strength over time, the ground settlements at JS10-1 to JS10-5 sta- JS3-5 and JS11-1 to JS11-5 at Section 3. The substantial ground
bilized gradually and finally stopped approximately 1 month after settlement development (up to 12–15 mm) at JS12-1 to JS12-5 and
the casting of the roof slabs. This aberrant postexcavation ground JS13-1 to JS13-4 during the excavation of the nearby Section 3
settlement development at J10-1 to J10-5 can be partially inter- indicates that although these two survey sections were distant, their
preted by the significant postexcavation wall deflection between settlements were still affected by the stress field created by the
Stages 7(a) and 8(a) observed at B20 in the proximity (see Fig. 4).
excavation in the proximity. Unfortunately, the final measurements
Unfortunately, the final measurement at B20 was taken at the com-
were taken before the casting of the middle slabs. Hence, the long-
pletion of the middle slabs, and the long-term postexcavation wall
term postexcavation ground settlements at these sections were
deflection development was unknown. Therefore, the aforemen-
tioned reason for why J10-1 to J10-5 experienced distinct long- unknown.
term postexcavation ground settlements cannot be verified. As Fig. 18 plots the relationships between the measured maximum
observed at JS9-1 to JS9-5, it may be attributable to the combined ground settlements, δ vm , at each survey section and excavation
effects of ancillary construction activities, traffic nearby, or other depth, H. At Sections 1 and 4, δvm was bounded by δ vm ¼ 0:03%H
unknown factors. and δvm ¼ 0:15%H, whereas at Sections 2 and 3 it fell between
Fig. 17 presents the development of ground settlements along δ vm ¼ 0:03%H and δvm ¼ 0:18%H. The developed boundaries
JS3-1 to JS3-5, JS11-2 to JS11-5, JS12-1 to JS12-5, and JS13-1 for δvm at this site were much lower than those of Clough and
to JS13-4 during excavation of Section 3. Unfortunately, the ground O’Rourke (1990) for excavations in stiff clay. The upper boundary
settlements at JS12-1 to JS12-5 and JS13-1 to JS13-4 at the com- at Sections 1 and 4 was significantly below the corresponding
pleted Section 4 were not monitored during the excavation of line of Hashash et al. (2008) in medium-stiff Boston clay. At the

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / JANUARY 2012 / 81

J. Geotech. Geoenviron. Eng. 2012.138:69-88.

 12-11-2007

12-21-2007

12-31-2007
12-1-2007

1-10-2008

1-20-2008

1-30-2008

2-19-2008

2-29-2008

3-10-2008

3-20-2008

3-30-2008

4-19-2008

4-29-2008
2-9-2008

4-9-2008
0.0
Date (mm/dd/year)
-5.0
Excavation

Settlement (mm)
-10.0 of section 2 Construction
of floors
-15.0
Excavation of section 3
-20.0 JS5-3
-25.0 JS5-4
-30.0 JS5-5
Downloaded from ascelibrary.org by WASHINGTON UNIV IN ST LOUIS on 08/05/13. Copyright ASCE. For personal use only; all rights reserved.

-35.0

12-11-2007

12-21-2007

12-31-2007
12-1-2007

1-10-2008

1-20-2008

1-30-2008

2-19-2008

2-29-2008

3-10-2008

3-20-2008

3-30-2008

4-19-2008

4-29-2008
2-9-2008

4-9-2008
0.0
-5.0 Date (mm/dd/year)
Settlement (mm)

-10.0
JS9-1
-15.0
JS9-2 Excavation Construction
of section 2 of floors
-20.0
JS9-3
-25.0 JS9-4
-30.0 Excavation of section 3
JS9-5
-35.0

Fig. 15. Development of the ground settlements at JS5-3 to JS5-5 and JS9-1 to JS9-5
12-31-2007
12-1-2007

1-30-2008

2-29-2008

3-30-2008

4-29-2008

5-29-2008

6-28-2008

7-28-2008

8-27-2008

9-26-2008
0.0
-5.0 Date (mm/dd/year)
JS4-1
-10.0
Settlement (mm)

Excavation of section 3 JS4-2

-15.0
Construction of floor slabs JS4-3
-20.0 at section 3
Excavation of JS4-4
-25.0
section 2 JS4-5
-30.0 Construction of floor slabs
at section 2
-35.0
-40.0
12-31-2007
12-1-2007

