Displacement-Based Design for Deep Excavations

Downloaded from ascelibrary.org by Hong Kong Polytechnic University on 02/11/24. Copyright ASCE. For personal use only; all rights reserved.

W. Allen Marr1 and Martin Hawkes2

1
Geocomp Corporation, 1145 Massachusetts Avenue, Boxborough, MA 01719; PH (978)
635-0012; FAX (978) 635-0266; [email protected]
2
Geocomp Corporation, 1145 Massachusetts Avenue, Boxborough, MA 01719; PH (978)
635-0012; FAX (978) 635-0266; [email protected]

ABSTRACT. Many structures located close to excavations are damaged by

excessive ground movements during construction. This is partly due to limitations of
commonly used methods to design the excavation support system. The authors
propose using a displacement-based design method that focuses on designing to keep
movements within allowable values. A step-by-step approach for the method is
described and illustrated for a deep excavation into soft clay.
INTRODUCTION
Limiting movements that result from constructing deep excavations is becoming a
significant design issue, especially in urban environments. Failure to control
deformations can cause significant damage to adjacent structures and utilities.
Consequences such as the following can become very costly:
 Collapse of Excavation Support System (ESS) with major damage and delays.
 Loss of bearing capacity of shallow foundations for an adjacent building.
 Cracking of adjacent buildings that cost time and money to resolve.
 Loss of factor of safety for basal heave, global instability, or bearing capacity
such that work must be stopped until remedial measures are implemented.
 Shut down of the project and withdrawal of permits by government officials
until a resolution can be found.
 Abandonment of the project, leaving an open hole in the ground.
 Poor relations with neighbors, which take management time.
 Lawyers and litigation.
Design is becoming more challenging due to increasing requirements for deeper
excavations on poorer sites, tighter limits on allowable displacements for adjacent
structures, and new methods of construction that extend beyond the experience base
used to develop historical design methods. Increasingly, designs are controlled by the
need to limit movements, which goes beyond the traditional approach of focusing on
required support loads and avoiding collapse. These conditions require a new
approach to design of excavation support systems, one that focuses on controlling
displacements.
Designing to limit movements places the design emphasis exactly where it should
be: How much movement is allowable? What is the optimal design to keep
movements within the allowable values? What mitigation steps are possible if the
allowable movements are exceeded?
Construction of a deep excavation involves unloading of the soil. For unloading,
the movements of soil are small until the factor of safety for any soil failure

82
Copyright ASCE 2010 2010 Earth Retention Conference (ER2010)
Earth Retention Conference 3
 EARTH RETENTION CONFERENCE 83
Downloaded from ascelibrary.org by Hong Kong Polytechnic University on 02/11/24. Copyright ASCE. For personal use only; all rights reserved.

mechanism drops below 1.1 to 1.2. Then the movements increase rapidly and the
factor of safety decreases quickly. This condition can worsen rapidly as the
excavation becomes deeper, a fact that catches many people by surprise and leaves
little time to take preventive actions.
Designing to control movements became possible with the development of non-
linear finite element analysis (FEA) in the early 1970’s. However, it was not a
practical design tool in its early versions. In recent years, finite element software has
greatly improved. There are more realistic stress-strain models, more stable numerical
methods, tools to ease the creation of the geometric model, and especially tools to
provide powerful and useful graphical output for quick interpretation and presentation
of results. Some products also compute a factor of safety against soil failure at any
stage of the excavation. When done correctly, this method automatically gives the
most critical failure mode for each level of excavation, whether it is global instability,
basal heave or localized bearing capacity. FEA calculates displacements of all types
at all locations, making a displacement-based design approach now possible. The
ease-of-use of these programs allows quick parametric studies to examine sensitivity
of the design to changes in key parameters and feasibility of various options to
optimize the design. Additionally, the pool of engineers capable of using these
programs has greatly increased. A large portion of graduating geotechnical engineers
enters practice with FEA skills.
CAUSES OF DISPLACEMENTS

