PERFORMANCE OF DIAPHRAGM WALL CONSTRUCTED USING

TOP-DoWN METHOD

By Chang-Yu Ou; Member, ASCE, Jui-Thng Liao/ and Horn-Da Lin,3 Member, ASCE

ABSTRACT: This paper presents the performance of an excavation using the top-down construction method.
Strut loads, wall displacement, wall bending moment, ground surface settlement, pore-water pressure and bottom
heave were measured. Results obtained from those observations are correlated with the construction activities.
Field observations indicate that strut loads, wall displacement, and ground surface settlement correspond to those
reported in the literature. Bending moments of the wall are studied based on the results of the rebar strain gauge
and inclinometer measurements. The supported waH and the soil near the wall have a deep inward movement,
which accounts for the magnitude of the lateral earth pressure acting on the wall. The behavior of the supported
wall and soil over time is consistent with the variation of pore-water pressure during excavation. Analysis of
Downloaded from ascelibrary.org by Ryerson University on 06/17/14. Copyright ASCE. For personal use only; all rights reserved.

excavations in soft clay should therefore consider the creep factors andlor pore-water pressure dissipation.

INTRODUCTION story building and has five basement levels. The site occupies
Deep excavation in soft clay normally causes a large wall an area of about 3,500 m2 , as shown in Fig. 1. This paper
deflection and large ground surface settlement. Excessive discusses strut loads, wall displacements, wall bending mo-
ground surface settlement frequently damages the adjacent ments, ground movements, pore-water pressures, and bottom
property in urban areas. The characteristics of wall deforma- heaves associated with construction.
tion and ground movement must be thoroughly understood to
GROUND CONDITIONS
protect the adjacent properties. Many investigators have pro-
vided studies of case histories, e.g., Karlsrud (1981), Mana As shown in Fig. 2, the subsurface conditions at the site
and Clough (1981), and au et al. (1993), to understand these consist of six layers of alternating silty clay and silty sand
deformation characteristics. Furthermore, Finno et al. (1989) deposits overlying a thick gravel formation. The first and sec-
developed an extensive monitoring program on the Chicago ond layers are a 5.6-m-thick silty clay (CL) and a 2.4-m-thick
subway excavation HDR-4 project. The observation items in- silty sand (SM), respectively. The third layer is a 25-m-thick
cluded surface and subsurface three-dimensional soil move- silty clay (CL), and it is mainly this layer that affects the
ments, pore-water pressures, sheet-pile deformations, and strut excavation behavior in this case. The liquid limit for this layer
loads. The strength and stress-strain behaviors of the soil at of clay ranges from 29 to 39, and the plastic index ranges
the site were also studied thoroughly (Finno and Nerby 1989). from 9 to 19. The silt and clay contents are in the range of
In that case, high pore-water pressures were monitored during 40% to 55% and 45% to 60%, respectively. The coefficient of
the sheet-pile driving, and very large ground movements were permeability (k) from one-dimensional consolidation tests is
observed. This study enhances the knowledge of braced ex- around 4 X 10- 6 cm/s. The coefficient of consolidation (c,,)
cavations in soft clay. ranges between 3 X 10- 3 cm% and 1.1 X 10-3 cm2/s. The
Most of the cases reported in the literature were constructed fourth and fifth layers are a 2-m-thick medium dense fine sand
using the bottom-up excavation method. This method uses and 2.5-m-thick medium to stiff clay. The sixth layer is an 8.0-
temporary steel struts to support the excavation wall. Instal- m-thick medium to dense silt or silty sand. A gravel formation
lation of the struts requires a relatively short period of time is located 46 m below the ground surface and has a standard
(generally one to two weeks), depending on the size of the penetration resistance N value greater than 328 blows/m.
excavation. The displacement behavior of the supported wall Fig. 3 shows variation of water content, effective overbur-
and soil may change little during the period of strut installation den pressure, and preconsolidation pressure with depth. The
because the pore-water pressure in the clay typically does not preconsolidation pressure appears to correspond well with the
dissipate quickly. On the other hand, the top-down excavation water content. The undrained shear strength was obtained from
method uses concrete floor slabs to support the wall and some- unconsolidated-undrained (UU) triaxial tests, field vane shear
times requires long periods of time between two successive (FV) tests, triaxial Ko-consolidated undrained compression
excavation stages to construct the floor slab. Dissipation of (CKoU - AC) tests, and extension tests (CKoU - AE), as
excess pore-water pressure or creep behavior in the soil can shown in Fig. 4. The drained friction angle (4)') equals 30°.
have significant effects on the deformation behavior of the In addition, three cone penetration tests with pore-water pres-
wall and soil. For these reasons, a comprehensive monitoring sure measurement (CPTU) were performed at the site. The
system was installed on the Taipei National Enterprise Center variation in undrained shear strength computed using the for-
(TNEC) excavation project, which was completed using the mula provided by Roberston and Campanella (1989) from one
top-down construction method. The TNEC structure is an 18- of the CPTU tests is also shown in Fig. 4, in which the em-
pirical cone factor, N k , is equal to 15.
'Prof., Dept. of Constr. Engrg., National Taiwan Univ. of Sci. and Because creep behavior of the silty clay may affect exca-
Techno!., Taipei, Taiwan, Republic of China. vation behavior, a series of triaxial compression and lateral
'Grad. Student, Dept. of Constr. Engrg., National Taiwan Univ. of Sci.
and Techno!., Taipei, Taiwan, Republic of China. extension creep tests was conducted. Both types of these were
'Prof., Dept. of Constr. Engrg., National Taiwan Univ. of Sci. and consolidated isotropically prior to the creep test. Singh and
Techno!., Taipei, Taiwan, Republic of China. Mitchell's parameters (1968), such as Alt m, and a, can be
Note. Discussion open until February I, 1999. To extend the closing obtained from the regression analysis of the test results. The
date one month, a written request must be filed with the ASCE Manager results of the tests indicate that the parameters obtained from
of Journals. The manuscript for this paper was submitted for review and lateral extension creep tests are close to those from compres-
possible publication on February 8, 1996. This paper is part of the Jour-
nal of Geotechnical and Geoenvironmentol Engineering, Vo!. 124, No. sion creep tests. This finding implies that the parameters from
9, September, 1998. ©ASCE, ISSN 1090-0241/98/0009-0798-0808/ the compression creep test can be used in the excavation anal-
$8.00 + $.50 per page. Paper No. 12572. ysis while considering the creep effects, in which lateral de-
798/ JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING / SEPTEMBER 1998

J. Geotech. Geoenviron. Eng. 1998.124:798-808.

 P R WATER CONTENT (515)
N 0 ...--r~"'T""'120r-,-.,-r-r....
00'h--r-r"T"':"r.- 30r-,-.,-r-r...,40r-,--r-r-r.,50
(j) _
WATER CONTENT
. . . .A
PRE-CONSOLIDATION PRESSURE
----- EFFECTIVE OVERBURDEN PRESSURE
AA ..

