Computers and Geotechnics 86 (2017) 141–149

Contents lists available at ScienceDirect

Computers and Geotechnics

journal homepage: www.elsevier.com/locate/compgeo

Research Paper

Assessment of strut forces for braced excavation in clays from numerical

analysis and field measurements
Anthony T.C. Goh c, Fan Zhang d, Wengang Zhang a,b,⇑, Otard Y.S. Chew e
a
Key Laboratory of New Technology for Construction of Cities in Mountain Area, Chongqing University, Ministry of Education, Chongqing 400045, China
b
School of Civil Engineering, Chongqing University, Chongqing 400045, China
c
School of Civil and Environmental Engineering, Nanyang Technological University, 639798, Singapore
d
IIIBIT-Sydney, Federation University, NSW 2000, Australia
e
Geotechnical and Tunnels Division, Land Transport Authority, 219428, Singapore

a r t i c l e i n f o a b s t r a c t

Article history: One important consideration in the design of a braced excavation system is to ensure that the structural
Received 6 November 2016 bracing system is designed both safely and economically. The forces acting on the struts are often deter-
Received in revised form 13 January 2017 mined using empirical methods such as the Apparent Pressure Diagram (APD) method developed by Peck
Accepted 16 January 2017
(1969). Most of these empirical methods that were developed from either numerical analysis or field
Available online 25 January 2017
studies have been for excavations with flexible wall types such as sheetpile walls. There have been only
limited studies on the excavation performance for stiffer wall systems such as diaphragm walls and bored
Keywords:
piles. In this paper, both 2D and 3D finite element analyses were carried out to study the forces acting on
Apparent pressure diagram
Braced excavation
the struts for braced excavations in clays, with focus on the performance for the stiffer wall systems.
Clay Subsequently, based on this numerical study as well as field measurements from a number of reported
Strut forces case histories, empirical charts have been proposed for determining strut loads for excavations in stiff
Finite element analysis wall systems.
Stiff walls Ó 2017 Elsevier Ltd. All rights reserved.

1. Introduction Ou [18] summarized Peck [20]’s work on APD to estimate the

Construction of a basement structure using a braced retaining
wall system will inevitably result in wall deflections and ground
settlement. Excessive ground settlement will frequently cause
damage to adjacent properties in urban areas. The total amount
of ground settlement associated with deep excavations is closely
related to the type of support system, the properties of the
in situ soils, and the excavation procedure. For excavations in clays,
basal heave stability also needs to be considered.
Another important design issue is to ensure the structural
safety of the bracing system. The forces acting on the struts are
often determined using empirical methods such as the Apparent
Pressure Diagram (APD) method. Terzaghi and Peck [22] and Peck
[20] recommended the widely used APD, to estimate the magni-
tude and distribution of prop loads. They proposed different APDs
for braced excavation in sands, stiff fissured clays, and soft to med-
ium clays. This method was developed based on field measured
data for braced excavations with flexible wall systems.
⇑ Corresponding author at: Key Laboratory of New Technology for Construction of
Cities in Mountain Area, Chongqing University, , Chongqing 400045, China.
E-mail address: zhangwg@ntu.edu.sg (W. Zhang).
magnitude and distribution of strut loads in different clays as
shown in Fig. 1, where the Rankine’s coefficient of lateral active
earth pressure Ka is expressed as:
4cu
Ka ¼ 1  m ð1Þ
cHe
where cu is the soil undrained shear strength (in kPa), c is the soil
unit weight (in kN/m3), He is the depth of the excavation (in meter)
and m is an empirical coefficient. Most of the commonly used
empirical methods that were developed from either numerical
results or field studies have been for excavations with flexible wall
systems such as sheetpile walls. To date, there have been limited
studies on the excavation performance for stiffer wall systems such
as diaphragm walls and bored piles.
Chang and Wong [4] proposed a modified APD for diaphragm
walls in deep clay deposits. The research was based on a case study
and a parametric finite element study. Their research showed that
strut loads computed using the Peck’s APD underestimated the
strut loads significantly. By introducing a strut force exceedance
ratio a, as functions of the soil stiffness ratio and undrained shear
strength, a modified APD was proposed. They commented that the
amended APD was derived from the cases with T/B ratio greater