Figure 1 shows the various failure modes for a braced excavation. Similar failure
modes occur for most types of lateral supports including struts, rakers, tiebacks and
soil nails. To limit large movements from any of these failure mechanisms, it is
necessary to provide an adequate factor of safety for each.
Excessive movements can occur without a failure mechanism occurring.
Prefailure movements can result from:
 Elastic displacements of soil and support system.
 Plastic displacements of soil without stability failure.
 Strains in the structural support system due to lower stiffness or strength, load
redistribution, or temperature changes.
 Slippage and give at the structural connections.
 Consolidation of some or all of the soil.
 Local loss of ground due to flow into the excavation.
 Installation of components behind the wall (tiebacks, soil nails, grouting for
water control) that result in loss of soil from behind the supporting wall.
The first five in this list can be evaluated with geotechnical FEA programs. The last
two cannot with any reliability. The best approach to preventing movements from
water flow and ground losses is to adopt construction practices that keep these
mechanisms from occurring. Once these mechanisms begin they are unpredictable
and difficult to control.

83
Copyright ASCE 2010 2010 Earth Retention Conference (ER2010)
Earth Retention Conference 3
 84 EARTH RETENTION CONFERENCE
Downloaded from ascelibrary.org by Hong Kong Polytechnic University on 02/11/24. Copyright ASCE. For personal use only; all rights reserved.

Figure 1: Failure modes for excavation support systems

A common practice in the US is for the designer of the excavation support

system, usually a structural engineer, to ask a geotechnical engineer for design earth
pressures. The geotechnical engineer will use the geotechnical data he has to develop
a design earth pressure diagram. The earth pressure diagram is often based on forces
computed from Rankine active and passive earth pressures or from an “equivalent
fluid” pressure diagram based on local practice. The method based on Rankine earth
pressures assumes the earth deforms to a failure state. This cannot happen if
movements are to be limited. The equivalent fluid pressure method gives no
indication of what movements might occur. Others may use the Terzaghi et al (1996)
apparent earth pressure diagrams (hereafter referred to as the TPM method) which are
considered to give an upper bound to the forces for design of the lateral support
system. However, no guidance was given on using these diagrams when movement
is to be controlled.
PAST APPROACHES TO PREDICTING DISPLACEMENTS
Using a displacement-based design approach requires a way to predict
displacements for trial designs. For many years predicting displacements was
empirical. Guidelines such as these were used:
 Ignore displacements in the design, and then modify construction if necessary.
 Use local experience from measurements on previous projects of similar type.
 Use rules of thumb such as “maximum lateral displacement may be up to
0.005 times the depth of the excavation, H.”

84
Copyright ASCE 2010 2010 Earth Retention Conference (ER2010)
Earth Retention Conference 3
 EARTH RETENTION CONFERENCE 85
Downloaded from ascelibrary.org by Hong Kong Polytechnic University on 02/11/24. Copyright ASCE. For personal use only; all rights reserved.