10

- AA
....
A'"
Rl[-lCl E 20 A.A
A _
'-'
I

-
l-
A • Ap5 e..
P6 ~ 30
AM A
• Inclinometer
• •
Downloaded from ascelibrary.org by Ryerson University on 06/17/14. Copyright ASCE. For personal use only; all rights reserved.

....
tJ. Heave Gauge
A Piezometer

• Combined Earth/Water Pressure Cell

IB Rebar Strainmeler

FIG. 1. Instrumentation Plan

o 5 lO(m)
~hE'
40
- A

50 i--'-...L....J.-..l.-+-~-'-...L....If-l-....I-L-J..-+-J.-..l.--.L--'--l-.L..!-L-J..-l
IF o 100 200 300 400 500
o Slrul EFFECTIVE OVERBURNDEN PRESSURE (kPo)
CL-2.8m F
5 CL-4.9m FIG. 3. Variations of Water Content, Preconsolldatlon Pres-
sure, and Effective Overburden Pressure with Depth
to
Su (kPa)
15 50 100 150 200

Diaphragm "an
30 SI-4 SI-3
A
10
Incllnomeler
A Heave Gaule
• Selllemenl Poinl
A Piezomeler
...... Combined Eerlb/Weler
Pressure cen
I20
SI-2 SI-I 1-1 I-
o~(m) 0...
W
SCALE
o
FIG. 2. Instrumentation Section