http://dx.doi.org/10.1016/j.compgeo.2017.01.012
0266-352X/Ó 2017 Elsevier Ltd. All rights reserved.
142 A.T.C. Goh et al. / Computers and Geotechnics 86 (2017) 141–149
0 2̚0 4γHe
(a) Soft to medium clays Clay
Fig. 1. APD for design of struts in: (a) soft to medium clays; (b) stiff clays (adapted
from [20].
than 1 (where T is the clay thickness below the final excavation
level, and B is the excavation width, as illustrated in Fig. 2). If T/B
is less than 1, they inferred that there would be strong restraining
effect from the hard stratum reducing the strut force.
Hashash and Whittle [8] compared Peck’s conventional APD
with their FE results which considered undrained strength aniso-
tropy and strength non-homogeneity. Their research indicated that
the conventional APD was smaller than the finite-element results
for diaphragm wall, especially for deep excavations. Also, the wall
stiffness plays an important role in the apparent earth pressures.
As the wall stiffness decreased, the apparent earth pressure
decreased.
Hsiung et al. [10] reported the well-instrumented strut behav-
ior of a 16-m deep excavation with seven level struts restrained
by a diaphragm wall in Taipei. They found that the Peck’s APD
underestimated the measured apparent pressure for this case.
Sze [21] carried out a series of centrifuge tests to investigate the
apparent earth pressure for an undrained excavation. According
to Sze’s test results, Peck’s APD underestimated the measured
apparent pressure by 30 % for the case of excavations supported
Fig. 2. Cross-section and plan view of the model for braced excavation.
by diaphragm walls. Wong et al. [25] observed that for the con-
struction of a major Singapore underground expressway project,
most of the data were within the vertical boundary of the apparent
earth pressure diagram proposed by Terzaghi and Peck [22]. How-
ever, they recommended that the vertical pressure diagram should
extend to the ground surface instead of decreasing to zero to fit all
the measured data.
Twine and Roscoe [23] enhanced Terzaghi and Peck’s work and
introduced the Distributed Prop Load (DPL) method based on 81
case histories and field measurements of prop loads. However, of
the 81 case histories, 28 cases are for flexible walls in soft to med-
ium clays (denoted as class AF) while only 2 cases are for stiff walls
(class AS). In addition, although there are 10 reported cases for stiff
walls in stiff clays (class BS), 5 of them are singly propped while 2
cases have two strut levels and only the remaining 3 cases have
three levels of struts. In view of these limited published data, it
is therefore relevant to reassess the DPL method for the class AS
and class BS excavation types.
All these studies outlined earlier generally indicated that Peck’s
APD under-predicted the apparent earth pressure of the braced
excavations, especially for those involving diaphragm walls and
large excavation depths. As various factors are likely to influence
the APD such as the clay thickness, soil strength and stiffness, wall
stiffness, excavation width, and strut stiffness, this paper explores
the performance of the strutting system, including the apparent
earth pressure, through a series of plane strain and three-
dimensional finite element analysis. Some differences were
observed between the numerical results and Peck’s APD.
As discussed previously, for the case of excavations in stiff wall
systems, the proposed distributed prop loads (DPL) by Twine and
Roscoe [23] were based on only very limited measured data. The
main contribution from this paper is to propose updated APD for
stiff wall systems based on extensive numerical analyses supple-
mented by additional measured data (8 cases in soft clays and 8
cases in stiff clays, with up to five levels of struts).
2. Details of numerical models
For this study, the finite element analyses were carried out
using the geotechnical software PLAXIS 2D (V9.0) and PLAXIS 3D
Foundation [1]. Fig. 2 shows a typical cross-section and plan view
 A.T.C. Goh et al. / Computers and Geotechnics 86 (2017) 141–149 143

