Computers and Geotechnics 69 (2015) 452461

Contents lists available at ScienceDirect

Computers and Geotechnics

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

Research Paper

Numerical investigation of spatial aspects of soil structure interaction

for secant pile wall circular shafts
Ali Chehadeh, Alper Turan, Farid Abed
American University of Sharjah, PO Box 26666, Sharjah, United Arab Emirates

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

Article history: Implementation of fast and cost-effective shoring systems has become very necessary to overcome the
Received 30 October 2014 technical challenges such as variable soil and rock proles, high groundwater tables and limitations
Received in revised form 10 May 2015 imposed by the built environment. Secant pile wall shoring systems allow the construction of overlapped
Accepted 14 June 2015
piles in almost all subsurface conditions. They are constructed in a circular plan layout to form a vertical
Available online 25 June 2015
shaft which provides unique advantages such as compression ring behavior. This paper presents a numer-
ical study to investigate various aspects of the behavior of circular shafts constructed using secant pile
Keywords:
walls. The studied aspects include the identication of earth pressure distributions exerted on circular
Diaphragm wall
Soil-structure interaction
shafts, the impact of excavation of single and multiple holes on the shaft stresses, and the stresses in
Circular shaft the shaft in the case of sloping bedrock. A three-dimensional nite element model is developed to con-
Hole excavation duct the present analyses taking into consideration the actual behavior of soils surrounding the walls. The
Sloping bedrock stress concentrations calculated for circular shafts were seen to vary from the results of the innite plate
with hole solution. The sloping bedrock was also seen to result in signicant deviations from the com-
pression ring behavior. A large increase in the maximum compressive stresses and emergence of some
signicant tensile stress zones were observed for bedrock inclinations larger than 20. The results pre-
sented in this study address some practical design concerns and were considered to be of interest to those
involved in design and construction of vertical shafts.
2015 Elsevier Ltd. All rights reserved.

1. Introduction procedure is repeated until the entire wall is completed. Piles

Diaphragm wall shoring systems are used to facilitate deep
excavations and can be formed with various layout options
depending on the shape of the required excavation. These systems
can be supported by various means as the excavation advances.
The common construction approaches utilized for these structures
include slurry walls and secant pile walls. The use of diaphragm
walls in a circular layout forms vertical shafts which perform as
a compression ring under lateral earth and hydrostatic pressures
acting on them. Such vertical shafts are formed by installing a ser-
ies of overlapping concrete piles to form a continuous, water tight
wall. The secant pile walls are constructed by installation of an
array of primary piles. The installation of primary piles is followed
by installation of secondary piles, which are cut into the primary
piles forming a continuous wall. Concrete with a slow rate of set-
ting is typically used. Before the concrete sets, the sharp edged sec-
ondary pile cases are driven and cut into the primary ones. This
Corresponding author.
E-mail address: fabed@aus.edu (F. Abed).
can be constructed of either structural or lean concrete.
This paper focuses on evaluation of various aspects such as the
realistic distribution of earth pressures and the external loads, the
stress concentrations caused by excavation of holes on the walls of
the circular shaft, the impact of socketing of the secant walls into
the bedrock and inclination of the bedrock surface are of signicant
importance for a successful design and implementation of a secant
pile wall circular shaft. The treatment of the earth pressures of dia-
phragm walls differ from that of other wall types because of the
unique installation approach used in their construction. Since the
diaphragm walls installed to form a circular shaft are allowed to
harden before the excavation proceeds, the in-situ stresses in the
ground are assumed practically to remain unchanged after the
excavation (e.g. [21]). Several recent studies suggested that the
earth pressures behind diaphragm walls are affected by the instal-
lation (e.g., [9,20,8,24,1,2]).
The interaction between the circular shaft and surrounding soil
was also studied by many authors using reduced scale model tests.
Muller-Kirchenbauer et al. [22] studied the earth pressures exerted
on a circular shaft by dry sand. The results of this study indicated

http://dx.doi.org/10.1016/j.compgeo.2015.06.014
0266-352X/ 2015 Elsevier Ltd. All rights reserved.
 A. Chehadeh et al. / Computers and Geotechnics 69 (2015) 452461 453