 Limit displacements to elastic values by keeping the base stability number,

/ less than 4. γt is average total unit weight above
bottom of excavation, qs is the surface load and sub is the average undrained
strength below the bottom of the excavation.
 Make estimates based on prior experience of summarized performance. For
example, from NAVFAC DM-7 (1982):
 Walls in sands and silts might displace laterally up to 0.002H.
 Walls in stiff clays might displace laterally up to 0.005H.
 Walls into soft clays might displace laterally up to 0.02H.
 Walls into very soft clays might displace laterally more than 0.02H
 Pre-loading the lateral supports can reduce these values by up to half.
 Poor construction methods can increase these values significantly.
 Maximum settlement of outside ground about ½ to 1 times the maximum
horizontal displacement of wall. However, if significant soil volume
reduction might result from nearby vibrations of loose sands or consolidation
of soft clays, a separate evaluation is required.
A major step forward occurred with Dr. Peck’s state-of-the art paper at the 7th
ICSMFE in Mexico (Peck, 1969). He compiled data from instrumented excavations
located around the world and created charts showing settlement divided by depth of
excavation versus distance from the supporting wall. He did the same for maximum
horizontal displacement. He divided the data into three sets, varying from sand and
hard clay to very soft clay. He differentiated the different sets using soil types and
the base stability number. His work gave a rational basis to estimate the maximum
lateral wall displacements and settlement behind the wall for different soil types. His
results provided recommended envelopes of maximum displacement. Consequently,
it generally should over predict movements where design and construction are
managed to limit movements. Additionally, many of the construction methods used
today, particularly stiffer walls and struts with preloading, wall embedment, pre-
stressed tiebacks and soil nails, were not a part of his database. His work is still a
valuable set of reference cases to evaluate the reasonableness of newer methods.
With the advent of computers in civil engineering in the middle 1960’s, a
numerical solution using the theory of “beam on elastic foundation” was developed to
compute the stresses, moments and displacements of a wall.” This approach became
widely used particularly by structural engineers. Soil behavior was bundled into a set
of springs that represented soil as an elastic material. It was never clear how to
determine the values of spring constants to represent realistic soil behavior. The
method typically uses Rankine active and passive earth pressure coefficients;
however, excavation support systems designed to limit movements don’t develop
Rankine earth pressure conditions. Due to the nature of the method, the factor of
safety of the excavation against global instability may be overlooked, resulting in
severe consequences. This method is increasingly discredited as an applicable design
tool (Poulos, 2000).
In the early 1970’s, finite element programs became available that could model
the non-linear stress-strain behavior of soils to include elastic and plastic components.
One of the early applications of these new methods was to predict displacements for

85
Copyright ASCE 2010 2010 Earth Retention Conference (ER2010)
Earth Retention Conference 3
 86 EARTH RETENTION CONFERENCE
Downloaded from ascelibrary.org by Hong Kong Polytechnic University on 02/11/24. Copyright ASCE. For personal use only; all rights reserved.

deep excavations (Clough et al, 1972 and Jaworski, 1973). After a few rounds of
improvements, FEA was used to develop parametric charts that predict the maximum
displacements resulting from an excavation. Principal among these were papers by
Clough and his many co-authors. These culminated in the Clough et al (1989) paper
with additional considerations by Clough and O’Rourke (1990). The key figure from
Clough et al (1989) is reproduced in Figure 2. The figure includes some added data
symbols that will be discussed later. This figure was especially significant because it
showed the important relationships among factor of safety, system stiffness, and
depth of excavation as the determinants of maximum displacement. The paper
referenced other papers that provide approximate ways to account for other
significant factors such as preloads, vertical strut spacing, lateral support stiffness,
support preload, anisotropy in soil shear strength, variable soil conditions, water
pressure, and construction influences. This chart uses Terzaghi’s (1943) factor of
safety against basal heave defined as ∙ where Terzaghi’s Nc
∙
value is typically replaced by Skempton’s (1951) Nc (see Figure 3). suu is the average
shear strength above the bottom of the excavation. f equals 1/D if D < 0.7B and
equals 1/0.7B if D > 0.7B where B is the width of the excavation and D is the
distance from the bottom of the excavation to firm soil.

Figure 2: Design Curves to Obtain Maximum Lateral Wall Movement for

Soft to Medium Clays (from Clough, et al 1989)
This chart is widely used to calculate displacement for a design. Clough et al
(1989) described it as a first estimate tool. It has limitations because the effects of
many factors on displacement are reduced to those in the chart. It uses a conservative

86
Copyright ASCE 2010 2010 Earth Retention Conference (ER2010)
Earth Retention Conference 3
 EARTH RETENTION CONFERENCE 87
Downloaded from ascelibrary.org by Hong Kong Polytechnic University on 02/11/24. Copyright ASCE. For personal use only; all rights reserved.

upper bound of the data as design values. It

is limited to the types of cases considered in
its formulation, i.e. soft soils with undrained
shear strength increasing with depth from
the top of the ground. Many cases fall
outside the conditions of the chart. The
chart also focuses on maximum lateral wall
movement and maximum settlement outside
the wall but does not give angular distortion
or horizontal strain that are required to
assess impact of movements on adjacent Figure 3: Skempton’s (1951)
structures. With the challenges facing Bearing Capacity Values
designers and contractors to create less
costly designs that meet stricter performance requirements in more complex
subsurface environments, we need an improved way to make accurate predictions of
displacement, horizontal strain and angular distortion that result from the construction
of deep excavations with nearby structures.
What is needed is a design method that is centered on controlling displacements
to values that minimize damage caused by movements. Modern FEA software
designed for geotechnical applications gives this capability.