for the cases considered. The parameters shown in the figure For the 2D analysis, the retaining wall is simulated using 5-
include: L = excavation length, B = excavation width, D = wall pen- noded elastic plate elements. The elastic behavior is defined by
etration depth, T = clay thickness below the final excavation level the following parameters: EA: normal stiffness, EI: bending stiff-
(FEL), SH = horizontal strut spacing, SV = vertical strut spacing and ness, m: Poisson’s ratio. For the 3D analysis, the wall is simulated
He = depth of final excavation, all in meters. In this study, the geo- using 8-noded quadrilateral plate elements with six degrees of
metrical properties considered include: excavation widths B fixed freedom per node.
at 20 m, excavation lengths L of 44, 68, and 84 m, excavation In this study, three wall types, with five different stiffness val-
depths He of 14 and 16 m, wall penetration depths D of 6 and ues were considered for each soil type, as listed in Table 1. Based
4 m, T = 32 m, horizontal strut spacings SH of 4 and 5 m, and verti- on the approach adopted by Finno et al. [7], the wall thickness of
cal strut spacings SV of 3 m. Vertical retaining walls along the exca- 0.42 m was set to an arbitrary (constant) value so that the moment
vation boundary were installed together with a five-level strut and of inertia I and area A were kept constant, and only the wall elastic
waling system. The thickness of the clay layers were varied for the modulus E was varied. A wall stiffness coefficient a was introduced
different cases studied. to represent walls with different rigidities [2] as shown in Table 1.
For 2D analyses, a half mesh was used due to geometrical sym- The baseline bending stiffness EI for the analysis is
metry. A very fine mesh size was used for 2D analysis to improve 5.04  105 kN m2/m, which refers to a wall of medium stiffness
the accuracy of FE calculations. For 3D analyses, a medium mesh based on the databases of Long [14] and Moormann [15]. There-
size in the horizontal direction and medium coarse mesh size in fore, a = 1.0 for cases with this medium wall stiffness. For flexible
vertical direction were used to reach a balance between processing
time and accuracy. For brevity, only a typical 3D half mesh is Table 2
shown in Fig. 3, comprising of 15,679 nodes and 4980 15-noded Properties of waling system.
wedge elements.
Parameter Struts Walers
For the 2D simulations, fourth order 15-noded triangular ele-
ments, which are considered to be very accurate elements, were Unit weight c (kN/m3) 78.5 78.5
Cross section area A (m2) 0.007367 0.008682
used to model the soil while the interface elements have 5 integra- Young’s Modulus E (kPa) 2.1E8 2.1E8
tion points. In 3D PLAXIS, the interface elements have 9 point Moment of inertia (m4) I3 5.073E5 1.045E4
Gauss integration with three translational degrees of freedom for I2 5.073E5 3.668E4
each node. This is described in greater detail in Van Langen [24]. I23 0 0

Table 3
waler Typical construction sequence for 2D analysis.

Phases Construction details

strut
Phase 1 Install the excavation wall
wall Phase 2 Excavate to 2 m below ground surface
Phase 3 Excavate to 3 m
Phase 4 Install strut system at 2 m below ground surface
Phase 5 Excavate to 5 m below ground surface
Phase 6 Excavate to 6 m below ground surface
Phase 7 Install strut system at 5 m below ground surface
Phase 8 Excavate to 8 m below ground surface
Phase 9 Excavate to 9 m below ground surface
Phase 10 Install strut system at 8 m below ground surface
Phase 11 Excavate to 11 m below ground surface
Phase 12 Excavate to 12 m below ground surface
Phase 13 Install strut system at 11 m below ground surface
Phase 14 Excavate to 14 m below ground surface
Phase 15 Excavate to 15 m below ground surface
Phase 16 Install strut system at 14 m below ground surface
Phase 17 Excavate to 16 m below ground surface
Fig. 3. 3D half mesh of the excavation from PLAXIS 3D foundation.

Table 1
Wall properties for 2D and 3D analyses.

Parameters Wall types

Flexible Medium Stiff
Plane strain (2D) FE parameters
a 0.1 0.2 1.0 2.0 10
System stiffness, S 32 64 320 320 3200
Wall stiffness EI (kN m2/m) 5.04  104 1.008  105 5.04  105 1.008  106 5.04  106
Compressive stiffness EA (kN/m) 3.427  106 6.854  106 3.427  107 6.854  107 3.427  108
Poisson’s ratio, m 0 0 0 0 0
Young’s Modulus (kPa) Three-dimensional (3D) FE parameters
E1 8.16  106 1.632  107 8.16  107 1.632  108 8.16  108
E2 4.08  105 8.16  105 4.08  106 8.16  106 4.08  107
E3 2.00  108 4.00  108 2.00  109 4.00  109 2.00  106
Shear Modulus (kPa) G12 4.08  105 8.16  105 4.08  106 8.16  106 4.08  107
G13 4.00  105 8.00  105 4.00  106 8.00  106 4.00  107
G23 1.333  106 2.666  107 1.333  107 2.666  108 1.33  108
Poisson’s ratio, m 0 0 0 0 0
144 A.T.C. Goh et al. / Computers and Geotechnics 86 (2017) 141–149