that rigid models do not allow recess and provide the ultimate the wall friction angle and other parameters (see [14] for complete
pressures whereas; exible models allowed wall deformations mathematical derivation).
and exhibited larger reduction in pressures. Lade et al. [15] used Stresses in a circular diaphragm walls can be estimated by
centrifuge testing to study the earth pressures exerted on a vertical determining the distribution of stresses in a cylinder subjected to
circular shaft in dry sand. The earth pressures were slightly uniform radial pressures, which change with depth, acting on the
exceeded at-rest values prior to the excavation of the soil inside outer surface of the shaft. The theory of thick cylinders is consid-
the shaft. However, earth pressures reduced substantially with ered if the wall radius to thickness ratio is more than 10. In this
the excavation of the soil and resultant inward movement of the case, the meridional and hoop stresses cannot be considered uni-
shaft walls. Shin and Sagong [27] performed model tests to deter- form throughout the thickness of the wall and the radial stress can-
mine the earth pressures on a circular shaft. It was shown that the not be considered negligible. The parametric analyses and the case
earth pressures on the shaft reduce substantially if wall deforma- studies presented in this paper comprise walls that fall under thick
tions larger that 1.5% of the shaft radius takes place. At small defor- walled cylinder category, thus the relevant wall stresses must be
mations, the earth pressures remained between at-rest and active determined (see Fig. 2). For a thick-walled cylinder, which is sub-
conditions. The analytical and numerical studies have shown that jected to a uniform external radial pressure, q, and longitudinal
the pressures acting on the exterior surfaces of a circular shaft pressure of zero (or externally balanced), the stresses and changes
are less than at-rest pressures assumed in plane strain conditions in cylinder dimensions can be expressed as:
due to the stresses relief during the excavation if sufcient defor-
r1 0 2
mations occur in the wall [27]. Horizontal arching effects also
inuence the distribution of earth pressures on the circular exte- 2
qa2 b r 2 q2a2
rior surface of the walls. However, due to the construction tech- r2 ; r2 max at r b 3
2 2
niques employed in the construction of secant pile walls, only r 2 a2 b a2  b
small deformations take place and the deviation of earth pressures
2
from at rest conditions toward active conditions due to the wall qa2 r 2  b
deformations remain insignicant even at large shaft diameters.
r3 2
; r3 max q at r a 4
r 2 a2  b
The stresses generated in the shaft walls are calculated with the
assumption of compression ring behavior, which results in purely r2 max qa2
compressive stresses. However, there are practical situations that smax 2
; at r b 5
2 a2  b
will result in deviations from these stress conditions. For example,
the excavation of a hole(s) on the circular shaft and existence of !  
2
sloping bedrock conditions cause the stress distributions in the qa a2 b q 2a2
Da 2
l ; Db 2
;
shaft that will result in tensile stresses that are not observed in a E a2  b E a2  b
pure compression ring behavior.  
qll 2a2
This paper presents the results of parametric numerical analy- DL 2
6
E a2  b
ses that investigate interaction between vertical circular shafts
and surrounding soil with the consideration of holes on the shaft where r1, r2 and r3 are stresses in the longitudinal, circumferential
walls. The impact of the sloping bedrock conditions are also con- and radial directions, respectively (see Fig. 2). Da, Db and DL are the
sidered parametrically in order to identify the deviations from changes in the dimensions of a, b and L, respectively. The major
compression ring behavior. Three-dimensional nonlinear nite ele- parameters to be considered for an analytical design for diaphragm
ments analyses were performed using the commercial nite ele- wall circular shafts are the values of maximum external radial pres-
ment package, ABAQUS. Elasto-plastic soil behavior of retained sure q, the stresses in the wall section and the change in the radius.
soil and a non-linear soil-shaft interface was assumed in the
analyses. 2.2. Thick walled cylinders with holes