PARAMETERS FOR FINITE ELEMENT ANALYSIS

Finite element analysis requires more information than traditionally used to
design the ESS. A frequent argument is that this additional information is not
available or will be too expensive to obtain. In consideration of the potential risks
created by construction deep excavations, the cost to obtain the added information is
small and is good engineering practice. Required parameters for FEA are:
 Geometry of the site, plan for the excavation, possible methods for ESS, and
proposed sequence of work for construction expediency.
 Subsurface information to establish a model of the subsurface conditions that
realistically simulates the actual conditions.
 Groundwater conditions with depth and as planned during the work.
 Soil parameters – density, Atterberg limits, drained and undrained shear
strength, permeability and stiffness.
 Desired wall and lateral support methods and materials (such as struts or
sheeting that already exist for the project or are readily available from
suppliers).
 Desired lateral support type and spacing both vertically and horizontally for
construction expediency.
Water pressure has a direct effect on strength of soils, pressure against the wall,
and uplift of soil in the bottom of the excavation. Many situations are not hydrostatic
before the work starts and change during the work. These conditions can be modeled
in FEA, provided the initial water condition is known. At a minimum, depth to
groundwater, pore pressure at the bottom of the excavation and pore pressure in any
pervious layer up to 0.5H below the base of the excavation should be known.

87
Copyright ASCE 2010 2010 Earth Retention Conference (ER2010)
Earth Retention Conference 3
 88 EARTH RETENTION CONFERENCE
Downloaded from ascelibrary.org by Hong Kong Polytechnic University on 02/11/24. Copyright ASCE. For personal use only; all rights reserved.

Soil properties, especially strength, require careful attention because they are a
key factor in the overall performance of the excavation support system. Soil strength
below the bottom of the excavation becomes increasingly important for Nb values
greater than 4. Special care should be used to define strength of low strength soils
when Nb is greater than 6. There are at least a dozen methods to measure undrained
shear strength and each method gives a different result. A qualified geotechnical
engineer with experience in the design of deep excavations should soil strength
parameters for the design.
Soft to medium clays may exhibit strength anisotropy, i.e. strength varies with the
orientation of the failure surface. Very few FEM programs available in the US market
can model strength anisotropy. Clough and Hansen (1981) showed that strength
anisotropy can be considered by using the average of strengths measured with triaxial
compression and triaxial extension tests on undisturbed samples or strengths
measured in a direct simple shear device. These conclusions are consistent with
Ladd’s SHANSEP approach (Ladd and Foote 1974, Ladd 1991) and result in more
realistic values of undrained strength for the FEA. Undrained strength values can also
be obtained with cone penetration tests or corrected field vane tests. For projects
where the consequences of a wrong prediction are high or there is little experience
working with the soils, the stress path method by Lambe and Marr (1979) to
determine soil strength and stiffness can be very helpful.
Table 1: Stiffness values for soils (modified from AASHTO 1996, 2002)
Soil Soil Description Range of E50 (kPa) Range of E50 (ksf)
Type
Clay  Soft sensitive  2,500 to 15,000  50 – 300
 Medium stiff to stiff  15,000 to 50,000  300 - 1,000
 Very stiff  50,000 to 100,000  1,000 –2,000
Loess  15,000 to 60,000  300 – 1,200
Silt  Silts, sandy silts, slightly cohesive  2,000 to 20,000  40 – 400
mixtures [400 (N1)60] [8 (N1)60]
Fine  Loose  8,000 to 12,000  160 – 240
sand  Medium dense  12,000 to 20,000  240 – 400
 Dense  20,000 to 30,000  400 – 600
[Clean fine to medium sands and [700 (N1)60] [14 (N1)60]
slightly silty sands]
Sand  Loose  10,000 to 30,000  200 – 600
 Medium dense  30,000 to 50,000  600 – 1,000
 Dense  50,000 to 80,000  1,000 – 1,600
[Coarse sands with little gravel] [1,000 (N1)60] [20 (N1)60]
Gravel  Loose  30,000 to 80,000  600 – 1,600
 Medium dense  80,000 to 100,000  1,600 – 2,000
 Dense  100,000 to 200,000  2,000 – 4,000
[Sandy gravels and gravels] [1,200 (N1)60] [24 (N1)60]