1-30-2008

2-29-2008

3-30-2008

4-29-2008

5-29-2008

6-28-2008

7-28-2008

8-27-2008

9-26-2008

0.0
-5.0 Date (mm/dd/year)
-10.0 Excavation of
Settlement (mm)

section 2
-15.0
-20.0 JS10-1
-25.0 JS10-3 Construction of
floor slabs
-30.0 JS10-4 at section 2 Excavation of
section 3
-35.0 JS10-5
Construction of floor slabs at section 3
-40.0

Fig. 16. Development of the ground settlements at JS4-1 to JS4-5 and JS10-1 to JS10-5

overexcavated Sections 2 and 3, it was close to that of Hashash settlements attributable to excavation. By plotting the ground
et al. (2008). settlements, δ v , against the distance, d, away from retaining walls
On the basis of the ground settlement data from excavations sup- (δ v and d are normalized by the excavation depth, H e ), Peck (1969)
ported by temporary sheet pile walls and soldier piles, Peck (1969) categorized the ground settlements into three zones: I, II, and III.
proposed the first practical approach for the estimation of ground Fig. 19(a) shows the relationship between the normalized ground

82 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / JANUARY 2012

J. Geotech. Geoenviron. Eng. 2012.138:69-88.

 3-20-2008

3-30-2008

4-19-2008

4-29-2008

5-19-2008

5-29-2008
4-9-2008

5-9-2008

6-8-2008
0.0
Date (mm/dd/year)
-5.0

-10.0

Settlement (mm)
-15.0 JS3-1

-20.0 JS3-2
JS3-3
-25.0
JS3-4
-30.0
Downloaded from ascelibrary.org by WASHINGTON UNIV IN ST LOUIS on 08/05/13. Copyright ASCE. For personal use only; all rights reserved.

JS3-5 Excavation of section 3 Construction of floor slabs

-35.0 3-20-2008

3-30-2008

4-19-2008

4-29-2008

5-19-2008

5-29-2008
4-9-2008

5-9-2008

6-8-2008
0.0
-5.0 Date (mm/dd/year)
Settlement (mm)

-10.0

-15.0
JS11-2
-20.0
JS11-3
-25.0
JS11-4
-30.0 Excavation of section 3 Construction of floor slabs
JS11-5
-35.0
3-20-2008

3-30-2008

4-19-2008

4-29-2008

5-19-2008

5-29-2008
4-9-2008

5-9-2008

6-8-2008
0.0
Date (mm/dd/year)
-5.0

-10.0
Settlement (mm)

Excavation of section 3
-15.0
JS12-1 Construction of floor slabs
-20.0 JS12-2

-25.0 JS12-3
JS12-4
-30.0
JS12-5
-35.0
3-20-2008

3-30-2008

4-19-2008

4-29-2008

5-19-2008

5-29-2008
4-9-2008

5-9-2008

6-8-2008

0.0
Date (mm/dd/year)
-5.0

-10.0
Excavation of section 3
Settlement (mm)

-15.0
Construction of floor slabs
-20.0 JS13-1

-25.0 JS13-2
JS13-3
-30.0
JS13-4
-35.0

Fig. 17. Development of the ground settlements at JS3-1 to JS3-5, JS11-1 to JS11-5, JS12-1 to JS12-5, and JS13-1 to JS13-5

settlements and normalized distances for this project. Both mea- ground settlements fell below δvm ¼ 0:2%H e specified for the pro-
surements at completion of the excavation and the final readings tection Grade 2 (see Table 1). To better understand excavation
were presented. This shows that most of the measured maximum behaviors in Shanghai soft clay, the ground settlement data from

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / JANUARY 2012 / 83

J. Geotech. Geoenviron. Eng. 2012.138:69-88.

 25
(1) (3) Excavation of Sections 4 and 1
(2)
JS1-1~JS1-5

H
H
(1) JS2-1~JS2-2

.2%
Maximum ground settlement, δvm (mm)

0.5%
20 JS6-1~JS6-5

=0

H
5%
JS7-1~JS7-5

δvm =

0 .1
δv
JS8-1~JS8-5

=
(1) Clough and O'Rourke (1990)

vm
δ
15 (2) Hashash et al. (2008)
(3) This study

10
(2)
H
05%
Downloaded from ascelibrary.org by WASHINGTON UNIV IN ST LOUIS on 08/05/13. Copyright ASCE. For personal use only; all rights reserved.