formation behavior is largely involved. Based on the test re-

30
- - CPTU o
sults, average AI> m, and a are equal to about 0.0037, 0.94, ccccc FIELD VANE
~~~~~ UU
and 5.04, respectively. Because m is smaller than 1.0, the in ••••• CKoU-AC
situ silty clay can be classified as having low to medium creep 0 0 0 0 0 CKoU-AE

potential. The coefficient of secondary compression (ell)

ranged from 0.48% to 0.6%. According to Mesri (1973), the 40 - ' - - - - - - - - - - - - - - - - - - - '
soil can thus be classified as having low to medium secondary
FIG. 4. Variation of Undrained Shear Strength
compressibility.

CONSTRUCTION SEQUENCE AND from 2.0 to 7.5 m away from the wall and at 1.5 to 3.0 m
INSTRUMENTATION spacings at distances of 7.5 to 49 m from the wall. This ar-
rangement allowed the settlement profile to be continuously
Fig. 1 shows the excavation site along with the monitoring measured.
locations. As shown in this figure, the shape of the excavation Fig. 2 also indicates that 16 combined earthlwater pressure
site is slightly irregular. A 90-cm-thick and 35-m-deep dia- cells (eight on the back and eight on the front) were installed
phragm wall was used as the earth-retaining structure. The on a panel of the diaphragm wall in the main observation
final excavation depth was 19.7 m, and was completed using section at eight different depths. Six piezometers were also
the top-down construction method. installed inside the excavation. Outside the excavation, pie-
As indicated in Fig. 2, inclinometer casings I-I, 1-2, and 1- zometers were installed at various depths at five locations (P3
3 were installed in the wall, and SI-l to SI-3 were installed to P7), and P4 to P7 were arranged along the main observation
along the main observation section outside the excavation section.
zone. In the main observation section, the settlement measure- Stresses in the reinforced steel of the supported wall were
ment points were positioned at 1.0 m spacings at a distance measured by rebar strain gauges. Sixteen rebar strainmeters
JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING / SEPTEMBER 1998/799

J. Geotech. Geoenviron. Eng. 1998.124:798-808.

 TABLE 1. Excavation Sequence of TNEC Case History TABLE 2. Observed Strut Loads at Final Stage
Interval Distance Strut Tributary Strut load per
Stage (d) Construction activities from corner' load width unit width
(1 ) (2) (3) Number (m) (kN) (m) (kN/m)
1-89 Construct diaphragm wall (1 ) (2) (3) (4) (5)
90-155 Construct pile foundation SA-l 23 (W.C.) 853.5 4.0 213.4
1 156-162 Excavate to elevation of -2.80 m SA-2 45 (W.C.) 2,324 3.1 749.5
2 164-169 Install H300 X 300 X 10 X 15 sections at first SA-3 38 (E.C.) 1,813 2.5 725.0
strut level (elevation of -2.0 m), preload = SA-4 20 (E.C.) 1,525 3.7 412.0
784.8 kN per strut
181-188 Excavate to elevation of -4.9 m ·W.C. denotes western corner; E.C. denotes eastern comer.
3
4A 217 Cast floor slab (BIF) at elevation of -3.5 m
4B 222-328 Demolish first level of the strut and cast ground
level of slab After the floor slab, B IF, was cast, the strut was demolished.
5 233-255 Excavate to elevation of -8.6 m The force in the strut was released, which subsequently caused
6 279 Cast floor slab (B2F) at elevation of -7.1 m the wall to move inward. Under such a circumstance, the
Excavate to elevation of -ll.8 m
Downloaded from ascelibrary.org by Ryerson University on 06/17/14. Copyright ASCE. For personal use only; all rights reserved.

7 318-337 safety of the supported system was checked. Field observation

8 352 Cast floor slab (B3F) at elevation of 10.3 m results also indicate that demolishing the strut also had only a
9 363-378 Excavate to elevation of -15.2 m
10 400 Cast floor slab (B4F) at elevation of -13.7 m slight effect on the outward wall deflection.
llA 419-423 Excavate to elevation of -17.3 m (center strip) At the second strut level, the horizontal strut spacing varied
12A 425-429 Install H400 X 400 X 13 X 21 sections at sec- from 2.5 to 6.0 m, and averaged 3.0 m. The average nominal
ond strut level (elevation of -16.5 m), preload axial stiffness of the strut per unit width was 64,363 kN/m/m.
= 1,177 kN per strut (center strip) The preloading force was 1,177 kN per strut, and was about
llB 430-436 Excavate to elevation of -17.3 m (side strips) 392 kN/m per unit width in the central region. The strut was
12B 437-444 Install H400 X 400 X 13 X 21 sections at sec-
ond strut level (elevation of -16.5 m), preload demolished at day 528 after the floor slab, B5F, was com-
= 1,177 kN per strut (side strips) pleted. Field observation results indicate that preloading and
13 445-460 Excavate to elevation of -19.7 m demolishing the strut had no effect on the wall deflection. This
457 Complete the superstructure may imply that the preloading force cannot overcome the lat-
14 464-468 Cast the foundation slab eral earth pressure acting on the wall at this stage.
15 506-520 Cast floor (B5F) slab at elevation of -17.1 m The strain gauge was not installed on the first level of the
16 528 Demolish second level of the strut
strut because the load in this level of the strut was not expected
to be large. For the second strut level, four sets of strain gauges
(eight on the back and eight on the front) were installed in the were installed on the struts (SA-I, SA-2, SA-3 and SA-4) in
main observation section, as shown in Fig. 1. the north-south direction. Table 2 lists the distance to the cor-
As listed in Table I, two levels of the struts were installed ner for these struts and the loads in the struts at the final stage
at some stages. To monitor the strut loads, vibrating wire strain of excavation. Since struts SA-I and SA-4 were close to the
gauges were attached on both sides of the web of the steel corners, the corner effects apparently have a significant effect
strut. The load in the strut could then be calculated on the on the strut loads. Therefore, the data were excluded from the
basis of the measured strain. following analyses.
Table 1 also presents the time sequence of construction ac- As indicated in Table 2, the loads in struts SA-2 and SA-3
tivities for this project. Construction of the concrete diaphragm per unit width, considered as the plane strain condition, were
wall commenced on August 13, 1991. For the convenience of 750 kN/m and 725 kN/m, respectively. The average value was
describing construction activities, the construction days are 737.5 kN/m. Because the apparent earth pressure envelopes
numbered starting from this day. As revealed in Table I, stages available are only for uniform soil, questions may arise re-
1, 3, 5, 7, 9, llA, lIB, and 13 represent excavation stages, garding how to treat the subsoil conditions shown in Fig. 2.
and stages 2, 4,6,8, 10, 12A, 12B, 14, and 15 represent stages Fig. 5 shows the apparent earth pressure envelopes proposed
for strut installation or concrete floor slab construction. To by Peck (1969), considering that the undrained shear strength
reduce the wall displacement, the excavation was carried out was 58.6 kPa at the position of strut installation, i.e., an ele-
with zoning after the concrete floor slab (B4F) located at an vation of -16.5 m, and the average value of the soil within
elevation of -13.7 m was cast. During the zoned excavation, the depth of excavation, i.e., 19.7 m (Fig. 4), was 34.5 kPa.
the center strip of the site (area PQRT in Fig. 1), was first This figure reveals that the observed value was close to the
excavated and then reinforced with the strut. Thereafter, the computed value based on the average undrained shear strength
side strips of the site were excavated, and supported with the in engineering practice.
strut.
DISPLACEMENT OF DIAPHRAGM WALL
STRUT PERFORMANCE Fig. 6 shows the horizontal displacement of the diaphragm
For the first strut level, the horizontal strut spacing varied wall at I-I, 1-2, and 1-3 at the final construction stage (stage