walls, the baseline bending stiffness EI of 5.04  105 kN m2/m was is described in greater details in Brinkgreve et al. [1]. For the braced
multiplied by smaller a factors of 0.1 and 0.2, while for stiff walls, excavations in this paper, the struts were placed horizontally at a
the baseline bending stiffness was multiplied by larger a factors of spacing of 4 or 5 m (for different case studies) in two directions
2 and 10. The system stiffness, S [5] in Table 1 is defined as to form a frame net. The walings were used to connect the excava-
tion wall and the struts. The material properties are tabulated in
EI
S¼ ð2Þ Table 2.
cw h4av g The boundary conditions for 2D and 3D cases were: (1) rollers
at side boundaries to allow vertical displacements and (2) pinned
where EI = wall stiffness, cw = unit weight of water, and havg = aver- at the base to restrain any movements. For both 2D and 3D cases,
age vertical strut spacing. the lateral boundaries in the side directions were set at least 100 m
The struts were simulated using node-to-node anchor elements away from the center of the excavation to eliminate the influence
in 2D analysis. For 3D analysis, the struts and walers were modeled of the boundary restraints on the ground movements. This ensures
as beam elements, which have six degree of freedom per node. This that the lateral boundaries are beyond the settlement influence
zone (which is typically 2 times of the excavation depth) induced
Table 4 by the excavation as suggested by Hsieh and Ou [9]. In this study,
Input HS soil parameters of three clays.
the clay thickness below the final excavation level T is assumed as
Parameter Unit A: soft clay B: medium clay C: stiff clay 32 m, which is regarded as fairly large. A typical staged construc-
(Chicago clay) (Taipei silty clay) (Gault clay) tion simulation is shown in Table 3. The original ground water
cunsat kN/m3 18.1 18.1 20 table was assumed to be 2 m below the ground surface in the
csat kN/m3 18.1 18.1 20 retained soil. The water table inside the excavation was progres-
Eref kN/m2 2350 6550 14,847
50 sively lowered with the excavation of the soil during each phase.
Erefoed kN/m2 2350 6550 14,847
The properties of three different types of clays which were con-
Erefur kN/m2 7050 19,650 44,540
C kN/m2 0.05 0.05 0.05 sidered in this parametric study are similar to the properties
u ° 24.1 29 33 assumed by Bryson and Zapata-Medina [2] and are tabulated in
W ° 0 0 0 Table 4. The soils are assumed to follow the Hardening Soil (HS)
mur [–] 0.2 0.2 0.2 model. The three soil types are: soft clay, medium clay and stiff
pref kN/m2 100 100 100
M [–] 1.0 1.0 1.0
clay. The clays are real soils whose properties have been exten-
Knc 0 [–] 0.59 0.55 1.5
Rf [–] 0.7 0.95 0.96
Rinter (interface [–] 1 1 1
Table 6
friction)
Typical maximum compressive strut force for He = 16 m and a = 10.0.

Clay Type Struts Maximum strut force (kN) Max. difference

between 2D
2D 3D
and 3D (%)
L/B = 2.2 L/B = 3.4 L/B = 4.2
Soft clay S1 1034.2 1074.6 1055.2 1046.4 3.9
S2 2342.7 1923.1 1953.6 1934.1 17.9
S3 5019.9 4096.8 4467.6 4529.3 18.4
S4 3931.5 3690.4 3957.1 4046.8 6.1
S5 1028.3 1143.9 1010.0 971.8 11.2
Medium S1 949.5 1023.6 993.4 992.3 7.8
clay S2 1475.3 1292.1 1275.8 1359.1 13.6
S3 3148 2554.7 2452.9 2899.1 22.1
S4 3159.2 3005.6 2671.6 3065.6 15.4
S5 915.7 902.9 767.9 492.3 46.2
Stiff clay S1 1403.1 1430.2 1476.7 1349 3.9
S2 1752.6 1630.9 1717.7 1577.2 10.0
S3 1818.4 1685.9 1762.6 1643.4 9.6
S4 1379.4 1303.5 1430.3 1262.6 8.5
S5 418.7 377.6 357.5 373.0 14.8
Fig. 4. Plan view: horizontal location of struts and walers.

Table 5
Typical maximum compressive strut force for He = 16 m and a = 1.0.

Clay type Struts Maximum strut force (kN) Max. difference between 2D and 3D (%)
2D 3D
L/B = 2.2 L/B = 3.4 L/B = 4.2
Medium clay S1 605.6 645.2 631.1 – 6.4
S2 1689.9 1453.4 1591.1 – 14.0
S3 2315.5 2094.7 2007.1 – 9.6
S4 2886.1 2598.2 2204.5 – 10.0
S5 1130.5 384.4 349.0 – 66.0
Stiff clay S1 944.2 631.1 920.5 888.2 33.2
S2 1670.0 1591.1 1609.5 1597 4.7
S3 1791.4 2007.1 1699.7 1722.6 12.0
S4 1487.1 2204.5 1524.3 1486.4 48.2
S5 492.1 349.0 440.0 482.3 29.1
 A.T.C. Goh et al. / Computers and Geotechnics 86 (2017) 141–149 145