δ vm = 0. (3)
5

δ vm = 0.03%H

0
0 2 4 6 8 10 12 14 16 18 20
Excavation depth, H (m)

25
(2) (3) Excavation of Sections 2 and 3
(1)
H JS3-1~JS3-5
(1)
.2%
JS4-1~JS4-5
H
Maximum ground settlement, δvm (mm)

0 .5 %

20 JS5-1~JS5-5
=0

H
5% JS9-1~JS9-5
%H
m
δvm =

δv

0.1
JS10-1~JS10-5
.1 8

=
=0

JS11-1~JS11-5
vm

15
δ

JS12-1~JS12-5
m
δv

JS13-1~JS13-5
(1) Clough and O'Rourke (1990)
10 (2) Hashash et al. (2008)
(2) (3) This study
%H
0 .0 5
δ vm =
(3)
5
H
δ vm = 0.03%

0
0 2 4 6 8 10 12 14 16 18 20
Excavation depth, H (m)

Fig. 18. Relationship between the maximum ground settlement and excavation depths

six metro excavations (Wang et al. 2005) and the 92 cut-and-cover along the measurement section; d = distance away from the dia-
excavations for building basements (Xu 2007) were also included phragm wall; and H e = final excavation depth. In this figure,
in this figure. The comparison of the measurements from metro the measurements at the completion of excavation and the final
excavations in Shanghai and those reported in literature showed measured data were plotted. Published case histories (Clough
that the metro excavation data in terms of magnitudes and distri- and O’Rourke 1990; Hsieh and Ou 1998; Wang et al. 2005; Xu
bution patterns in Shanghai soft clay were consistent with the Cen- 2007; Hashash et al. 2008) were also included for comparison.
tral Artery/Tunnel (CA/T) Project (11.3 km long) in medium-stiff Clough and O’Rourke (1990) pointed out that when the ground
Boston clay (Hashash et al. 2008), but far away from Zone III of settlements in Peck (1969) for excavation in soft to medium clay
Peck (1969) for very-soft-to-soft clay. Except for a few scattering are plotted as fractions of maximum settlements, the settlement dis-
data points, most of the data from the 92 building excavations also tribution is bounded by a trapezoidal envelope in which the zones
fell within Zone I of Peck (1969). However, the maximum ground of movements can be identified. When 0 ≤ d=H e ≤ 0:75, the maxi-
settlements attributable to the building excavations were much mum settlements occur; when 0:75 < d=H e ≤ 2:0, a transition
greater than those attributable to the metro excavations. This com- zone in which the settlements decrease from maximum to negli-
parison demonstrates that a long and narrow multistrutted excava- gible values exists. On the basis of the field measurements in Taipei
tion retained by rigid diaphragm walls is capable of suppressing soft clay, Hsieh and Ou (1998) defined the zone of d=H e ≤ 2:0 as
excavation-induced movements. In sharp contrast with the metro the primary influence zone and the zone of 2:0 < d=H e ≤ 4:0 as the
excavation data, the building excavation data exhibited an apparent secondary influence zone. Analyzing the field measurements from
tendency to decrease rapidly with distance from the excavation pits. those excavations in Shanghai soft clay (Wang et al. 2005; Xu
This discrepancy indicates that a much longer excavation might 2007; this study) showed that the ground settlements attributable
incur a much wider influencing zone behind the pit. to the building basement excavations (Xu 2007) were consistent
Fig. 19(b) presents the relationships between the normalized with those of Clough and O’Rourke (1990) for soft-to-medium clay
ground settlements, δ v =δ vm , and the distance ratio, d=H e for this and with those of Hsieh and Ou (1998) for Taipei soft clay. In con-
project, in which δ v = ground settlement along the surveyed section trast, the distribution ranges of the ground settlements during the
behind the diaphragm wall; δ vm = maximum ground settlement metro-station excavations were much wider. Fig. 19(b) shows the

84 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / JANUARY 2012

J. Geotech. Geoenviron. Eng. 2012.138:69-88.

 d / He
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
0.0

0.2

δv / He (%)
0.4 I
0.6 II
I - Sand and soft to hard clay
0.8
III II - Very soft to soft clay
III - Very soft to soft clay to a great depth
1.0
Downloaded from ascelibrary.org by WASHINGTON UNIV IN ST LOUIS on 08/05/13. Copyright ASCE. For personal use only; all rights reserved.