from 6.0 to 11.0 m, and was 8.0 m on average. The average 13). As indicated in this figure, the wall displacements at these
nominal axial stiffness per unit width for the first strut level inclinometers are similar. This finding exhibits that the wall at
was 14,980 kN/m/m. The preloading force was 784.8 kN per 1-1, 1-2, and 1-3 is in the plane strain condition.
strut, and was about 98.1 kN/m per unit width in the central Figs. 7(a and b) show displacements of the diaphragm wall
region (i.e., plane strain condition). Because the preloading and transverse horizontal ground deformations in the main ob-
force can push the supported wall outward toward the unex- servation section after the completion of each excavation
cavated side and may damage adjacent buildings, the magni- stage. As shown in Table I, the site was first excavated down
tude of wall movement usually is a concern in excavation de- to 2.8 m below the ground surface (elevation -2.8 m). The
sign. Field observation results indicate that the preloading strut was not installed and the floor slab was not constructed
force appeared to have an insignificant effect on the wall de- at this time. At the subsequent stage, the struts were installed
flection. This insignificant effect may be due to the relatively at an elevation of -2.0 m; the site was then excavated down
low preloading forces in the struts. to 4.9 m. The nominal axial stiffness per unit width, 14,980
800 I JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING I SEPTEMBER 1998

J. Geotech. Geoenviron. Eng. 1998.124:798-808.

 IF r- BXCAVATION ZONB
51-4 51-3 51-2 51-1 11
o
.. .",.,,-;7 o ,'u
:~SI-( I) f
l.V SI-J ~~ l
)SI- :\ 'ff:: PI

:': \,~\1 I-I

.; I.'
',' I

,.,,-- '" :! !j' : SI-1 .

(5;-
.... ....
'" ~~t1
,I /'. i( i: 1
'j I"
i!l 1\\/
,~

iI, :
,,
.;.;
-10

,."
.;.;
I
I ll!
it
I' \:
i) i i 1 I \;
/ . : : I j,
I! Ij J\ Ww
"I i ii : / : . " • I I .
. ( :: I : / "i /
I ~-20 !
I'J I'
I II
ii :, / !il/
Su=34.5kPa : Su=58.6kPa
I ~w ! '.J : I~
:. :1/
I 0_
30 j ':. : '(
I
. if : ,'.
I
I STAGE I ---
I STAGB l -'-
-12
-40
I
1
STAGB'
STAGB 7 _.-
-"-
oi sp lace.ent
(a) o -=-=._
Downloaded from ascelibrary.org by Ryerson University on 06/17/14. Copyright ASCE. For personal use only; all rights reserved.

I 5(clI)
1
SCALB
-50
-25 -15 -5 5 15 25
DISTANCE FROM THE WALL(m)
-16

51-4 51-3 51-2

0

·20. I :,---~---:-:-:~----7--
I,I f---117.7kPa---:
1
~ SI-3

I 221.8kPa I GL-19.7m -10

V
i
H (9)
245.8kPa
if
i'
(liB)
(13
P
------- Computed - - - Observed 1- 20
1
I
FIG. 5. Comparison of Observed Apparent Earth Pressure >-
0..
and Peck's Earth Pressure Envelope W
0-30

DISPLACEMENT (em)
5 10 15 -40
Displacement
(b) o _::-=:_ S(CI'l)
_50+--:"'-r-~...,.-_~~.,---.:..-+_-+_--.,.---=S:...:C:r'L:.:.B_""'---:l
-25 -15 -5 5 15 25
DISTANCE FROM THE WALL(m)
10 FIG. 7. Wall Displacement and Soli In the Main Observation
section, with the Location of 1-1 Shown Inside Excavation for II-
IU8tration Purposes: (a) Stages 1-7; (b) Stages 9-13

...... 20 the position of the floor slabs because the high axial stiffness
E
....... of the slabs prevented the wall from moving at these positions.
I Deep inward movement thus developed on the wall, with the
f-
0... maximum wall displacement occurring near the excavation
W
o 30 surface. Soil at Sl-I and SI-2 also had a deep inward type of
movement. Soil at SI-3 and SI-4 continued to behave like a
cantilever.
Deep inward movements continuously developed on the di-
aphragm wall for the subsequent construction stages. The
40 amount of wall movements increased with excavation depth.
-1-1 The maximum wall deformation for each stage occurred near
-......... 1-2
........... 1-3 the excavation surface. Soil at SI-l had a deformation pattern
similar to that of the wall at all stages. Soil at SI-2 at stage 5
50 -"- ...J began to deform inwardly at a depth below the ground surface
(deep inward movement), and the deformation became more
FIG. 6. Displacements of the Wall at Inclinometers 1-1 ,1-2, and pronounced after stage 9. Soil at SI-3 at stage lIB began to
1-3 at Final Excavation Stage deform inwardly at a depth. Soil at SI-4 behaved like a can-
tilever at all stages.
kPa, which is considered relatively low, might not be able to As shown in Figs. 7(a and b), the maximum deformations
prevent the wall movement. The wall and soil deformations at of the wall were close to those of the soil at SI-I at all stages.
these two stages behaved as a cantilever, in which the maxi- For instance, at stage 13 (excavation depth = 19.7 m), the
mum horizontal displacement occurred at the top level of the maximum wall deformation was 10.6 cm; the maximum soil
wall, as Fig. 7(a) shows. deformation at SI-l was 10.5 em. However, the horizontal de-
As the excavation proceeded to stage 5 (excavation depth = formation of the top level of the wall was obviously smaller
8.6 m), the surface level and the first level of the floor slab than that of the ground surface at SI-l at all stages. The line
were constructed. The diaphragm wall rotated with respect to of the locations of maximum deformation at all inclinometer
JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING / SEPTEMBER 1998/801

J. Geotech. Geoenviron. Eng. 1998.124:798-808.

 casings (1-1, SI-l, SI-2, SI-3, SI-4) might be the potential fail- difference to the soil strength. Higher values of the stress level
ure surface. The ratios of maximum wall deformation to ex- of the soil normally have more pronounced creep behavior.
cavation depth for stages 1 and 3 were about 0.89%, where Table 3 lists the deflection of the diaphragm wall at depths
the maximum wall deformation occurred at the top level of of 10 and 20 m below the ground surface during the period
the wall; the ratios for all other stages ranged from 0.51 to when the excavation depth remained unchanged. Since the ex-
0.57%, which are higher than the general trends found by Ou cavation after stage 10 was conducted with zoning, differen-
et al. (1993). tiating the deflection for the period when the excavation depth
Fig. 8 shows the wall movements with time at different remained unchanged from the total deflection is difficult.
depths. As shown in this figure, the wall movement increased Therefore, only deflection data from the beginning of exca-
with time while the excavation depth remained unchanged. vation to the day of measurement before stage 10 (excavation
This is because the top-down construction generally requires depth = 15.2 m) are used in the table. The table shows that
considerable time to erect the mold and to pour the concrete the accumulated deflections (8 1) at the depths of 10 and 20 m
floor slab before the next stage of excavation. Durations of 30 during the periods when the excavation depths remained un-
to 60 days were frequently encountered for the TNEC project, changed were 18.88 and 29.29 mm, respectively; the corre-
as shown in Table 1. Both soil creep and excess pore-water sponding total deflections (8 2) at these depths were 62.54 mm
Downloaded from ascelibrary.org by Ryerson University on 06/17/14. Copyright ASCE. For personal use only; all rights reserved.

pressure dissipation may partially contribute to some extent. and 80.56 mm (Fig. 8), respectively. The percentages of ac-
Fig. 9(a) shows the relationship between the maximum de- cumulated deflection during the periods when the excavation
flection rate for the wall at I-I, 1-2, and 1-3 and the excavation depths remained unchanged, 8 1/8 2 , were 30% and 36%, re-
depth. In this figure, the deflection rate (A'O/At) is defined as spectively. These percentages were fairly large compared with
the ratio of deflection increment (A'O) to the period of the ex- the immediate deflection induced by excavation.
cavation depth remaining unchanged (At). Since an inclinom- Mana and Clough (1981) examined the time effects on lat-
eter casing has many measurement points and thus has many eral wall deflection in several case histories in San Francisco