sively reported in the literature. The properties of the soft clay with
average cu = 20 kPa are based on the Upper Blodgett soft clay
reported by Finno et al. [6]. The medium clay with average cu = 45 -
kPa are based on the Taipei silty clay found at the Taipei National
Enterprise Center (TNEC) project [19]. The Gault clay at Cambridge
[16,17] with average cu = 125 kPa was used as the model for the
stiff clay.
A series of 2D and 3D analyses were carried out covering vari-
ous cases of wall stiffness a, soil types, excavation geometries
and different strut levels and locations using the Hardening Soil
(HS) model. For brevity, only the main findings of the numerical
results are presented in the following sections. For all the 2D and
3D cases, the height of wall Hw is fixed at 20 m, so the depth of wall
penetration D decreases as He increases. Apart from slight differ-
ences with regard to the excavation depth He, identical construc-
tion procedures were applied as described in Table 3.

3. Numerical results

In this section, the general trends of the strut forces with differ-
Fig. 6. Cross-section: assumed tributary loading areas of struts for He = 16 m.
ent wall stiffness and for the three clay types are presented. Fig. 4

Fig. 5. Maximum compressive strut forces for various L/B ratios in stiff clay (a = 0.1, 1.0 and 10).
146 A.T.C. Goh et al. / Computers and Geotechnics 86 (2017) 141–149

Table 7
Summary of case histories.

Case histories Soil type Unit weight He (m) Wall type No. of Strut Horizontal strut Vertical strut References
(kN/m3) levels spacing (m) spacing (m)
BTG residual soils Soft clay 18.5–19.2 19.5– D-wall/SBP wall (stiff 4–6 4.2–8.5 4.0–6.0 This study
24.0 wall)
Taipei (CIRIA report) AS1 Soft clay 17.5 14.4 D-wall (stiff wall) 4 5.5 4.0 Twine and
5.5 3.8 Roscoe [23]
5.5 2.9
5.5 2.3
Taipei (CIRIA report) AS2 Soft clay 19.2 14.1 D-wall (stiff wall) 4 6.0 2.7 Twine and
6.0 2.9 Roscoe [23]
6.0 3.4
6.0 3.4
(CIRIA report) BS4 Stiff clay 19.6 9.5 D-wall (stiff wall) 2 4.5 6.0 Twine and
4.5 2.9 Roscoe [23]
(CIRIA report) BS6 Stiff clay 20.0 10.0 D-wall (stiff wall) 3 1.7 2.6 Twine and
1.7 3.0 Roscoe [23]
1.7 3.0
(CIRIA report)BS7 Stiff clay 19.5 12.4 D-wall (stiff wall) 3 1.7 4.3 Twine and
1.7 3.0 Roscoe [23]
1.7 3.3
(CIRIA report) BS9 Stiff clay 18.9 15.3 D-wall (stiff wall) 3 3.6 6.3 Twine and
3.6 4.1 Roscoe [23]
3.6 3.1
(CIRIA report) BS10 Stiff clay 18.8 16.6 D-wall (stiff wall) 2 4.2 4.3 Twine and
4.2 5.3 Roscoe [23]
Singapore Circle Line Soft clay 16.0 18.0 D-wall (stiff wall) 4 – – Cham and Goh
projects [3]
Singapore Circle Line Stiff clay 20.0 22.0 D-wall (stiff) 4–5 – – Cham and Goh
projects 22.0 20.0 [3]
22.0 21.0
22.0 18.0
Singapore Old Alluvium Soft clay – – D-wall (stiff) – – – Li [13]
soil sites
Singapore Old Alluvium Stiff clay – – D-wall (stiff) – – – Li [13]
soil sites
Shanghai CBD deep Soft clay 17.2 17.5 D-wall (stiff) 4 – 4.4 Lau et al. [12]
excavation 4.6
4.0
4.6
Singapore C907 site, CST Soft clay 18.0 16.0 1.1 m diameter CBP 5 5 1.0 Jadhav [11]
3A wall (stiff) 3.0
3.0
3.0
3.0
Singapore C907 site Soft clay 18.0 15.0 1 m thick D-wall 4 6 2.5 Jadhav [11]
(stiff) 3.0
3.5
4.0
Singapore Central Soft soil overlying 18.0 Up to Arbed wall, CBP wall 3–5 5.0–6.0 3.0–5.0 Wong et al.
Expressway Phase II stiff soils 21 and D-wall [25]

Notations: D-wall: diaphragm wall; SBP wall: Secant bored pile wall; CBP wall: contiguous bored pile wall; BTG: Bukit Timah granite.
A: soft and medium clays; B: stiff and very stiff clays; F: flexible wall; S: stiff wall; AF: flexible wall in soft and medium clays; AS: stiff wall in soft and medium clays; BS: stiff
wall in stiff clays, the above are used for classifications of the ground and wall types.