At the end of excavation (this study)

Final measurements (this study) (a)
Wang et al. (2005)
Xu (2007)
Clough and O'Rourke (1990) for stiff to hard clay
Hashash et al. (2008) for medium stiff clay
Zones I, II, III after Peck (1969)

d / He
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
0.0

At the end of excavation (this study)

Final measurements (this study)
Wang et al. (2005)
δ v / δ vm

0.5 Xu (2007)
Clough and O'Rourke (1990)
Hashash et al. (2008)
Hsieh and Ou (1998)
Wang et al. (2005)
1.0
(b)
Max. settlement Transition zone

Primary influence zone Secondary influence zone

Fig. 19. Observed settlement profiles versus those predicted by empirical methods

measured δ v =δ vm at this site spanned across the boundaries of the Distance ratio, dh/L
transition zone defined by Clough and O’Rourke (1990) and the 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
primary influence zone defined by Hsieh and Ou (1998). Most 0.0
of the measured δ v =δ vm lay between 0.5 and 1.0 at this site and 0.1

JS1-1~JS1-5
did not show an apparent tendency to decrease within the distance 0.2

JS2-1~JS2-5
JS6-1~JS6-5
JS8-1~JS8-5

JS9-1~JS9-5
JS7-1~JS7-5

of 0:3 ≤ d=H e ≤ 1:6. Similar to the observation from the CA/T

JS5-1~JS5-5

0.3
JS3-1~JS3-5
JS4-1~JS4-5
JS10-1~JS10-5

JS11-1~JS11-5

Project (Hashash et al. 2008), δ v =δvm at this site seemed to extend

δvm/He (%)

0.4
to a great distance behind the retaining walls. Unfortunately, the
0.5
ground settlements beyond 25 m (1:6H e ) from the retaining walls
were not surveyed, and hence at what distance δ v =δ vm would 0.6
decrease remained unknown. The significantly wider distributions 0.7 North side (this study)
South side (this study)
of δ v =δvm observed in the Shanghai metro excavations and the 0.8
Yishan Road-side 1
CA/T Project in Boston than those observed in literature for build- 0.9 Yishan Road-side 2
ing basement excavations indicate that a much longer excavation 1.0
would incur a much wider influencing zone behind the pit.
Fig. 20. Relationship between the normalized maximum ground
Three-Dimensional Effects on the Maximum Ground settlement and the distance ratio
Settlements
To examine the 3D geometric effects of the excavation on the excluded. The measured ground settlements during the excavation
ground settlements, the normalized maximum ground settlement, of Yishan Road Metro Station in Shanghai (Liu et al. 2005) were
δ vm =H e , at each monitored critical section behind the diaphragm also included in this figure for comparison. Similar to the observed
walls was plotted against the distance ratio, d h =L, in Fig. 20. H e wall deflections in Fig. 9, the variation of δ vm =H e with d h =L was
= final excavation depth; d h = horizontal distance between a survey very limited, and no apparent corner effect was observed at either
point and the left-most point of the excavation; and L = total length excavation site. The relatively larger δvm =H e observed at the central
of the excavation in the longitudinal direction. Because the ground standard segments of this excavation may arise from overexcava-
settlements at JS12-1 to JS12-5 and JS13-1 to JS13-5 were not tion. Further examination showed that the measured ground settle-
surveyed during excavation of Section 4, their measurements were ments at Section 3 were slightly greater than those of Section 2 with

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / JANUARY 2012 / 85

J. Geotech. Geoenviron. Eng. 2012.138:69-88.

 0.5
Field data (this study)
the basis of the excavation projects in soft clays in San Francisco;
Field data (Wang et al. 2005) Oslo, Norway; and Chicago. They observed that most of the rela-

hm
tionships fell between δ vm ¼ 0:5δ hm and δ vm ¼ δ hm . On the basis of

δ
Mana and Clough (1981)

=
0.4

vm
Wang et al. (2005) the collected data from six metro-station excavations in Shanghai,

δ
δ hm
This study 0.8 Wang et al. (2005) found that most of the data lay between δ vm ¼
=
δ vm 0:2δ hm and δ vm ¼ 0:6δ hm . Fig. 21 summarizes δ vm =H e and δ hm =H e

hm
δ
0.3
.6δ h
δvm / He (%)

collected from Wang et al. (2005) and this project, where H e = final

1.2
m
=0

=
δ vm excavation depth. For this excavation project, the data were

vm
δ
bounded by δ vm ¼ 0:8δ hm and δ vm ¼ 1:2δ hm , which was greater
0.2 .5 δ h m than those of Mana and Clough (1981) and Wang et al. (2005).
=0
δ vm The greater δ vm =H e and the comparatively smaller δ hm =H e at this
Downloaded from ascelibrary.org by WASHINGTON UNIV IN ST LOUIS on 08/05/13. Copyright ASCE. For personal use only; all rights reserved.