deflection rates for a given excavation depth, the maximum Bay mud. They concluded that higher creep rates are associ-
value of the deflection rate (A'O/At) was used to study the de- ated with lower factors of safety against basal heave and that
formation behavior as a function of time. As shown in this the wall deflection rate decreases rapidly with time. In their
figure, the maximum deflection rate generally increased with study, the deflection rates were in the range of 0.3 to 30 mmI
excavation depth. Except for stage 1 (excavation depth = 2.8 d, values much larger than those in this study. The reason for
m), most of the maximum deflection rates were in the range such a difference may be the different stress levels occurring
of 0.1 mm/d to 0.6 mm/d. in the soils. Sheet pile was used in Mana and Clough's case
Fig. 9(b) shows the relationship between the maximum de- histories, where the excavation depth ranged from 9.1 to 13.5
flection rate of inclinometer casings SI-1, SI-2, and SI-3 and m. The ratios of maximum lateral wall deflection to maximum
the excavation depth. As indicated in this figure, the maximum excavation depth for the case histories observed by Mana and
deflection rates at SI-l were close to those of the diaphragm Clough (1981) (0.5% to 3.0%) are generally larger than in the
wall. The maximum deflection rate at SI-4 remained nearly TNEC excavation project (0.5%). The soil near the excavation
unchanged with time. The maximum deflection rate for the zone in Mana and Clough's case histories was presumed to be
inclinometer casings decreased with increasing horizontal dis- near the failure condition. The stress levels of the soils near
tance from the diaphragm wall. This is perhaps attributed to the excavation zone in their case histories would be expected
the fact that for a given depth, the stress level of the soil to be higher than those in the TNEC case history. A higher
decreases with an increasing horizontal distance from the wall, stress level of the soils normally produces a higher potential
with the stress level defined as the ratio of the principal stress creep behavior. This may account for the larger maximum de-
flection rates in Mana and Clough's study than in this study
DISPLACEMENT (em) and why the deflection rate of the wall in the TNEC excavation
project increased with excavation depth.
o 5 10 15
o-r~~~--'--'--'-""""""'...J.....LI
BENDING MOMENT OF DIAPHRAGM WALL
A schematic diagram for computing the bending moment of
the diaphragm wall based on measurements from the rebar
10 strainmeters is presented in the insert of Fig. 10. This com-
putation assumes that the variation of stresses over a cross
section of the wall is linear. The concrete is assumed to be
capable of sustaining a tension force until the tension stress in
the concrete exceeds the allowable tension stress. The wall
20 bending moment at various stages can then be obtained on the
E
'-'
basis of this computational procedure. Fig. 10 shows the com-
I
putational results for stages 5, 9, and 13.
f-
a.. The bending moment can also be computed using the equa-
w tion M = ElIr from the curvature radius of the wall deflection
o 30
curve, where M is the bending moment of the wall, E is
Young's modulus, and I is the moment of inertia. Fig. 10 also
4.9m; doy189 shows the wall bending moments at stages 5, 9, and 13 based
4.9m; doy221 on wall displacements. This figure shows that the bending mo-
8.6m; doy256
40 .........- H= 8.6m; doy281 ments computed from the rebar strain gauges are generally
........- H= 15.2m; doy379 smaller than those from the wall deflection curve, particularly
.............. H= 15.2m; doy40 1
- H=19.7m; doy464 for the location where the maximum lateral wall deflection
........- H=19.7m; doy470 occurred and its neighboring location. This is because the
bending moment from the wall deflection curve is computed
50..l-------------' without considering cracking concrete so that the moment of
FIG. 8. Variation of Wall Displacement with Time (1-1) inertia (l) is not reduced. As a matter of fact, some cracks
8021 JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING 1 SEPTEMBER 1998

J. Geotech. Geoenviron. Eng. 1998.124:798-808.

 1.0
000001-1
AAAAA 1-2
(a)
".....,
00000 1-3
~0.8 -
'0
..........
A (10)
E (6) 0
E
W'0.6 -
~
0::
Z (jlO)

~ 0.4 - ~ .(jIO'
--.~
(30)~"O)
u (22)
Downloaded from ascelibrary.org by Ryerson University on 06/17/14. Copyright ASCE. For personal use only; all rights reserved.

W _- A (20)
-1
l.L. ~
W
---441r A (30)

Cl 0.2 - ~- A (23) (18)0 (jU)

( ) A (25) ----'- - A - (8)
(10)0 '~~lg 42
_·t~~ Q--- --(70)&'4.)(48)
-lIJ.JJ..----.-------
0.0 I I I I I I I I I I
0 2 4 6 8 10 12 14 16 18 20 22
DEPTH(m)
1.0
(10) (b)
".....,

~0.8 - ODD a D
51-1
"0 + + + + + 51-2
.......... * * * * * 51-3
E
E
(lO)a
wO.6 -
~
0::
Z
::::---~
~
(8)

~ 0.4 -
U
+ (24)
51-1 J .- ----;:.::::=-
-;:::.-~
--;....----~
W
-1
.-'- ...
l.L. (16)a _(2J2Y~-L 51-2
W .---.--'
.---...--'
Cl 0.2 _ '--.-A~
(44) ---- - ---*(')
(23)a --:::-• ...--(24)a _---- --------
(42)\l--1~~J
:n:r
(70),(44) (70)---- - -- - --
(36)1~~~ _ _, _ -' •••
$1-.3 L
*(4')
0.0 I I I I I I I I
0 2 4 6 8 10 12 14 16 18 2'0 22
DEPTH(m)
FIG. 9. Variation of Maximum Lateral Deflection Rate of the Wall and Soli with Time, with Numbers In Parentheses Denoting At: <a> I-
1,1-2,1-3; (b> SI-1, SI-2, SI-3

actually exist in the concrete because of the deformation of LATERAL EARTH PRESSURE ON DIAPHRAGM WALL
the concrete wall, subsequently thereby reducing the moment
of inertia. The reduction factor for the moment of inertia (R) The total lateral earth pressures and pore-water pressures
is then defined as the ratio of the moment obtained from the acting on the unexcavated side and excavated side of the wall
rebar strain gauge to the moment obtained from the inclinom- can be obtained by observing the combined earth/water pres-
eter measurement. Fig. 11 presents the variation of the reduc- sure cells. The observation details and complete monitoring
tion factor with depth for these stages. As shown in this figure, results were presented by Ou and Liao (1995). Compared with
the reduction factor decreases with excavation depth. This in- the theoretical lateral at-rest (Ko) earth pressure prior to ex-
formation is valuable in the structural design of a diaphragm cavation, the initial readings were not exactly in the Ko con-
wall as well as in predicting wall deformations using numer- dition. The theoretical total lateral earth pressure at rest (O'h)
ical tools. was computed according to the equation O'h = K o O'~ + U. K o
JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING / SEPTEMBER 1998/803

J. Geotech. Geoenviron. Eng. 1998.124:798-808.

 TABLE 3. Relationship between Wall Deflections and Time actual initial pressure, the observed earth pressure at each stage
Elevation = -10m Elevation = -20 m can thus be adjusted.
Construction Figs. 12(a and b) compare the observed lateral earth pres-
Depth day' lit 1l811i. t b 118 1l8/llt b 118 sures (adjusted) and the theoretical Rankine active and passive
(m) (d) (d) (mm/d) (mm) (mm/d) (mm) pressures for stages 3 and 7 and for stages 9 and 13, respec-
(1 ) (2) (3) (4) (5) (6) (7)
tively. As shown in these figures, the lateral earth pressure at
4.9 189-233 45 0.09 4.05 0.07 3.15 the shallow depth (about 12 m) on the unexcavated side in-
8.6 256-318 63 0.07 4.41 0.10 6.30 creased with excavation depth to the at-rest Ko condition. This
1l.8 338-363 26 0.21 5.46 0.31 8.06
15.2 379-409 31 0.16 4.96 0.38 11.78
behavior can be accounted for by the fact that the wall had
deep inward movement, thereby causing it to rotate at shallow
Note: Accumulated deflection is 18.88 mrn at -10 m and 29.29 mrn depths and move to the unexcavated side (i.e., push into the
at -20 m.
"The range 189-233 denotes construction day 189 to 232, etc. soil). For soil at deeper levels, the lateral earth pressure on the