Strength values for stiff to very stiff clays and for silts and sands are less critical
to the design of excavation support systems with soft clay below the bottom. For
most cases, values can be estimated from empirical correlations with SPT tests or
cone penetration tests. For a project involving stiff to very stiff clays, or silts and

88
Copyright ASCE 2010 2010 Earth Retention Conference (ER2010)
Earth Retention Conference 3
 EARTH RETENTION CONFERENCE 89
Downloaded from ascelibrary.org by Hong Kong Polytechnic University on 02/11/24. Copyright ASCE. For personal use only; all rights reserved.

sands with no soft cohesive soils, strengths can be estimated from SPT tests or cone
penetration tests and local experience.
Finite element methods require values of
soil stiffness to make reasonable predictions
of displacement; however, soil stiffness is less
important than soil strength for values of Nb
above 4. Soil stiffness values can be
reasonably estimated from the values given in
Table 1. Stiffness data can often be
determined from the data used to obtain
design strength. For clays, Figure 4 from
Duncan and Buchignani (1976) is very useful.
K is E50/su. Values from Table 1 and Figure 4
are used as the secant Young’s modulus at a
shear stress of half the shear strength, E50
which is twice the initial tangent modulus.
The unloading modulus is typically 3 to 5
times E50. Kulhawy and Mayne (1990) and Figure 4: Generalized Undrained
Modulus Ratio for Clays
Sabatini et al (2002) contains extensive (after Duncan and Buchignani 1976)
compilations of methods to estimate or
measure stiffness and strength.
PROPOSED APPROACH TO DISPLACEMENT-BASED DESIGN
The proposed approach is to determine the allowable displacements that minimize
potential damage, establish a trial design and analyze the performance of this design
with FEA. Then compare results to the allowable displacements, revise the design
and rerun the FEA until the predicted displacements are less than allowable values.
The FEA software must have the following capabilities:
 Model non-linear stress-strain behavior for soils, including drained and
undrained soil behavior for loading and unloading stress paths.
 Compute factor of safety against global instability.
 Model structural components of the wall and support system and their
interaction with the soil, including slippage between soil and wall.
 Compute groundwater pressures and their change with time.
 Support removal of elements and correctly adjust nodal forces.
 Model the sequence of excavation to closely follow the steps of dewatering,
excavation, support installation and support pre-stressing.
Current programs used in the US with these capabilities include PLAXIS, FLAC,
SIGMA/W, midasGTS, CRISP, and others.
The following sections describe step-by-step guidelines to establish allowable
displacements, then develop a design that deforms less than these values.
Allowable Displacements
Designing to limit movements requires knowledge of how much each structure
within the influence zone can deflect without incurring excessive damage. Very few
building codes give explicit limits for allowable displacements and these limits are

89
Copyright ASCE 2010 2010 Earth Retention Conference (ER2010)
Earth Retention Conference 3
 90 EARTH RETENTION CONFERENCE
Downloaded from ascelibrary.org by Hong Kong Polytechnic University on 02/11/24. Copyright ASCE. For personal use only; all rights reserved.