site may be attributed to its long shape because (1) a larger-scale

δ hm
0.1 δ vm = 0.2 stress field was induced in the soils surrounding a longer excavation
pit and (2) a longer multipropped stiff concrete diaphragm wall
could better restrain the wall deflections.
0.0
0.0 0.1 0.2 0.3 0.4 0.5 Settlements of Adjacent Buildings and Utility Pipelines
δhm / He (%)
The monitored building and utility settlements during construction
Fig. 21. Relationship between the maximum wall deflections and the were less than the allowable values, and no damage to buildings or
maximum ground settlements at the end of excavation utilities was reported during construction. Because all the high-rise
buildings and utility pipelines settled similarly during construction,
only typical data are presented in Fig. 22. The surveyed data
a similar geometry. This difference should be primarily attributable showed that for the one- and two-story brick structures supported
to the excavation sequence. The excavation of Section 3 was not by shallow foundations, the measured maximum total settlements
initialized until the permanent structures were completed at Sec- were no more than 10 mm and the maximum differential settle-
tions 1, 2, and 4 (i.e., the diaphragm walls were almost fully fixed ments were less than 6 mm. In contrast, the high-rise buildings rest-
in these sections). Therefore, the soil movements around the pit due ing on deep foundations experienced much smaller and more
to the excavation of Section 3 would be concentrated on Section 3. uniform settlements. Their maximum settlements were approxi-
mately 2 mm, and the maximum differential settlements were less
Relationship between Maximum Wall Deflections and
than 1 mm. This discrepancy can be interpreted by the following
Maximum Ground Settlements
facts: (1) structural integrity of those high-rise concrete structures
Mana and Clough (1981) investigated the relationship between the was much better than the one- and two-story brick buildings; (2) the
maximum wall deflections at the completion of the excavation to high-rise buildings were supported by ACIP piles that were toed
the final level, δ hm , and the maximum ground settlements, δ vm , on through the upper soft clay layers into the deep hard bearing
6/11/2007

6/21/2007

7/11/2007

7/21/2007

7/31/2007

8/10/2007

8/20/2007

8/30/2007

9/19/2007

9/29/2007

10/9/2007
6/1/2007

7/1/2007

9/9/2007

8
Date (mm/dd/year)
6 F1 F2
4 F3 F4
2
Settlement (mm)

F5 F6
0
-2 F7 F8
-4
-6
-8
-10
-12 Gas-pressure control station
-14
6/11/2007

6/21/2007

7/11/2007

7/21/2007

7/31/2007

8/10/2007

8/20/2007

8/30/2007

9/19/2007

9/29/2007

10/9/2007
6/1/2007

7/1/2007

9/9/2007

8
6 Date (mm/dd/year) F13
4 F14
2
Settlement (mm)

F15
0
-2 F16
-4 F17
-6 F18
-8
-10
Le-Kai building
-12
-14

Fig. 22. Typical settlement development of the adjacent buildings

86 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / JANUARY 2012

J. Geotech. Geoenviron. Eng. 2012.138:69-88.

 stratum; and (3) the high-rise buildings were located relatively 6. Ground settlements increased with the progress of excavation
further from the excavation. to a deeper layer. With the construction of floor slabs, the
Similar to the observed settlements of the high-rise buildings, ground settlements tended to stabilize over time. Because of its
the nearby utility pipelines experienced relatively small and uni- long length, this excavation caused a farther extension of the
form settlements during construction. The observed maximum ground settlement trough than those reported in literature (e.g.,
settlements were less than 10 mm, and the maximum differential Clough and O’Rourke 1990; Hsieh and Ou 1998), but it was
settlements were less than 2 mm. consistent with the observation in the Boston CA/T project
(Hashash et al. 2008). The measured maximum ground settle-
ments, δ vm , attributable to excavation lay between δ vm ¼
Conclusions 0:03%H and δ vm ¼ 0:15%H at Sections 1 and 4, whereas it
fell between δ vm ¼ 0:03%H and δvm ¼ 0:18%H at the overex-
Downloaded from ascelibrary.org by WASHINGTON UNIV IN ST LOUIS on 08/05/13. Copyright ASCE. For personal use only; all rights reserved.