2. Background Some literature exists on the behavior of thick walled cylinders

2.1. Analytical evaluation of earth pressures and wall stresses
The vertical shaft behavior is signicantly inuenced by the
gravitational forces and resultant stress state in the surrounding
soil [29]. Vertical, radial and tangential stress components are usu-
ally developed around and along the shaft due to the 3D nature of
the stress concentrations near a vertical circular shaft (Fig. 1a). A
common shape assumption for a failure wedge is depicted in
Fig. 1b, where b = 45 + h/2 and h is dened in terms of the wall
friction angle.
Design of circular diaphragm walls is typically performed using
closed form analytical solutions assuming a uniform external pres-
sure distribution to represent the earth pressures, hydrostatic pres-
sures and surcharge. Kim et al. [14] provided a comprehensive
summary of the earth pressure assumptions typically used in prac-
tice. In this regard, the total soil pressure (Pi) was calculated as
follows:
Pi K wa rv 1
where rv is the effective vertical overburden pressure and Kwa is the
coefcient of radial earth pressure which was dened in terms of
with holes. For example, Makulsawatudum et al. [18] presented
elastic stress concentration factors for internally pressurized thick
walled cylindrical vessels with radial, offset circular and elliptical
cross holes. Laczek et al. [16] performed elasto-plastic analysis of
stressstrain state in the vicinity of a hole in a thick walled cylin-
drical pressure vessel using the nite element method. Nihous
et al. [25] studied elastic stress concentration factors for internally
pressurized thick walled cylinders with oblique circular cross holes
using the nite element method considering various wall
cross-hole ratios. Li et al. [17] employed inelastic FE analyses to
study the effect of autofrettage on the stress levels in thick walled
cylinders with a radial cross-hole. Duncan et al. [5] determined the
effect of a cross hole on the inelastic response by considering the
shakedown and ratcheting behavior of plain thin and thick walled
cylinders with a radial cross-hole, subjected to constant internal
pressure and cyclic thermal loading. No literature was encountered
on the stress regimes around a hole in a large diameter thick
walled cylinder such as the secant pile wall circular shaft. The fac-
tors such as the external pressure distributions around a circular
shaft and the location of the hole along the height of the shaft
are the factors complicating the stress regime around the hole
making an independent evaluation necessary.
454 A. Chehadeh et al. / Computers and Geotechnics 69 (2015) 452461
Fig. 1. Behavior of the soil around vertical circular shaft: (a) mode of yielding and (b) slip surface.
Fig. 2. Stress components in thick cylinders (a: outer radius, b: inner radius).
The stress concentration factors determined using the nite ele-
ment analyses were compared with those obtained using the circu-
lar hole in an innite plate solution [26]. Fig. 3 depicts the stresses
on the plate. The maximum stresses under uniaxial conditions,
where r2 = 0, were 3r1 and 1r1 at the maximum compression
and tension points, A and B respectively.
3. Methodology
The effect of different shaft parameters on the mechanics and
overall response of the circular shaft system were studied
Fig. 3. Elastic stress, in-plane normal stress.
numerically using the Finite Element (FE) Method with the consid-
eration of SoilStructureInteraction (SSI). A series of
three-dimensional (3D) non-linear FE models were developed
using the commercial nite element software ABAQUS to simulate
the structural response of a circular shaft inserted in soil. The FE
model was rst veried by comparing displacements, vertical pres-
sure distribution, and circumferential stresses with existing ana-
lytical solutions and eld data. The veried FE model was then
utilized to conduct a parametric study, investigating various design
aspects for vertical shafts.
The rst step in simulating the realistic SSI behavior was con-
ducted by considering an initial geostatic step to dene the gravity
loading and also to establish an initial stress regime in the ground.
Once the geostatic equilibrium was established within the ground,
the soil excavation within the shaft was conducted in multiple
stages after the complete construction of the circular shaft.
A benchmarking step was then conducted by studying the effect
of different shaft parameters such as the depth, diameter and
thickness on the mechanics and overall response of an intact circu-
lar shaft system (i.e., without holes). Next, the stress concentra-
tions resulting from the excavation of a hole on the circular shaft
was investigated. Various shaft dimensions, hole radii and loca-
tions were studied. The interaction between the multiple holes
was also studied parametrically. The nal stage of the present
study included the investigation of the effect of inclined bedrock
conditions on the stresses that occurred in the shafts that were
socketed into the bedrock. The deviation from the compression
ring behavior as a result of sloping rock conditions was also evalu-
ated. More details about the FE analysis steps including the model
geometry, calibration, meshing and element types, and various
soils properties are explained next.
3.1. Geometric description
The basic geometry of soil was modeled as a block that contains
a circular hole, which extends according to the shafts basic geom-
etry for a specic height and thickness (see Fig. 4). The dimensions
of the soil block depend on the radius and depth of the concrete
shaft, and extend far enough away to eliminate any boundary
effect. For the present FE study, a range of shaft radii between
5 m and 20 m with a constant wall thickness along the depth
was simulated. The corresponding dimensions of the soil surface
ranged from 30 m  30 m to 200 m  200 m, while the depths
 A. Chehadeh et al. / Computers and Geotechnics 69 (2015) 452461