'Values are from Fig. 8. unexcavated side decreased with excavation depth to values
smaller than theoretical Rankine active earth pressure. As gen-
erally known, Rankine's earth pressure would not take into
Downloaded from ascelibrary.org by Ryerson University on 06/17/14. Copyright ASCE. For personal use only; all rights reserved.

MOMENT (kN-m) account the friction between the soil and the diaphragm wall.
-2000 -1000 0 1000 2000 3000 However, trench excavation normally causes a rugged bound-
ary surface between a trench and soil; the diaphragm wall
I
surface would be expected to be rough. Therefore, assuming

10 r:h that no friction exists between the soil and concrete wall will
overestimate the lateral earth pressure in the unexcavated side.
~ I T", T.
As shown in Fig. 12, the observed lateral earth pressure on
the excavation side at excavation stage 3 is markedly smaller

~
I20 .,," L:ON than the theoretical Rankine passive earth pressure. As de-
~ scribed in the preceding section, the maximum wall deflection
~
Q...
W j cOUP.
at this stage, which occurred near the ground surface, was 4.0
o
Co'
•
i
L.
t cm. This occurrence implies that the magnitude of wall move-
30 M-C ..·L.+C.·L ....T ..·L.
"'T.·L ..
ments does not cause failure of the soil on the excavated side.
The amount of difference between the observed and the the-
- COMPUTED rROM REBAR STRAINMETER oretical value diminished with excavation depth. This differ-
---- COMPUTED mOM INCLINOMETER
40...!.-----------'---------------J ence became small at the final stage (stage 13), except for the
FIG. 10. Comparison of Wall Bending Moments from the Re- LATERAL EARTH PRESSURE(kPo)
bar Strain Gauge and the Inclinometer
-600 -400 -200 o 200 400 600
0-+-.l-.JL-..J.--'--'--'----.l...-l-J..--'-..J...,*-L-..J.-I.......L.-...L...-.l-.JL-..J.-I.......L.---'--1
R
0.2 0.4 0.6 0.8 1.0 -4.9m (STAGE 3)
o4----'---'---'-----'---L.----'----'-----i

10
.....-.. STAGE 5
........ STAGE 9
- - STAGE 13

10 I
b::
w
20
o

I
t-
o... 20 30
W
o

LATERAL EARTH PRESSURE(kPo)

30 -400 -200 o 200 400 600

- K, CONDITION (INITIAl.)
~ • - - Ko CONOITION~OWPut~O)
- - Kp CONOITION STAGE 9~
..;;;;.
_
~M~~DI~e" ~~61' )
STAGE 1J (DBSERVEb)
FIG. 11. Variation of Reduction Factor for Wall Bending Mo-
ment at Various Stages of Excavation 10
-15.2m (STAGE 9)
was computed from laky's equation Ko = 1 - sin <1>', <1>' is
the drained friction angle; (T~ is the effective overburden pres- :r:
sure; and U is the pore-water pressure and can be assumed to b:: 20
w
be the hydrostatic condition. o
The difference between theoretical and observed values may
occur because the initial readings were affected by the jacking
operation while installing the pressure cells to make a good 30
contact with the sidewall of the trench. Strictly speaking, such
observed values cannot represent the actual lateral earth pres-
sure at rest (DiBiagio and Roti 1972; Moh and Huang 1993; FIG. 12. Observed Lateral Earth Pressure versus Theoretical
Karlsrud 1981). If the theoretical Ko pressure is treated as the Rankine Earth Pressure: (a) Stages 3 and 7; (b) Stages 9 and 13

804/ JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING / SEPTEMBER 1998

J. Geotech. Geoenviron. Eng. 1998.124:798-808.

 soil at 32.0 m below the ground surface. The soil at this po-
sition should be far from the failure condition because the wall 200 i
deflection at the same level was only 2.8 cm.
~ ISO Ir'~.--_..:­
PORE-WATER PRESSURE RESPONSE ~
w
As shown in Figs. 1 and 2, 35 electronic-type piezometers ~'oo
V1
were installed in the soil on the excavation and the unexca- V1
W
vated sides. Pore-water pressure acting on the wall can be g: 50

obtained by observing the combined earth/water pressure cells.

Some of the measurement results have been presented by Ou D

and Liao (1995). A more refined discussion is presented

herein. -SO)+:-,......:..,l.....-:so,--.-~-ri:-r-T"~':1,:f-rl-'";;2;C0~~2;;rSO;;+-'~?;;;~~
TIME (day)
Fig. 13 shows the pore-water pressure contours ob~erv~d
from the piezometers for stages 5 and 13. As shown 10 thiS FIG. 14. Variation of Pore-Water Pressure with Time
figure, excavation caused the pore-water. pres~ures to decrease
Downloaded from ascelibrary.org by Ryerson University on 06/17/14. Copyright ASCE. For personal use only; all rights reserved.

with increasing depth and with decreasmg distance from the excavation. Piezometer P4-20 was 20 m deep and 2.0 m away
wall. The pore-water pressure in the zone bounded by the hor- from the wall outside the excavation. As revealed in this fig-
izontal distance greater than 2 m from the wall and by the ure, the pore-water pressure dropped significantly during ~x­
depth approximately less than the excavation surface (d.epth = cavation. However, the pore-water pressure gradually In-
8.6 m) changed only slightly at stage 5, compared With the creased with time, except for the initial stages. This behavior
preexcavation condition. A similar response was found for is attributed to the fact that negative excess pore-water pres-
stage 13 and the other excavation stages. Thi.s can be :ac- sures dissipated with time. This observation corresponds to the
counted for the maximum lateral wall defleCtion occumng field observations on wall deflection and settlement, in which
near the excavation surface for most of the stages. The soil the deformation increased with time. Note that there was no
below the excavation surface, even far from the wall, was dewatering used for this case because the subsurface soil pro-
subjected to relatively larger shear stresses. file within the depth of excavation was composed mainly of
Fig. 13 also reveals that the pore-water pre.ssure contours cohesive soil.
inside the excavation moved down as excavatiOn proceeded.
The decreased pore-water pressure level near the center of the
site was larger than that near the wall. The co~tours at ~e GROUND SURFACE SETTLEMENT
center part of the excavation zone were nearly honzont~. ThiS Fig. 15 shows the ground surface settlement profile at the
is because the decrease of the pore-water pressure was directly main observation section at key excavation stages. As revealed
affected by releasing the overburden pressure during excava- in this figure, the settlement increased with excavation depth.
tion. The soil near the wall experienced the release of over- The maximum ground surface settlement after the completion
burden pressure, which caused the pore-water pressure to de- of the final excavation stage (stage 13, excavation depth = 19.7
crease, as well as lateral compression, which resulted in an m) was 7.8 cm. The ratios of maximum ground surface settle-
increase in pore-water pressure. This occurrence explains why ment to maximum horizontal wall deflection at all excavation
the pore-water pressures near the wall were higher than those stages ranged from 0.56 to 0.78, which are generally within
at the center part. the range of the findings by Clough and O'Rourke (1990).
Because the excavation was carried out using the top-down The maximum ground surface settlement occurred near the
construction method, considerable time was required to erect diaphragm wall at the first stage. This occurrence is perhaps
the molds and to pour the concrete floor slab before the next attributed to the fact that the wall behaved like a cantilever at
stage of excavation. The pore-water pressure would be ~x­ this stage. As the excavation proceeded, the maximum ground
pected to change during this period. Fig. 14 shows the varia- surface settlement occurred at some distance behind the wall.
tion of pore-water pressure with time. As indicated in this fig- The ratios of the location of the maximum surface settlement
ure, piezometers SP-5 and SP-6 were installed on the wall at behind the wall to the depth where the maximum lateral wall
a depth of 20 m in the unexcavated and excavated sides, re- deflection occurred are in the range of 0.63 and 0.78. This