Through a long-term comprehensive instrumentation program, this cavated Sections 2 and 3.

study investigated the performance of a long sword-shaped deep 7. In this project, most of the measured maximum ground settle-
excavation in soft clay constructed by the cut-and-cover technique ments, δ vm , at the end of excavation ranged between δ vm ¼
in the metropolitan area of Shanghai, China. On the basis of the 0:8δ hm and δ vm ¼ 1:2δ hm , which were significantly greater than
analyses of the field data, the following major conclusions can be those reported in literature. These greater δ vm =H e and compara-
drawn: tively smaller δhm =H e may be attributed to its long and narrow
1. Although the central standard segments of this metro station shape because (1) a larger-scale stress field was induced in the
were overexcavated, the magnitudes of wall deflections, soils surrounding a longer excavation pit and (2) a long and nar-
ground settlements, and settlements of the adjacent buildings row multipropped excavation supported by concrete diaphragm
and utilities were relatively small and less than the specified walls could better restrain the wall deflections.
protection levels. This primarily was the result of three bene-
ficial factors: (1) the use of rigid concrete diaphragm walls;
(2) quick excavation; and (3) the rapid cast of rigid concrete Acknowledgments
floor slabs following completion of the excavation. Many organizations contributed to the success of this project, and
2. For cut-and-cover excavations in the sensitive and creep special thanks are owed to Shanghai Geotechnical Investigations &
Shanghai soft clay, (1) excavation should be finished within Design Institute Co., Ltd. for providing the detailed information
a short time, followed by propping struts promptly and prop- of the soil conditions at the project site; Shanghai Institute of
erly; (2) stiff concrete base slabs should be constructed right Geological Science (SIGS) for field data collection; and Dr. Ye Lu
after the excavation, and the middle slabs should be cast as of Shanghai University for her substantial contributions to improve
soon as possible; and (3) overexcavation should be treated the quality of this paper. The financial support from the National
with caution and avoided at deeper layers. Otherwise, long Natural Science Foundation of China (NSFC Grant No. 50908172
construction duration, long-time unpropped wall exposure, or and No. 51109125) and Kwang-hua Funds for College of Civil
overexcavation would induce significant retaining wall and Engineering at Tongji University are gratefully acknowledged.
ground movements. The construction of base slabs and middle Finally, the two anonymous reviewers, the editor Dr. Iraj Noorany,
slabs helps stabilize the soil-structure system by suppressing and the editorial board members are sincerely appreciated for their
postexcavation wall deflections and ground movements. great comments and suggestions, which substantially improved the
3. If excavation was executed according to design, the measured presentation of this paper.
maximum wall deflections lay between δh ¼ 0:02%H and
δ hm ¼ 0:13%H. Until H reached 12 m, δ h at the overexcavated
sections followed the same developing patterns as that at the References
normally excavated sections. However, when excavation pro-
ceeded to a greater depth, δ h at the overexcavated sections Blackburn, J. T., and Finno, R. J. (2007). “Three-dimensional responses
abruptly jumped into the range between δh ¼ 0:13%H and observed in an internally braced excavation in soft clay.” J. Geotech.
δ h ¼ 0:23%H. Geoenviron. Eng., 133(11), 1364–1373.
4. For this project, if excavation were carried out exactly as Boscardin, M. D., Cording, E. J., and O’Rourke, T. D. (1978). “Case
designed, the location of the maximum wall deflection along studies of building behavior in response to adjacent excavation.” Final
the depth, H m , fell between H m ¼ H
6 m and H m ¼ H. Rep. Prepared for U.S. Dept. of Transportation Rep. No. UMTA-IL-06-
0043-78-2, University of Illinois, Urbana, IL.
However, if the pit was overexcavated, H m lay between H m ¼
Clough, G. W., and O’Rourke, T. D. (1990). “Construction induced move-
H
6 m and H m ¼ H þ 3 m. This means overexcavation in- ments of in-situ walls.” Geotechnical special publication: Design and
creased the chance that the maximum wall deflections occurred performance of earth retaining structures (GSP 25), ASCE, Reston,
below the excavation surfaces. VA, 439–470.
5. For this long and narrow rectangular excavation, no obvious Clough, G. W., Smith, E. M., and Sweeney, B. P. (1989). “Movement
corner effect was observed. The comparisons of this study control of excavation support system by iterative design.” Proc., Foun-
and those from Wang et al. (2005) and Xu (2007) demonstrate dation Engineering: Current principals and practices, Vol. 2, ASCE,
that a long and narrow multipropped excavation supported by New York, 869–884.
concrete diaphragm walls is capable of constraining lateral Finno, R. J., Atmatzidis, D. K., and Perkins, S. B. (1989). “Observed
wall movements and associated ground movements during performance of a deep excavation in clay.” J. Geotech. Eng., 115(8),
1045–1064.
excavation. However, a longer excavation may lead to a much
Finno, R. J., Blackburn, J. T., and Roboski, J. F. (2007). “Three-
wider distribution of ground settlements behind the pit because dimensional effects for supported excavations in clay.” J. Geotech.
it will induce a stress field of wide range. The behaviors of Geoenviron. Eng., 133(1), 30–36.
deep excavations in Shanghai soft clay are affected not only Finno, R. J., and Bryson, L. S. (2002). “Response of building adjacent to
by construction duration (time effects), but also by their geo- stiff excavation support system in soft clay.” J. Perform. Constr. Facil.,
metries (space effects). 16(1), 10–20.