Fig. 4. Problem geometries for (a) soil block, and (b) circular concrete shaft.
extend enough below the shaft. The shaft thickness in all analyses
was considered to be 0.84 m.
The compression ring behavior of a secant pile shaft (without
hole) in cohesionless soil was rst investigated for comparisons
and normalization purposes. Several combinations of shaft radii
and depths were considered as listed in Table 1.
Fig. 5 illustrates the soil geometry that was modeled to investi-
gate the existence of inclined rock layers interfering with the con-
crete shaft. A range of inclination angles (0, 10, 20, and 30) for
the rock layer was investigated while the socketed shaft distance in
the rock was limited to 2.5 m.
The geometries of the concrete shafts containing one and two
holes are presented in Fig. 6. Circular-hole openings of diameters
ranging between 1.0 m and 8.0 m, and for different shaft radii,
5 m, 10 m, 15 m, and 20 m were studied and modeled in the pre-
sent FE analysis. For the case of a two-hole shaft, two hole diame-
ters of 1 m and 2 m, and two shaft diameters of 5 m and 10 m were
considered. The effect of the distance between the adjacent holes
was also investigated, and a range of 2r14r (where r is the hole
radius) was covered.
3.2. Material properties
The soil stratigraphy comprised a single uniform layer of cohe-
sionless soil with isotropic properties. An elasto-plastic model with
MohrCoulomb failure criterion was utilized to represent the
behavior of the soil while a linear elastic behavior was assumed
for the concrete material that was used to model the secant pile
wall and also for the rock. Three common cohesionless soil proper-
ties were considered. Table 2 summarizes the material properties
of the soils, rock and concrete used in the present FE simulations.
The ground water table was assumed to be deep and its inuence
on the interaction between the secant pile wall and the surround-
ing soil was neglected for the sake of simplicity. A constant prole
of soil elastic modulus was assumed and dilation mechanism was
Table 1
Geometric parameters investigated for circular shafts without holes.
Shaft radius R (m) Shaft depth H (m)
3 3, 6, 9, 12, 15, 18, 21, 24
5 5, 10, 15, 20, 25, 30, 35, 40
10 10, 20, 30, 40
455
neglected in the analyses as their inuence on the mechanical
behavior of such low deformation systems is very minor. Fig. 7
depicts the normalized earth pressure distribution on a 40 m deep
shaft using two different sets of data. Soil properties of soil 2 in
Table 2 are adopted in the investigation. First set of data assumes a
constant value of soil modulus (E = 30 MPa) along the shaft while
the other one assumes a varying soil modulus along shaft depth,
where the average modulus was equal to the value of constant
modulus (30 MPa).
The elasto-plastic behavior of the cohesionless soil considered
in this study was modeled using the MohrCoulomb failure crite-
rion, which illustrates the failure conditions of an isotropic mate-
rial [12]. The experiments show that the criterion is appropriate
when principal stresses are all compressive while there shall be
some modications for the tensile principal stresses [23].
Coulomb [4] assumed a linear failure relationship between shear
stress (s) and compressive stress (r), while Mohr [19] assumed
that the failure is only related to minimum principle stress (r1)
and maximum principle stress (r3). The MohrCoulomb failure cri-
terion is already implemented in the ABAQUS formulation. The
plastic behavior can be modeled using either the MohrCoulomb
surface or the Rankine surface for the tension cutoff (for more
details see ABAQUS Documentation [6]).
3.3. Boundary conditions and contacts
The soil boundaries in each model were assumed to be at least
ve times shaft depth (H) from the edge of the circular shaft to
ensure no effects on the stresses in and around the shaft. A soil
depth of at least one shaft depth (H) was maintained below the
bottom of the shaft in all models. Displacement boundary condi-
tions were xed at the sides in both horizontal directions and unre-
strained in the vertical direction, while they were xed in both
horizontal and vertical directions at the model base. Normal and
tangential surface-to-surface contacts with an allowed nite slid-
ing were dened between the concrete shaft and the surrounding
soil. The tangential contact was dened using a stiffness (penalty)
method that permits some relative motion of the surfaces, with a
coefcient of friction set to 0.3. The normal contact, on the other
hand, was dened using the classical Lagrange multiplier method
of constraint enforcement Hard, with an option to allow separa-
tion after contact. The proper selection of contact elements helps in
simulating the real behavior of the SoilStructureInteraction (SSI)
and also in accurate calculation of stress distribution along the
shaft.
3.4. Element types and mesh sensitivity analysis
A free mesh algorithm was adopted for the soil and concrete
shaft as shown in Fig. 8. The concrete shaft was dened as a con-
ventional shell model and was meshed using 3D linear triangular
shell elements. The soil block, on the other hand, was modeled as
a solid part using 3D tetrahedral linear elements as shown in
Fig. 8. A mesh sensitivity analysis was also conducted to select
an optimum mesh with a proper number of elements to be used
for the present parametric study. For that reason, ve mesh cong-
urations with approximate element sizes of 1.5, 1.0, 0.5, 0.3, and
0.2 were considered as shown in Fig. 9(a). The contour lines of
radial displacements along the shaft surface obviously indicate
that Meshes 4 and 5 provide similar numerical results compared
to those of coarse mesh, so no more mesh renement is required
(Fig. 9b). Thus, Mesh 4 was selected for the present parametric
study as it provides acceptable accuracy and minimum computa-
tional time.
456 A. Chehadeh et al. / Computers and Geotechnics 69 (2015) 452461