spectively (Fig. 2). Piezometer Pl-2l was 21 m below the value is larger than observations by Nicholson (1987). Note
ground surface and 20.0 m away from the wall inside the from Fig. 7(a) that deep inward movement began to develop
on the wall after stage 3, which may account for the maximum
surface settlement occurring at some distance from the wall at
-5
these stages.
5---5---~
-10
DISTANCE (m)
E -'5 15 _
10 20 30 40 50 60

-...- -20

:c 20 -==: -2
b: -2~
W
Cl
E
-3D ~
I- -4
Z
-35 W
:::l!
w
-J -6
~
-40 -STAGE 3
-STAGE 5
Ul t+++< STAGE 7
_ _ STAGE 9
- 4~4L .L1.--LL-•.LJDLL....t-l....L•..1....u...L-.L.l-I_'-'-0L-.L..L..L.:>JoL-.L.......'"'-:,':-o.J....1....1-L"!::,O -8 ........... STAGE 118
o 2D 2
-STAGE 13
DISTANCE FROM THE WALL em)
-lO..L.--------------------'
FIG. 13. Variation of Pore-Water Pressure (Expressed In kPa)
Inside and outside the Excavation FIG. 15. Settlement Profiles Induced by Excavation

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING / SEPTEMBER 1998/805

J. Geotech. Geoenviron. Eng. 1998.124:798-808.

 As shown in Fig. 15, the soil movement extended to a con- DISTANCE (m)
siderable distance behind the wall. The soil settled 1.2 cm even 0 10 20 30 40 50 60
0
at a distance of 49 m from the wall. In other words, the influ-
ence zone caused by excavation may be more than 49 m.
-2
Buildings or facilities within the influence zone could be dam-
aged, depending on the degree of distortion and the condition E
~
of the structures. The apparent influence range (AIR) was then f- -4
Z ............ H= 4.9m; doy21 4
w
defined as the horizontal distance behind the wall to the lo- ::::0 ---.. H= 4.9m; doy233
w ~ H= 8.6m; doy259
cation where the settlement becomes uniform (Ou et al. 1993). ...J -6
.............. H= 8.6m; doy282
As shown in this figure, the AIR does not vary with excavation E
V1
...........-. H=15.2m; doy380
~- H=15.2m; doy400
depth. The magnitude of the AIR for each stage was roughly -8 ~ H=19.7m; doy464
30 m. This value is slightly larger than that calculated from .~ H=19.7m; doy505
the equation AIR = L tan (45 0 - <1>/2) (Ou et al. 1993), where -101...L------------- --..1
L is the depth of diaphragm wall and <I> is the friction angle
of the soil. FIG. 17. Variation of Ground Surface Settlement with Time
Downloaded from ascelibrary.org by Ryerson University on 06/17/14. Copyright ASCE. For personal use only; all rights reserved.

Fig. 16(a) presents the relationship between normalized set-

1.0
tlement and distance from the wall for the final excavation
'>:
stage. This figure also contains the three categories of typical 0
~0.8
excavation performance defined by Peck (1969). Field obser- E
E
vation results indicate that the relationship for the final stage
falls into category I, although Peck's categories were estab- ~0.6
lished for excavations with braced sheet-pile walls or soldier J-
~0.4
piles with laggings, which differ markedly from the case his- ::l!

tory presented herein. Fig. 16(b) shows the relationship be- '"
EO.2
tween normalized settlement (8j8....,) at various distances and VI

normalized distance from the wall (dlH) for the key stages. As 0.0
indicated in this figure, the normalized settlement also falls a 5 10 15 20
DEPTH(m)
inside the envelope identified by Clough and O'Rourke
(1990). FIG. 18. Variation of Settlement Rate for Soli 13 m from Wall
Fig. 17 shows the ground surface settlement profile as a with Time, with Numbers In Parentheses Denoting ~t
function of time for some stages. As shown in this figure, the
ground surface settlement also increased with time while the TABLE 4. Relationship between Ground Surface Settlement
excavation depth remained unchanged. As stated in the pre- and Time
ceding section, both effects of soil creep and excess pore-water Depth Construction day' M ~S/~tb ~S
pressure dissipation may partially contribute to some extent (m) (d) (d) (mm/d) (mm)
because considerable time is normally required for the top- (1 ) (2) (3) (4) (5)
4.9 189-233 45 0.116 5.22
DISTANCE/EXCAVATION DEPTH 8.6 256-318 63 0.10 6.30
0.5 1.0 1.5 2.0 2.5 3.0 11.8 338-363 26 0.238 6.19
.!. 15.2 379-419 22 0.350 7.70
I
f-
Note: Accumulated settlement is 25.41 mm.
a. "The range 189-233 denotes construction day 188 to 232, etc.
w
0 bValues are from Fig. 17.
z
0
n
i=
:ffi
<I:
u II down construction method. Fig. 18 shows the relationship be-
w
x tween settlement rates (as/at) at a distance of 13 m from the
.......
f-
z
-2 wall and excavation depth. As shown in this figure, the settle-
W
w
::::0 ment rate increased with excavation depth. The settlement
...J
rates were in the range of 0.1 mm/d to 0.4 mm/d.
S
V1
(a) Table 4 shows the settlement at a distance of 13 m from the
-3
wall during the periods when the excavation depths remained
DISTANCE/EXCAVATION DEPTH unchanged from the beginning of excavation to the end of
0.0 0.5 1.0 1.5 2.0 2.5 3.0 stage 10. As shown in this table, the accumulated settlement
f-
0.0 (Sd during the periods when the excavation depths remained
w
Z
unchanged was 25.41 mm; the corresponding total settlement
::::0
(S2) was 58 mm (Fig. 15). The ratio of SI to S2 is 44%. This
S
w
-0.5 ratio was fairly large compared with the immediate settlement
V1
induced by excavation.
~ -1.0 Proposed by Clough ond O'Rourke( 1990)
.......
f- EXCAVATION BOTTOM HEAVE
Z
w -STAGE 5
::::0
~ -1.5
........ STAGE
_STAGE
7
9
Very limited field observations regarding heave at the ex-
~
w
- - - STAGE
- STAGE
11 B
13
cavation surface have been reported in the literature. As in-
V1
(b) dicated in Fig. 2, a heave gauge was installed at 20 m from
-2.0 the southern wall at a depth of 21.5 m below the ground sur-
FIG. 16. Observed Settlement Profile versus Computed Set- face (maximum excavation depth = 19.7 m). Fig. 19 shows
tlement Profile Using Empirical Methods: (a) Final Excavation the variation of bottom heave with excavation depth. The ar-
Stage; (b) Key Stages abic number in parentheses in the figure denotes the period of
806/ JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING / SEPTEMBER 1998

J. Geotech. Geoenviron. Eng. 1998.124:798-808.

 10
shallow depth. For the soil at deeper levels, the lateral
earth pressure on the unexcavated side decreased with
E 8 excavation depth to values smaller than the theoretical
~ Rankine active earth pressure. This behavior can perhaps
w
be attributed to the fact that, for Rankine's earth pressure
~:I:
6
theory, no friction between the wall and soil is consid-
:::IE
0 4 ered.
~
0
5. The influence range caused by excavation extended to a
lD considerable distance from the wall. However, the ap-
2
parent influence range, where the building may be af-
fected by the excavation, was equal to about 30.0 m. This
0 value did not vary with excavation stage. The maximum
0 5 10 15 20
EXCAVATION DEPTH (m) ground surface settlement occurred at a distance 0.63 to
FIG. 19. Variation of Excavation Bottom Heave with Depth,
0.78 times the depth where maximum lateral wall de-
with Numbers In Parentheses Denoting At flection occurred. This value also did not vary with ex-
Downloaded from ascelibrary.org by Ryerson University on 06/17/14. Copyright ASCE. For personal use only; all rights reserved.

cavation depth. The normalized settlement profile with

the excavation depth remaining unchanged. As excavation pro- respect to the maximum value was in good agreement
ceeded to the final stage, the magnitude of the heave at ex- with Clough and O'Rourke's studies. The normalized
cavation bottom was equal to 9.7 cm. The ratio of bottom settlement profile with respect to the final excavation
?e~ve to maximum ground surface settlement for each stage
depth was near the boundary of zones I in Peck's dia-
IS In the range of 0.94 to 1.42. The implications of the bottom
gram. The magnitude of the bottom heave was equal to
heave with respect to the conditions of an excavation require 9.7 cm at the final excavation depth. The ratio of bottom