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / JANUARY 2012 / 87

J. Geotech. Geoenviron. Eng. 2012.138:69-88.

 Finno, R. J., and Calvello, M. (2005). “Supported excavations: Observa- structures.” U.S. Dept. of Transportation Rep. No. DOT-TST-76T-22,
tional method and inverse modeling.” J. Geotech. Geoenviron. Eng., University of Illinois, Urbana, IL.
131(7), 826–836. O’Rourke, T. D., and McGinn, A. J. (2006). “Lessons learned for ground
Finno, R. J., and Harahap, I. S. (1991). “Finite element analyses of HDR-4 movements and soil stabilization from the Boston Central Artery.”
excavation.” J. Geotech. Eng., 117(10), 1590–1609. J. Geotech. Geoenviron. Eng., 132(8), 966–989.
Finno, R. J., and Roboski, J. F. (2005). “Three-dimensional responses of Osouli, A., Hashash, Y. M. A., and Song, H. (2010). “Interplay between
a tied-back excavation through clay.” J. Geotech. Geoenviron. Eng., field measurements and soil behavior for capturing supported excava-
131(3), 273–282. tion response.” J. Geotech. Geoenviron. Eng., 136(1), 69–84.
Finno, R. J., Voss, F. T., Jr., Rossow, E., and Blackburn, J. T. (2005). Ou, C. Y., and Chiou, D. C. (1993). “Three-dimensional finite element
“Evaluating damage potential in buildings affected by excavations.” analysis of deep excavation.” Proc., 11th Southeast Asia Geotechnical
J. Geotech. Geoenviron. Eng., 131(10), 1199–1210. Conf., Institution of Engineers Malaysia, Kuala Lumpur, Malaysia,
Fu, Y. B., Zhu, H. H., and Yang, J. (2009). “Experimental study of 769–774.
Ou, C. Y., Hsieh, P. G., and Chiou, D. C. (1993). “Characteristics of ground
Downloaded from ascelibrary.org by WASHINGTON UNIV IN ST LOUIS on 08/05/13. Copyright ASCE. For personal use only; all rights reserved.