Soil Rock

(a) (b)
Fig. 5. (a) Soil layer overlying inclined rock layer, and (b) Cross section at the middle of the block.

Fig. 6. Geometries of the circular concrete shaft with one circular hole and two
circular holes.
3.5. FE Model verication
3.5.1. Verication with experimental results
RWH Engineering, a deep foundation construction company in
Ontario, Canada, which has carried out the design and construction
of many vertical shafts, provided monitoring data for verication. A
6 m diameter and 20 m deep secant pile shaft which was part of
the Waterloo Landll Pumping Station No. 2 in Ontario was used
for FE model calibration. The depth of excavation inside the shaft
was 18.2 m and overlapping of piles was 0.45 m, which resulted
in a wall thickness of 0.84 m. Fig. 10 shows some images depicting
the nal construction stage of the secant pile compression ring
shaft. Inclinometers were installed in two locations on the
opposing sides of the shaft walls and the readings in the radial
and circumferential directions were recorded. The monitoring data
for this shaft were utilized in this study to verify the present FE
model.
The material properties used in this FE analysis (Table 2) were
determined based on the factual data presented in the LVM
geotechnical investigation report dated April 28, 2011. The ground
stratigraphy considered in the FE analysis comprised a single layer
of cohesive deposit extending to the entire depth of the shaft. An
effective angle of internal friction of 27 was assumed in accor-
dance with Kulhawy and Mayne [13]. A drained analysis was per-
formed. The concrete compressive strength used for the shaft was
taken as 10 MPa. Low concrete modulus assumed in the analyses
reects practical conditions. As the circular secant pile wall exhi-
bits a compression ring behavior, which results in a purely com-
pressive, small magnitude stresses on the wall sections, low
quality concrete is typically used for project economy (see
Figs. 11 and 12). The changes in the shaft diameters calculated
using the FE model were compared with the monitoring results
as shown in Fig. 11, where the positive value refers to the inward
shaft movement. A fair agreement between the numerical and
monitoring results at low shaft depths was observed while there
was a good agreement in the maximum wall deformation,
0.2 mm. The results of model calibration indicated that the FE
model was able to replicate the soilstructureinteraction taking
place between the secant pile wall and surrounding soil.
3.5.2. Verication with analytical results
The circumferential (hoop) stresses and the lateral earth pres-
sure along the shaft predicted by the FE model were also compared

Table 2
Material properties for soil, rock and concrete.