further study. heave to maximum ground surface settlement was in the
range of 0.94 to 1.42
6. Wall deflection and settlement generally increased with
CONCLUSIONS time while the excavation depth remained unchanged.
This paper presents the characteristics of movements of the The magnitudes of the deflection and settlement rates
diaphragm wall and soil caused by the TNEC excavation increased with an increasing excavation depth. In this
project using the top-down construction method. The behavior project, the maximum deflection rate was in the range of
of the diaphragm wall and soil with time was also studied. 0.1 to 0.6 mm/d. The effects of both soil creep and ex-
Records of construction activities as well as observations of cess pore-water pressure dissipation may partially con-
excavation performance were fairly complete. Therefore, these tribute to this behavior. As excavation proceeded to a
observations may not only facilitate a more thorough under- depth of 15.2 m, the percentage of accumulated wall de-
standing of the general excavation behavior, but also provide flection at depths of 10 and 20 m during the periods
a good case history to calibrate and verify numerical tools. In when the excavation depths remained unchanged were
addition, the following observations and conclusions can be 30% and 36%, respectively. The percentage of settlement
made on the basis of the work presented herein: for the soil at a distance of 13 m from the wall was 44%.
These percentages were fairly large compared with the
immediate deformation induced by excavation. There-
1. According to the field observations, the observed strut fore, analysis of excavation in soft clay should consider
loads were close to the values computed using Peck's ap- the creep factors and/or pore-water pressure dissipation.
parent earth pressure diagram for the average undrained 7. The pore-water pressure in the soil outside the excava-
shear strength used. For the struts close to the excavation tion decreased significantly, except for the soil near the
corners, the load in the strut was relatively small. The
wall and above the excavation surface. The magnitude
corner effects apparently had a significant effect on the
of the pore-water pressure decrease near the center of the
strut loads.
site was less than that near the wall. This response is due
2. The maximum horizontal displacement of an inclinom- to the soil near the wall experiencing the release of over-
eter casing 2 m away from the wall was close to that of
burden pressure as well as the lateral compression of the
the wall. Deep inward movements developed for the wall
wall. The pore-water pressure dropped significantly dur-
as well as the soil near the wall. This behavior became
ing the period of excavation. However, pore-water pres-
less pronounced with increasing horizontal distance from
sures gradually increased with time, except for the initial
the wall. The maximum lateral wall displacement oc-
stages. This phenomenon was consistent with the field
curred near the excavation surface. The ratio of maxi-
observations on wall deflection and settlement, in which
mum lateral wall displacement to excavation depth
the movements increased with time.
ranged from 0.51 % to 0.57%, except for the initial ex-
cavation stages.
3. The ratio of the wall bending moment obtained from the ACKNOWLEDGMENTS
rebar strain gauges to the moment obtained from the in- The writers would like to thank the National Science Council at Tai-
clinometer measurements decreased with increasing ex- wan for financial support of this research work under Contract No.
cavation depth. This response may be because some NSC82-04 lO-EO I 1-23.
cracks exist in the concrete wall as a consequence of the
deformation of the concrete wall. Therefore the struc- APPENDIX I. REFERENCES
t~ral design of the diaphragm wall as well as'the predic-
tIOn of the wall deformation should consider this factor. Clough, G. W., and O'Rourke, T. D. (1990). "Construction-induced
4. Field observations indicated that the lateral earth pressure movements of insitu walls." Proc., Design and Peiformance of Earth
at shallow depths (Le., less than about 12.0 m) on the Retaining Structure, ASCE, Reston, Va., 439-470.
DiBiagio, E., and Roti, J. A. (1972). "Earth pressure measurements on a
unexcavated side increased with excavation depth to the braced slurry-trench wall in soft clay." Proc., European Conf. on Soil
at-rest Ko condition because the wall had deep inward Mech. and Found. Engrg., 473-483.
movement, which caused it to push toward the soil at Finno, R. J.• Atrnatzidis, D. K.. and Perkins, S. B. (1989). "Observed

JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING 1 SEPTEMBER 1998/807

J. Geotech. Geoenviron. Eng. 1998.124:798-808.

 performance of a deep excavation in clay." J. Geotech. Engrg., ASCE, Cc = compressive force in concrete;
115(8),1045-1064. C, = compressive force in steel;
Finno, R. J., and Nerby, S. M. (1989). "Saturated clay response during
braced cut construction." J. Geotech. Engrg., ASCE, 115(8), 1065-
Tc = tension force in concrete;
1084. T, = tension force in steel;
Karlsrud, K. (1981). "Performance and design of slurry walls in soft c. = coefficient of consolidation;
clay." Proc., ASCE Spring Convention, ASCE, Reston, Va. E = Young's modulus;
Mana, A. I., and Clough, G. W. (1981). "Prediction of movements for H = depth of excavation;
braced cut in clay." J. Geotech. Engrg. Div. ASCE, 107(8), 759-777. I = moment of inertia;
Mesri, G. (1973). "Coefficient of secondary compression." J. Soil Mech.
and Found. Div., ASCE, 121-137. k= coefficient of permeability;
Moh, Z. C., and Hawang, R. N. (1993). "Earth pressures on walls of a Ko = coefficient of lateral earth pressure at rest;
deep excavation." Proc., 3rd Int. Conf. on Case Histories in Geotech. L = depth of diaphragm wall;
Engrg., 1-5. M = bending moment of wall;
Nicholson, D. P. (1987). "The design and performance of the retaining Nk = empirical cone factor;
wall at Newton Station." Proc., Singapore Mass Rapid Transit Conf.,
147-154. R = reduction factor for moment of inertia;
Ou, C. Y., Hisieh, P. G., and Chiou, D. C. (1993). "Characteristics of U = pore-water pressure;
!J.t = period during which excavation depth remained un-
Downloaded from ascelibrary.org by Ryerson University on 06/17/14. Copyright ASCE. For personal use only; all rights reserved.

ground surface settlement during excavation." Can. Geotech. J., Ot-

tawa, Canada, 30(5), 758-767. changed;
Ou, C. Y., and Liao, J. T. (1995). "Variations of earth and porewater 8. = accumulated deflection during periods when excava-
pressures during deep excavation." J. Chinese Inst. of Civ. and Hydr. tion depths remained unchanged;
Engrg., 7(3), 253-262.
Peck, R. B. (1969). "Deep excavation and tunneling in soft ground." 82 = total deflection;
Proc., 7th Int. Conf. on Soil Mech. and Found. Engrg., 225-290. !J.8/!J.t = deflection rate;
Robertson, P. K., and Campanella, R. G. (1989). "Guidelines for geo- S. = accumulated settlement during periods when exca-
technical design using the cone penetrometer test and CPT with pore vation depths remained unchanged;
pressure measurement." Hogentogler Co., Inc. S2 = total settlement;
Singh, A., and Mitchell, J. K. (1968). "General stress-strain-time function
for soils." J. Soil Mech. and Found. Div., ASCE, 94(1), 21-46. !J.S/!J.t = settlement rate;
4>' = drained friction angle;
APPENDIX II. NOTATION E. = coefficient of secondary consolidation;
(J'~ = effective overburden pressure;
The following symbols are used in this paper:
(J'h = total lateral earth pressure;

Ah m, a = Singh and Mitchell's creep parameters; 8.= ground surface settlement behind wall; and
c = cohesion intercept; =
8.... maximum ground surface settlement.

808/ JOURNAL OF GEOTECHNICAL AND GEOENVIRONMENTAL ENGINEERING / SEPTEMBER 1998

J. Geotech. Geoenviron. Eng. 1998.124:798-808.