depicts the plan view arrangement of the struts and walers
(depicted schematically in red1) for the 3D analyses. The symbol x
denotes the horizontal distance from the center of the excavation.
There is no strut located at the middle section of the excavation,
so the simulations of both the struts and the walers of 3D excava-
tions are symmetrical. For each level of struts, the struts at the
x = 2 m and x = 2 m section are identical because of symmetry.
These struts which are the closest to the middle section are sub-
jected to the largest force among all the struts at the same level.
Hence only the strut forces at location x = 2 m section are presented
and discussed. These values will be used in the subsequent section to
assess the APD.
1
For interpretation of color in Figs. 4, 7, and 8, the reader is referred to the web
version of this article.
For brevity, only some of the results from the 2D and 3D anal-
yses are presented. Tables 5 and 6 show some typical results of
the strut forces from the 2D and 3D analyses. S1 denotes the top
strut and S5 is the lowest strut level. Generally, the struts S2, S3
and S4 are subjected to larger axial forces. The general trend was
for the forces in struts S2 to S5 to decrease with increasing soil
strength. It can be observed from Tables 5 and 6 that the differ-
ences of maximum compressive strut forces depend on the clay
types, L/B ratios, and the strut levels. Fig. 5 shows the maximum
axial force of different strut levels for various L/B ratios in stiff clay
(a = 0.1, 1.0 and 10). Generally, the strut forces in struts S2 to S5
increased with increasing wall system stiffness which is similar
to the trend observed by Hashash and Whittle [8]. It is obvious
from this plot that the differences of maximum compressive strut
forces for various L/B ratios (and 2D vs 3D) in stiff clay are less
significant.
 A.T.C. Goh et al. / Computers and Geotechnics 86 (2017) 141–149
4. Apparent earth pressure envelope for stiff walls
For design of braced deep excavation in soils, apparent earth
pressure diagrams are usually used to calculate the minimum hor-
izontal strut loads for each supporting level during vertical excava-
tion. Herein the results from the 2D and 3D maximum strut forces
for He = 14 m and He = 16 m were used to derive the apparent earth
pressures for excavations in stiff walls. The tributary area load dis-
tribution procedure proposed by Peck [20] was used. The assumed
loading area for each strut is plotted in Fig. 6 taking as an example
the case with He = 16 m, where hsi represents the loading area
depth for the corresponding strut.
For the force on the strut P, since the horizontal spacing of the
struts SH is 4 m, the apparent earth pressure rh can be obtained:
P P
rh ¼ ¼ ð3Þ
SH hsi 4hsi
The normalized apparent earth pressure (rh/cHe) and the nor-
malized depth (z/He) are then plotted. Only excavations in stiff
walls for different clays are considered as explained in the Intro-
duction because of the limited studies that have been carried out
so far and also because these walls are becoming more common
for very deep excavations in congested urban environments. For
comparison, Peck’s APD, the CIRIA DPL recommendations [23],
together with various more recent case histories from the litera-
ture as listed in Table 7, are also plotted alongside the FE results.
For the soft clay, the Ka value is calculated using Eq. (1).Following
the approach from CIRIA for excavations with flexible wall sys-
tems, this paper provides an update for excavations with stiff
walls. The proposed normalized apparent earth pressures for exca-
vations with stiff walls in soft clays and stiff clays are shown in
Figs. 7 and 8, respectively. Excavations in medium clays have been
excluded due to limited field measurements from the literature.
For comparison, Peck’s APD for soft clays (cu = 20 kPa) assuming
Fig. 7. Normalized apparent earth pressure versus normalized depth for stiff walls in soft clays.
147
m = 0.4 is also shown in Fig. 7. For stiff clays (Fig. 8), two dash lines
are used to define the upper and lower bound of Peck’s APD. Also
shown in Fig. 8, represented by the blue dash line is the character-
istic distributed prop load diagram of DPL=cHe ¼ 0:5 reported by
CIRIA [23] for stiff walls in stiff to very stiff clays. Fig. 8 also