Material h E (Mpa) c (kg/m3) v c (MPa)

Soil 1 40 50 1900 0.25 0
Soil 2 35 30 1800 0.3 0
Soil 3 30 10 1700 0.35 0
Bedrock 35 200 2500 0.3 0.2
Fig. 7. Normalized pressures for constant vs. varying soil elastic modulus (E) along
Concrete 15,000 2300 0.25
the shaft.
 A. Chehadeh et al. / Computers and Geotechnics 69 (2015) 452461
Fig. 8. Free mesh algorithm used for (a) soil using 3D tetrahedral solid elements, and (b) shaft using 3D triangular shell elements.
with the analytical calculations. The analytical approach used to
calculate the lateral earth pressure (qlateral) is explained by the fol-
lowing well-known equation:
qlateral K  c  z
3
where c = 21.5 kN/m is the soil unit weight, z is the shaft depth,
K = 1  sinh is the lateral earth pressure coefcient where the soil
friction angle h is taken as 30. The maximum lateral earth pres-
sures in both the analytical and FE analyses were around 200 kPa
at 18 m depth as shown in Fig. 12(a). The pressure distribution var-
ied linearly along the shaft depth except for a small deviation that
occurred at about one third of the total shaft height from its base.
The small deviation may be attributed to numerical instabilities
existing in the FE simulations.
The circumferential stresses r2 which represent the hoop stres-
ses through the circular wall were calculated analytically using Eq.
(3), where a is take as the outer radius (a = 3.84 m), b is the inner
radius (b = 3 m), and x is the point of interest (x = 3 m). It is very
clear from the comparison presented in Fig. 12(b) that the circum-
ferential stresses increase with the shaft depth, where a maximum
value of 1 MPa was recorded at the shaft base. The hoop stress rep-
resents the compression effect on the wall thickness which is basi-
cally resisted by the compressive strength of the concrete.
4. Results and discussion
4.1. Earth pressures on circular shaft (without holes)
The earth pressures, p, determined in this FE study were nor-
malized (p/Rc) with respect to the soil density, c, and the shaft
radius, R, and were plotted against various normalized depth ratios
(h/R) in order to be consistent with the depiction of similar data in
the literature. Fig. 13 presents the variation of normalized lateral
earth pressures for three different soil types (see Table 2) and for
up to a maximum normalized depth ratio of h/R = 8, which covers
most practical shaft dimensions. The result indicate that the calcu-
lated earth pressures can be represented by at rest pressure distri-
bution for / = 30 soil. However, the earth pressures increased
from active condition at (h/R) of 03, to an at-rest condition up
to (h/R) of 8 for / = 40 soil.
It should be mentioned here that the circular shaft exhibited
compression ring behavior under balanced earth pressures and
457
resultant wall movements were not large enough to reduce the lat-
eral stresses from the at-rest condition to active condition for the
shaft diameter, thickness and material properties considered in this
study. The results depicted in Fig. 13 indicated that the practical
7
shaft dimensions and material properties resulted in very small wall
deformations that were not enough to generate active earth pres-
sure conditions and as expected no yielding of the retained soil
was observed due to these small deformations. For all analyzed
cases, the simulated deformations remained well below the defor-
mation criterion recommended by Fujii et al. [7], Imamura et al.
[11], Herten and Pulsfort [10] and Chun and Shin [3] for dense sands
and the one recommended by Tobar and Meguid [28] for loose sand.
4.2. Effect of hole excavation on the stresses in the circular shaft
A series of analyses were performed to study the impact of a
hole opening on the stresses generated within the shaft. The holes
were modeled at the middle of the shaft in order to eliminate the
boundary effects. The excavation of a hole in a circular shaft
resulted in compressive stresses that are larger than what would
be expected in a symmetrically loaded circular shaft (compression
only). The tensile stresses also occur in the vicinity of the holes
after the excavation of holes. The stress concentration factor
(rmax/r11), which is dened as the ratio of maximum stress in a
shaft without hole to the corresponding stress for a shaft with a
hole, was calculated and presented against the (r/R) ratios.
4.2.1. Single hole on the shaft
Fig. 14 shows the stress concentration factors calculated on top
of and on the side of the hole excavated on the wall of a circular
shaft for various (r/R) ratios. The theoretical stress concentration
factor for the compression zones above the hole was shown as 3
(innite plate solution). The innite plate solution indicates that
the hole results in tensile stresses on the sides of the hole with a
stress concentration factor of 1. These values were plotted in
Fig. 14 along with the results of nite element analyses. As can
be observed from Fig. 14, the analytical solution for determining
the stress concentration factor for a hole on an innite plate gives
constant values for compression and tension. The FE results
depicted in Fig. 14 showed that the stress concentration factors
were a function of (r/R) ratios for curved surfaces such as the walls
of circular shaft. As the (r/R) ratio approaches 0, the stress
458 A. Chehadeh et al. / Computers and Geotechnics 69 (2015) 452461

Mesh 1 Mesh 2 Mesh 3 Mesh 4 Mesh 5

(a)

Mesh 1 Mesh 2 Mesh 3 Mesh 4 Mesh 5

(b)
Fig. 9. Mesh sensitivity analysis (a) mesh congurations, and (b) contours of the radial displacement.