time-dependent properties and pore water pressure of soft soil under

unloading.” Chin. J. Rock Mech. Eng., 28(Supp. 1), 3244–3249 (in surface settlement during excavation.” Can. Geotech. J., 30(5),
Chinese). 758–767.
Hashash, Y. M. A. (1992). “Analysis of deep excavations in clay.” Ou, C. Y., Liao, J. T., and Lin, H. D. (1998). “Performance of diaphragm
Ph.D. thesis, Dept. of Civil Engineering, Massachusetts Institute of wall constructed using the top-down method.” J. Geotech. Geoenviron.
Technology, Cambridge, MA. Eng., 124(9), 798–808.
Hashash, Y. M. A., Marulanda, C., Ghaboussi, J., and Jung, S. (2006). Ou, C. Y., and Shiau, B. Y. (1998). “Analysis of the corner effects on
“Novel approach to integration of numerical modeling and field obser- excavation behaviors.” Can. Geotech. J., 35(3), 532–540.
vations for deep excavations.” J. Geotech. Geoenviron. Eng., 132(8), Ou, C. Y., Shiau, B. Y., and Wang, I. W. (2000). “Three-dimensional
1019–1031. deformation behavior of the Taipei National Enterprise Center (TNEC)
Hashash, Y. M. A., Osouli, A., and Marulanda, C. (2008). “Central artery/ excavation case history.” Can. Geotech. J., 37(2), 438–448.
tunnel project excavation induced ground deformations.” J. Geotech. Peck, R. B. (1969). “Deep excavation & tunneling in soft ground. State-
Geoenviron. Eng., 134(9), 1399–1406. of-the-art-report.” Proc., 7th Int. Conf. of Soil Mechanics and Founda-
Hashash, Y. M. A., and Whittle, A. J. (1996). “Ground movement predic- tion Engineering, International Society for Soil Mechanics and
tion for deep excavations in soft clay.” J. Geotech. Eng., 122(6), Geotechnical Engineering, Mexico City, 225–281.
474–486. Roboski, J., and Finno, R. J. (2006). “Distribution of ground movements
Hashash, Y. M. A., and Whittle, A. J. (2002). “Mechanisms of load transfer parallel to deep excavations in clay.” Can. Geotech. J., 43(1), 43–58.
and arching for braced excavations in clay.” J. Geotech. Geoenviron. Schuster, M., Kung, G. T. C., Juang, C. H., and Hashash, Y. M. A. (2009).
Eng., 128(3), 187–197. “Simplified model for evaluating damage potential of buildings adja-
Hsieh, P. G., and Ou, C. Y. (1998). “Shape of ground surface settlement cent to a braced excavation.” J. Geotech. Geoenviron. Eng., 135(12),
profiles caused by excavation.” Can. Geotech. J., 35(6), 1004–1017. 1823–1835.
Kung, G. T. C., Juang, C. H., Hsiao, E. C. L., and Hashash, Y. M. A. (2007). Shanghai Municipal Engineering Authority. (2000). “Specification for
“Simplified model for wall deflection and ground-surface settlement excavation in Shanghai metro construction.” SZ-08-2000, Shanghai,
caused by braced excavation in clays.” J. Geotech. Geoenviron. China (in Chinese).
Eng., 133(6), 731–747. Son, M., and Cording, E. J. (2005). “Estimation of building damage due to
Lee, F. H., Yong, K. Y., and Quan, K. C. N. (1998). “Effects of corners in excavation-induced ground movements.” J. Geotech. Geoenviron. Eng.,
strutted excavations: Field monitoring and case histories.” J. Geotech. 131(2), 162–177.
Geoenviron. Eng., 124(4), 339–349. Tan, Y., and Li, M. W. (2011). “Measured performance of a 26 m deep top-
Liu, G. B., Ng, C. W. W., and Wang, Z. W. (2005). “Observed performance down excavation in downtown Shanghai.” Can. Geotech. J., 48(5),
of a multipropped excavation in Shanghai soft clays.” J. Geotech. 704–719.
Geoenviron. Eng., 131(8), 1004–1013. Terzaghi, K. (1943). Theoretical soil mechanics, Wiley, New York.
Liu, J. H., and Hou, X. Y. (1997). Excavation engineering handbook, Terzaghi, K., and Peck, R. B. (1967). Soil mechanics in engineering
Chinese Construction Industry, Beijing (in Chinese). practice, 2nd Ed., Wiley, New York, 729.
Long, M. (2001). “Database for retaining wall and ground movements due Wang, Z. W., Ng, C. W. W., and Liu, G. B. (2005). “Characteristics of wall
to deep excavations.” J. Geotech. Geoenviron. Eng., 127(3), 203–224. deflections and ground surface settlements in Shanghai.” Can. Geotech.
Mana, A. I., and Clough, G. W. (1981). “Prediction of movements for J., 42(5), 1243–1254.
braced cuts in clay.” J. Geotech. Eng. Div., 107(6), 759–777. Whittle, A. J., Hashash, Y. M. A., and Whitman, R. V. (1993). “Analysis of
Moormann, C. (2004). “Analysis of wall and ground movements due to deep excavation in Boston.” J. Geotech. Eng., 119(1), 69–90.
deep excavation in soft soils based on a new worldwide database.” Soils Wong, L. W., and Patron, B. C. (1993). “Settlements induced by deep
Found., 44(1), 87–98. excavation in Taipei.” Proc., 11th Southeast Asia Geotechnical Conf.,
O’Rourke, T. D. (1981). “Ground movements caused by braced excava- Institution of Engineers Malaysia, Kuala Lumpur, Malaysia, 787–791.
tions.” J. Geotech. Eng. Div., 107(9), 1159–1178. Xu, Z. H. (2007). “Deformation behavior of deep excavations supported by
O’Rourke, T. D., Cording, E. J., and Boscardin, M. (1976). “The ground permanent structures in Shanghai soft deposit.” Ph.D. thesis, Shanghai
movements related to braced excavation and their influence on adjacent Jiao Tong Univ., Shanghai, China (in Chinese).

88 / JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING © ASCE / JANUARY 2012

J. Geotech. Geoenviron. Eng. 2012.138:69-88.