includes the proposal by Yau and Sum [26] who numerically anal-
ysed the strut loads of typical excavation and lateral support work
for two case histories in Hong Kong and proposed the maximum
prop load line of 0.65cHe.
Fig. 7 shows that for excavations in soft clays with stiff walls,
the FE results (represented by open symbols) indicate the apparent
earth pressure is generally less than 0.3cHe for z/He 6 0.25 while
for z/He P 0.25, the apparent earth pressure is generally less than
0.9cHe. Based on the measured strut forces from the case histories
(denoted as solid symbols), for z/He 6 0.25, the apparent earth
pressure is generally less than 0.4cHe, while for z/He P 0.25, the
apparent earth pressure is generally less than 1.0cHe. Numerical
and field measurements indicate that Peck’s APD (which was
developed solely based on flexible walls) can sometimes underes-
timate the apparent earth pressure for soft clays with stiff walls.
From Fig. 7, it can also be observed that the line proposed by Yau
and Sum [26] overestimates the apparent earth pressure
for z/He 6 0.25 and underestimates the apparent earth pressure
for z/He > 0.25. Thus, for excavations in soft clays with stiff walls,
following the similar approach adopted by CIRIA for excavations
with flexible wall systems, a threshold apparent earth pressure of
0.45cHe for z/He 6 0.2 and threshold apparent earth pressure of
0.9cHe for 0.2 < z/He 6 1.0 is proposed, as indicated in Fig. 7, by
the red solid lines.
For excavations in stiff clays with stiff wall systems, for the
numerical results the apparent earth pressure is generally less than
0.4cHe for z/He 6 0.25, while for z/He P 0.25, the apparent earth
pressure is generally less than 0.5cHe, which fits the CIRIA charac-
teristic DPL of 0.5cHe well. Based on the measured strut forces
148 A.T.C. Goh et al. / Computers and Geotechnics 86 (2017) 141–149
Fig. 8. Normalized apparent earth pressure versus normalized depth for stiff walls in stiff clays.
Table 8
Recommended Ka values for stiff walls in clays.
Clay type Ka
Soft clay 0.45 (z/He 6 0.20)
0.9 (0.20z/He 6 1.0)
Stiff clay 0.4 (z/He 6 0.20)
0.6 (0.20z/He 6 1.0)
from the case histories, for z/He 6 0.25, most of the data are less
than 0.5cHe, while for z/He P 0.25, the apparent earth pressure is
also generally less than 0.6cHe, except for two points (with
0.63cHe and 0.61cHe, respectively), which are slightly larger than
0.6cHe but within the limit of 0.65cHe as proposed by Yau and
Sum [26]. Thus, for excavations in stiff clays with stiff walls, a
threshold apparent earth pressure of 0.4cHe for z/He 6 0.2 and
threshold apparent earth pressure of 0.6cHe for 0.2 < z/He 6 1.0 is
proposed, as indicated in Fig. 8, by the red solid lines.
Table 8 summarizes the recommended APD for excavations
using stiff wall systems. It is expressed in the form of KacHe.
5. Summary and conclusions
A detailed study of strut forces were carried out through a series
of 2D and 3D finite element analyses for excavations in clays. The
general trend was for the forces in struts to decrease with increas-
ing soil strength and to increase with increasing wall system stiff-
ness. The results indicated that the differences in the strut forces
between different L/B (and 2D vs 3D) depend on the clay types,
L/B ratios, and the strut levels. For various L/B ratios in stiff clay,
the differences are less significant. Based on the numerical study
as well as field measurements from a number of reported case his-
tories, empirical charts have been proposed for excavations in clay
soils with stiff wall systems as shown in Figs. 7, 8 and listed in
Table 8.
Acknowledgements
Part of this research is supported by the LTIF project titled
‘‘Braced Excavation-induced Ground Movements”, funded by the
Land Transport Authority (LTA), Singapore. The authors would like
to acknowledge the financial support from LTA. The corresponding
author is also grateful to the support by the National Natural
Science Foundation of China (Nos. 51608071 and 51420105013).
References
[1] Brinkgreve LBJ, Swolfs WM, Engin E. Plaxis manual. Netherlands: PLAXIS BV;
2011.
[2] Bryson L, Zapata-Medina D. Method for estimating system stiffness for
excavation support walls. J Geotech Geoenviron Eng 2012;138:1104–15.
http://dx.doi.org/10.1061/(ASCE)GT.1943-5606.0000683.
[3] Cham WM, Goh KH. Observed apparent pressure diagrams from actual strut