Fig. 10. Secant pile construction photos for the Waterloo Landll Pumping Station No. 2 in Ontario, Canada.

concentration factors converge to the theoretical value of 3 for the theoretical value of 1 reported for an innite plate and thus they
compression. Stress concentration factors reduce to values as low were unconservative.
as 1.5 for large (r/R) ratios. The stress concentration factors for ten-
sile stresses on the sides of the hole were calculated as 1.3 for low
4.2.2. Multiple holes on the shaft
(r/R) ratios and as high as 2 for large (r/R) ratios. The stress concen-
Fig. 15 shows the stress concentration factors calculated for com-
tration factors calculated for tensile stresses were larger than the
pression (on top) and tension (on the side) zones of the hole
 A. Chehadeh et al. / Computers and Geotechnics 69 (2015) 452461 459

0.25 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

FE results 0
0.2 =30deg Acve
Monitored data 1
Sha deformaon (mm)

=35deg Acve
=40deg Acve
0.15
2 =30deg At Rest
=35deg At Rest
0.1 3 =40deg At Rest
=30deg ABAQUS

h/r
0.05 4 =35deg ABAQUS
=40deg ABAQUS
0 5
0 5 10 15 20 25
-0.05 6

7
-0.1
Depth (m) 8
p/r
Fig. 11. FE model verication with measured shaft deformations.
Fig. 13. Normalized earth pressure distribution for three different soil properties.

excavated on the wall of a circular shaft for various normalized hole

radius (r/R) values considering two identical holes positioned at 90
to each other. The FE results depicted in Fig. 15 show that the stress
concentration factors were a function of normalized hole radius (r/R)
for curved surfaces such as the walls of circular shafts. As the (r/R)
ratios approach 0, the stress concentration factors approached 2.5
for the compression. Stress concentration factors reduce to values
as low as 1.3 for large values of (r/R) ratios. The stress concentration
factors for tensile stresses on the sides of the hole were calculated as
1.7 for low (r/R) ratios and as high as 2 for large (r/R) ratios. It can be
observed from the results in Fig. 15 that the existence of second hole
on the shaft wall reduced concentration factors for compression and
increased them for tension.
The effect of the distance between the holes on the stress con-
centration factors were also studied parametrically. Fig. 16 depicts
Fig. 14. Theoretical solution of hole on plate versus nite element results for a shaft
the variation of stress concentration factors with the hole-to-hole with a single hole.
distance. The results indicate that the values of stress concentra-
tion factors were almost independent of hole-to-hole distance.
The results of FE analyses that investigated the impact of a hole m = r, 2r, 3r, 4r, etc. The results indicated that the normalized com-
excavation on the wall of a circular shaft resulted in stress regimes pressive pressure was not affected by the distance m.
that differed signicantly from what would be expected for a com-
pression ring. Thus, the stress concentrations should be taken into
4.3. Effect of sloping bedrock on the stresses in the circular shaft
consideration in the design of circular shafts.
The existence of sloping bedrock, where the secant pile wall cir-
4.2.3. Effect of shaft boundary cular shafts are typically installed in/on if they are in reasonable
The holes are typically excavated near the bottom of the shafts, depths, may result in wall stresses that would be signicantly dif-
which are typically near the bottom of the shafts. The boundary ferent from what would be expected from a symmetrically loaded
effects on the stress concentration factors were studied. Fig. 17 circular shaft (compression only). Figs. 1820 depict the results of
shows the effect of the distance between the hole and the lower shaft analyses that assume bedrocks inclined at 10, 20 and 30 (cases 1,
boundary of the shaft and stress concentration factors for two differ- 2 and 3) along with the results of a plane bedrock case. The results
ent (r/R) ratios. The distance (m), the clear distance between shaft indicate that the circumferential stresses in the wall of circular
bottom boundary and the hole, is a function of hole radius (r), i.e., shaft were purely compressive and they increased with increasing

250 1.2
FE (ABAQUS) FE (ABAQUS)
Circumferenal Stress (MPa)
Lateral Earth Pressure (kPa)

1
200 Analycal approach Analycal approach
0.8
150
0.6
100
0.4

50 0.2

0 0
0 3 6 9 12 15 18 0 3 6 9 12 15 18
Depth (m) Depth (m)
(a) (b)
Fig. 12. FE model verications; (a) lateral earth pressures, and (b) circumferential stresses, along the depth.
460 A. Chehadeh et al. / Computers and Geotechnics 69 (2015) 452461

Circumferenal Stresses in Sha (MPa)

-3 -2 -1 0 1 2 3
0

10

h
15

20 Along Axis 1
Along Axis 2
Along Axis 3
Circumferenal Stresses for Plane Bedrock
25
Fig. 15. Theoretical solution of hole on plate versus nite element results for a shaft
with a two holes. Fig. 18. The circumferential stresses in the walls of circular shaft (bedrock
inclination A = 10).

5
Stress Concentraon Factor - Compression
4.5 Stress Concentraon Factor - Tension
Circumferenal Stresses in Sha (MPa)
Normalized Pressure ( max/11 )

4 -3 -2 -1 0 1 2 3
0
3.5

3
5
2.5

2
10
1.5
h

1
15
0.5

0
0 2 4 6 8 10 12 14 16 20 Along Axis 1
Clear Distance between TWO excavated holes (xr) Along Axis 2
Along Axis 3
Circumferenal Stresses for Plane Bedrock
Fig. 16. Effect of distance between holes on stress concentration factors for two 25
holes on circular shaft.
Fig. 19. The circumferential stresses in the walls of circular shaft (bedrock
inclination A = 20).
3.5
Normalized Compressive Pressure

3
Circumferenal Stresses in Sha (MPa)
-3 -2 -1 0 1 2 3
2.5 0
(max/11 )

5
1.5

1
r/R = 0.2 10
0.5
h

r/R = 0.4
0 15
0 2 4 6 8
Clear distance between hole and sha edge, m (xr)
20 Along Axis 1
Fig. 17. Effect of distance between hole and shaft boundary. Along Axis 2
Along Axis 3
Circumferenal Stresses for Plane Bedrock
25
depth. However, signicant variations were observed in the stres-
ses with various inclinations of bedrock. For example, Fig. 18 Fig. 20. The circumferential stresses in the walls of circular shaft (bedrock
depicts the variation of circumferential stresses along axes 1, 2 inclination A = 30).
and 3 for bedrock sloping at 10 (case 1). The results show that
the maximum compressive stresses along axes 1, 2 and 3 remained remain below the maximum compressive stress of the plane bed-
below the maximum compressive stress of plane bedrock case rock case. However, at shallow depths and along axis 1 (side with
(compression ring). The circumferential stresses in the shaft were deep soil) tensile stresses occurred.
observed to be almost fully compressive for case 1. Fig. 20 shows the variation of circumferential stresses along
Fig. 19 depicts the variation of circumferential stresses along axes 1, 2 and 3 for bedrock sloping at 30 (case 3). The results show
axes 1, 2 and 3 for bedrock sloping at 20 (case 2). The results show that the compressive stresses that occurred along axis 2 exceeded
that the maximum compressive stresses along axes 1, 2 and 3 the maximum compressive stress of the plane bedrock case
 A. Chehadeh et al. / Computers and Geotechnics 69 (2015) 452461
signicantly. On the other hand, signicant tensile stresses
occurred along the axes 1 and 3. It was observed that the largest
compressive stresses occurred along axis 2. Thus, it should be
noted that the circular shafts that are located on sloping bedrocks
with inclination of 20 and larger should be designed to account for
both the increased levels of compressive stresses as well as the
newly emerged tensile stresses resulting from the bedrock
inclination.
5. Summary and conclusions
Parametric numerical analyses were performed to investigate
the interaction between secant pile wall circular shafts and sur-
rounding soil with the consideration of hole excavation on the
shaft walls as well as the existence of sloping bedrock.
Three-dimensional total stress nonlinear nite element analyses
were performed using the commercial package, ABAQUS.
Elasto-plastic soil behavior of retained soil and a non-linear
soil-shaft interface were assumed in analyses. The following con-
clusions may be drawn from this study.
1. For the cases where no sloping bedrock existed, the results
showed that the calculated earth pressures can be represented
by at rest pressure distribution for soil with / = 30. However,
the earth pressures were at active condition between (h/a) of
03 and they increased to the at-rest value between (h/a) of
38 for soil with / = 40.
2. The circular shaft exhibited compression ring behavior under
balanced earth pressures and resultant wall movements were
not large enough to reduce the lateral stresses from at-rest con-
dition to active condition for the shaft diameter, thickness and
material properties considered in this study.
3. The stress concentration factors for compression (on top of the
holes) were found to be a function of (r/R) ratios for curved sur-
faces such as the walls of a circular shaft. It was observed that as
the (r/R) ratio approaches 0 (plane strain conditions), the stress
concentration factors approach the theoretical value of 3 for the
compression (solution hole in a plate). Stress concentration fac-
tors were seen to reduce to values as low as 1.5 for large values
of (r/R) ratios.
4. The stress concentration factors for tensile stresses (on the sides
of the holes) were calculated as 1.3 for low (r/R) ratios and
reached values as high as 2 for large (r/R) ratios. The stress con-
centration factors calculated for tensile stresses were larger
than the theoretical value of 1 (innite plate solution), making
the use of hole on the plate solution unconservative for the
design of holes on circular shafts.
5. For a shaft with two holes, the results indicated that the value of
stress concentration factor is almost independent of
hole-to-hole distance. The stress concentration factors were
also seen to be independent of the distance between hole and
shaft boundary. The existence of a second hole on the shaft wall
reduced concentration factors for compression and increased
them for tension.
6. Results showed that in case of sloping bedrock at the bottom of
circular shaft, signicant deviations from a compression ring
behavior was observed. The circular shafts that were installed
on sloping bedrocks with inclination angles larger than 20
should be designed to account for both the increased levels of
compressive stresses as well as the newly emerged tensile
stresses resulting from the bedrock inclination.
461
Acknowledgement
The contribution by HC Group and American University of
Sharjah is gratefully acknowledged.
References
[1] Ariizumi K, Kumagai T, Kashiwagi A. Behaviour of large-scale cylindrical earth
retaining wall. In: Int symposium on geotechnical aspects of underground
construction in soft ground, Balkema; 2000.
[2] Arai Y, Kusakabe O, Murata O, Konishi S. A numerical study on ground
displacement and stress. Comput Geotech 2008:791807.
[3] Chun B, Shin Y. Active earth pressure acting on the cylindrical retaining wall of
a shaft. South Korea Ground Environ Eng J 2006;7(4):1524.
[4] Coulomb C. Sur une application des regles maximis et minimis a quelques
problems de statique, relatives a larchetecture. Acad Sci Paris Mem Math phys
1776;12:34382.
[5] Duncan C, Donald M, Robert H. Shakedown of a thick cylinder with a radial
cross hole. J Press Vess Technol Trans ASME 2009;131. 011203-208.
[6] Dassault Systmes Simulia C. Abaqus reference manual, RI; 2011.
[7] Fujii T, Hagiwara T, Ueno K, Taguchi A. Experiment and analysis of earth
pressure on an axisymmetric shaft in sand. In: Proceedings of the 1994
international conference on centrifuge, Singapore; 1994. p. 7916.
[8] Goto S, Muramatsu M, Sueoka T, Saka F, Yabe H, Watanabe H. Ground
movement earth and water pressures due to shaft excavations. In: Int
symposium on underground construction in soft ground, Balkema; 1995.
[9] Gunn M, Satkunananthan A, Clayton C. Finite element modeling of installation
effects. Ice conference on retaining structures. Cambridge: Robinson College;
1992. p. 4655.
[10] Herten M, Pulsfort M. Determination of spatial earth pressure on circular shaft
constructions. Granul Matter 1999;2(1):17.
[11] Imamura S, Nomoto T, Fujii T, Hagiwara T. Earth pressures acting on a deep
shaft and the movements of adjacent ground in sand. In: Kusakabe O, Fujita K,
Miyazaki Y. (Eds.), Proceedings of the international symposium on
geotechnical aspects of underground construction in soft ground, Tokyo,
Japan; 1999. p. 64752.
[12] Jaeger J, Cook N. Fundamentals of rock mechanics. London: Chapman & Hall;
1979.
[13] Kulhawy FH, Mayne PW. Manual on estimating soil properties. Report EL-
6800, Electric Power Res. Inst., Palo Alto; 1990.
[14] Kim D, Cha M, Lee D, Kim K, Lee I. Earth pressure acting on vertical circular
shafts considering arching effects in soils. Tunell Technol 2009;12(2):12944.
[15] Lade P, Jessberger H, Makowski E, Jordan P. Modeling of deep shafts centrifuge
test. In: International conference on soil mechanics and foundation
engineering, Stockholm; 1981.
[16] Laczek S, Rys J, Zeilinski AP. Load capacity of thick walled cylinder with a radial
hole. Int J Press Vessels Pip 2010;87:4339.
[17] Li H, Johnston R, Meckenzie D. Effect of auto frottage in thick walled cylinder
with a radial cross bore. J Press Vess Technol, Trans ASME 2010;132. 011205-1.
[18] Makulsawatudum P, Meckenzie D, Hamilton R. Stress concentration at cross
holes in thick cylindrical vessels. J Strain Anal 2004;39:47181.
[19] Mohr O. Welche Umstnde bedingen die Elastizittsgrenze und den Bruch
eines Materials? Zeit des Ver Deut Ing 1900:152430.
[20] Moor ED. An analysis of bored pile/diaphragm wall installation. Geotechnique
1994;45(4):7535.
[21] Moor ED, Stevenson M. Evaluation of the performance of a multi-propped
diaphragm wall during construction. In: Int. symposium on geotechnical
aspects of underground construction in soft ground, Balkema; 1996.
[22] Muller-Kirchenbauer H, Walz B, Klapperich UH. Experimentelle und
theoretische untersuchungen zum erddruck-problem auf radial
symmetrische senkkasten und schachte[R]. Veroff. des Grundbauinstitutes
der TU Berlin; 1980.
[23] Nadai A. Theory of ow and fracture of solids. New York: McGraw Hill; 1950.
[24] Ng C, Yan R. Stress transfer and deformation mechanisms around a diaphragm
wall panel. J Geotech Geoenviron Eng 1998;124(7):63848.
[25] Nihous GC, Kinoshita CK, Masutani SM. Stress concentration factors for oblique
holes in pressurized thick walled cylinders. J Press Vess Technol, Trans ASME
2008;130. 021202-1.
[26] Roark RJ, Young WC, Budynas RG. Roarks formulas for stress and strain. New
York: McGraw-Hill; 2002.
[27] Shin Y, Sagong M. Ground pressure acting on cylindrical retaining wall of a
shaft in soft ground. Chin J Rock Mech Eng 2007;26.
[28] Tobar T, Meguid MA. Comparative evaluation of methods to determine the
earth pressure distribution. Tunn Undergr Space Technol 2010;25(2):18897.
[29] Wong RCK, Kaiser PK. Design and performance evaluation of vertical shafts:
rational shaft design method and verication of design method. Can Geotech J
1988;25(2):32037.