monitoring of excavations in Circle Line project. In: Proceedings of
underground singapore, Singapore, 8–9 December 2011. p. 289–97.
[4] Chang JD, Wong KS. Apparent pressure diagram for braced excavations in soft
clay with diaphragm wall. In: Proceedings of international symposium on
geotechnical aspects of underground construction in soft ground. A.A.
Balkema; 1996. p. 87–92.
[5] Clough GW, Smith EM, Sweeney BP. Movement control of excavation support
systems by iterative design. Proceedings of foundation engineering congress
on current principles and practices, New York, vol. 2. p. 869–84.
[6] Finno RJ, Bryson S, Calvello M. Performance of a stiff support system in soft
clay. J Geotech Geoenviron Eng 2002;128:660–71. http://dx.doi.org/10.1061/
(ASCE)1090-0241(2002)128:8(660).
[7] Finno RJ, Blackburn JT, Roboski JF. Three-dimensional effects for supported
excavations in clay. J Geotech Geoenviron Eng 2007;133:30–6. http://dx.doi.
org/10.1061/(ASCE)1090-0241(2007)133:1(30).
[8] Hashash YMA, Whittle AJ. Mechanisms of load transfer and arching for braced
excavations in clay. J Geotech Geoenviron Eng 2002;128:187–97. http://dx.doi.
org/10.1061/(ASCE)1090-0241(2002)128:3(187).
[9] Hsieh PG, Ou CY. Shape of ground surface settlement profiles caused by
excavation. Can Geotech J 1998;35:1004–17. http://dx.doi.org/10.1139/t98-
056.
[10] Hsiung BCB, Nash DFT, Tsai WS, Hwang RN. Observed behaviour of a deep
excavation in Taipei. Proceedings of civil and environmental engineers
conference-new frontiers and challenges, vol. II. p. 55–66.
[11] Jadhav AS. Field measurements of strut loads in LTA contract C907. In:
Proceedings of underground singapore, Singapore, 8–9 December 2011. p.
267–76.
 A.T.C. Goh et al. / Computers and Geotechnics 86 (2017) 141–149
[12] Lau CS, Chiu SL, Lo KL, Chu KKN. Ground response in deep excavation in soft
soil in Shanghai. In: Proceedings of the 30th annual seminar geotechnical
division, Hong Kong, 6 May 2010. The Hong Kong Institution of Engineers;
2010. p. 149–61.
[13] Li W. Braced excavation in old alluvium in Singapore [Ph.D.
Thesis]. Singapore: Nanyang Technological University; 2001.
[14] Long M. Database for retaining wall and ground movements due to deep
excavations. J Geotech Geoenviron Eng 2001;127:203–24. http://dx.doi.org/
10.1061/(ASCE)1090-0241(2001)127:3(203).
[15] Moormann C. Analysis of wall and ground movements due to deep excavations
in soft soil based on a new worldwide database. Soils Found 2004;44:87–98.
http://dx.doi.org/10.3208/sandf.44.87.
[16] Ng CWW. An evaluation of soil-structure interaction associated with a multi-
propped excavation [Ph.D thesis]. UK: University of Bristol; 1992.
[17] Ng CWW, Yan RWM. Stress transfer and deformation mechanisms around a
diaphragm wall panel. J Geotech Geoenviron Eng, ASCE 1998;124:638–48.
http://dx.doi.org/10.1061/(ASCE)1090-0241(1998)124:7(638).
[18] Ou CY. Deep excavation: theory and practice. London: Taylor & Francis Group;
2006.
[19] Ou CY, Liao J, Lin H. Performance of diaphragm wall constructed using top-
down method. J Geotech Geoenviron Eng, ASCE 1998;124:798–808. http://dx.
doi.org/10.1061/(ASCE)1090-0241(1998)124:9(798).
149
[20] Peck RB. Deep excavations and tunneling in soft ground. In: Proceedings of the
7th international conference on soil mechanics and foundation engineering,
Mexico City. p. 225–90.
[21] Sze S. Ground movements due to excavation in clay: physical and analytical
models. Doctor of Philosophy: University of Cambridge; 2010.
[22] Terzaghi K, Peck RB. Soil mechanics in engineering practice. 2nd ed. New
York: John Wiley & Sons; 1967. p. 729.
[23] Twine D, Roscoe H. Temporary propping of deep excavations-guidance on
design. London: CIRIA C517, CIRIA; 1999.
[24] Van Langen H. Numerical analysis of soil-structure interaction [PhD
thesis]. The Netherland: Faculty of Civil Engineering, Delft University of
Technology; 1991.
[25] Wong IH, Poh TY, Chuah HL. Performance of excavations for depressed
expressway in Singapore. J Geotech Geoenviron Eng 1997;123:617–25. http://
dx.doi.org/10.1061/(ASCE)1090-0241(1997)123:7(617).
[26] Yau PKF, Sum AHL. An analytical review of excavation and lateral support, case
history in Hong Kong. In: Proceedings of the 30th annual seminar geotechnical
division, Hong Kong, 6 May 2010. The Hong Kong Institution of Engineers;
2010. p. 47–53.