not quantitative. Exceptions are codes for Shanghai, South Korea and some railway
agencies in the US. Most of the codes that address this issue put the responsibility on
the Builder to avoid damaging all neighboring structures. If not thoroughly familiar
with local Code requirements, it is best to consult with a construction attorney
familiar with these requirements.
Displacement limits depend on several factors, including the details of the
foundation, loads, type of structural framing and exterior shell, and the existing
condition of these elements. Many older buildings may have already undergone
significant movement such that floor joists have limited bearing areas. For others, the
structural members may have been significantly weakened by rot or insects, or their
condition may be totally unknown because they are not visible for inspection. The
limit values also depend on the consequences of any significant movement. A
building where movement of one inch might cause a collapse and loss of life will
Table 2: Allowable Settlement and Tilt of Structures

Type of
Limiting Factor Maximum Settlement
Movement
Drainage 150 – 300 mm
Access 300 – 600 mm
Total Masonry walled structure 25 – 50 mm
settlement Framed structures 50 – 100 mm
Smokestacks, silos, mats 75 – 300 mm
Tilting of smokestacks, towers 0.004L
Stacking of goods, rolling of trucks, or similar 0.01L
Machine operation-cotton loom 0.003L
Machine operation – turbogenerator 0.0002L
Crane rails 0.003L
Drainage of floors (0.01 to 0.02)L
Framed buildings and reinforced load bearing walls:
Structural damage 1/150(1) 1/250(2) 1/200(3)
Cracking in walls and partitions 1/300(1) to 1/500(2)
Open frames 1/300(6)
In filled frames 1/1000(6)
Framed buildings 1/300(7)
Tilting/
High continuous brick walls (0.0005 to 0.001)L
Differential
movement One-story brick mill building, wall cracking (0.001 to 0.002)L
Plaster cracking (gypsum) 0.001L
Reinforced-concrete building frame (0.0025 to 0.004)L
Reinforced-concrete building curtain walls 0.003L
Steel frame, continuous 0.002L
Simple steel frame 0.005L
Unreinforced load bearing walls:
Sagging 1/2500(2)
L/H < 3; 1/3500 – 1/2500(3)
L/H < 5; 1/2000 – 1/1500(3)
1/2500 at L/H = 1(5)
1/1250 at L/H = 5(5)
Hogging 1/5000 at L/H = 1(5)

90
Copyright ASCE 2010 2010 Earth Retention Conference (ER2010)
Earth Retention Conference 3
 EARTH RETENTION CONFERENCE 91
Downloaded from ascelibrary.org by Hong Kong Polytechnic University on 02/11/24. Copyright ASCE. For personal use only; all rights reserved.

1/2500 at L/H = 5(5)

(1)
Skempton &Macdonald (1956), 2 Meyerhof (1956),
( )
Note: Data from Sowers (1962) unless otherwise indicated
(3)
require
Polshin much (1957), (4)Bjerrum
& Tolkarsmaller than a(5)Burland
limits (1963), building where
& Wroth six (6)inches
(1975), wall(7)movement
Meyerhofof(1953), has
Grant et al (1974).
no consequence to anything or anyone.
There are a number of publications on the allowable settlement and tilt for
different types of structures. Table 2 summarizes recommendations from some of
these as a useful reference to help set limit values for displacements. Boscardin and
Cording (1989) developed a plot relating angular distortion and horizontal strain to
building damage. Their plot is provided in Figure 5. It is useful in establishing
allowable values for a structure.

Figure 5: Angular Distortion and Horizontal Strain Limits

(after Boscardin and Cording, 1989)

The recommended approach to establishing displacement limits is as follows:

1. Review project information, local codes, ordinances and the Owner’s
requirements. Call these values ΔHmax1 for maximum allowable lateral movement
of the wall, ΔVmax1 for maximum allowable settlement behind the wall, εH max1 for
maximum allowable horizontal strain in of building behind the wall and Δαmax1
for maximum allowable angular distortion of building outside the wall.
2. If displacements limits are not provided in Step 1, perform a site-specific
evaluation. An engineer qualified in structural assessment of constructed facilities
should inspect each building and utility within 4H of the ESS to determine its
present condition and tolerance for additional settlement, horizontal strain and
angular rotation. A larger zone should be evaluated if any significant alteration to
the groundwater conditions outside 4H is anticipated. The result of this work
should be limits for horizontal displacement of the ESS and settlement, angular
distortion and horizontal strain of the ground outside the excavation for each
structure. Call these values ΔHmax2, ΔVmax2, εH max2 and Δαmax2.
3. If specific displacement limits are not provided from Steps 1 or 2, use the
following method for each structure that might experience damage:
a. Determine maximum allowable angular distortion, Δαmax3, from Table 2
depending on building type and condition. Poor condition will lower the
allowable values in Table 2.

91
Copyright ASCE 2010 2010 Earth Retention Conference (ER2010)
Earth Retention Conference 3
 92 EARTH RETENTION CONFERENCE
Downloaded from ascelibrary.org by Hong Kong Polytechnic University on 02/11/24. Copyright ASCE. For personal use only; all rights reserved.

b. Enter chart by Boscardin and Cording (1989), Figure 5, with allowable

angular distortion and allowable damage to determine maximum allowable
horizontal strain, εHmax3.
c. Multiply εHmax3 by 3H for an estimate of the maximum allowable horizontal
displacement of the wall that limits building damage, ΔHmax3.
d. Estimate maximum allowable ΔVmax3 from values in Table 2 or from
∆ ∆ ∙ 0.16 ∙ . which is based on Clough et al (1989).
4. Take the smallest values from Steps 1-3 to obtain ΔHall, ΔVall, εH all and Δαall.
Establish FEA Input
Finite element analysis is a very useful tool for design of ESS, but considerable
care must be used by someone familiar with the software, its limitations and with the
behavior of soils. The following gives a set of steps that can produce a trial design
with 1 or 2 re-runs of the software once the finite element geometric model is correct.
1. Establish the subsurface conditions, including material properties. Especially
important are strength and stiffness of soils below the bottom of the excavation
and groundwater conditions before and during excavation. See section
“Parameters for FEA” for guidance. Use expected value for each parameter.
2. Layout an ESS that works for constructability, i.e. type of wall, type of supports
and horizontal and vertical locations for lateral supports. People with experience
in constructability of ESS should participate in this step.
3. Compute FOSBH using above equation. If less than 1.5, extra attention should be
given to reduce any uncertainty about the shear strength of the soil below the
excavation.
4. Use Terzaghi et al (1996), reproduced in Figure 6, to estimate lateral support
loads per unit length of wall. Note that TPM for soft to medium soils has been
revised from original Terzaghi and Peck method with the addition of ΔK based on
. .
Henkel (1971). ∆ 1 ). “d” is the lesser of 1.41B and D. ΔK
can increase K for soft soils considerably. Use TPM lateral support loads and
allowable axial stress to compute the structural area, As, for each support,
including the effect of inclination of rakers or tiebacks.

Figure 6: Apparent Earth Pressure Diagrams from Terzaghi et al (1996)

92
Copyright ASCE 2010 2010 Earth Retention Conference (ER2010)
Earth Retention Conference 3
 EARTH RETENTION CONFERENCE 93
Downloaded from ascelibrary.org by Hong Kong Polytechnic University on 02/11/24. Copyright ASCE. For personal use only; all rights reserved.

5. Compute stiffness of each support per unit length of the wall using ks=EsAs/L,
where ks is the stiffness of the support member per unit of length, Es is the
modulus of elasticity for the support member, As from Step 4, and L is the length
of the support member (half length for struts).
6. Compute moments in wall between each strut level and between lowest strut and
the bottom of the excavation assuming hinges at each support location and at the
bottom of the excavation. This can be done using M=wl2/8, where M is the
maximum moment, w is the average stress acting on the segment of length, l.
Take the largest value to size the wall. For cantilevered walls and walls with
embedment of more than the average strut spacing, h, the maximum moment will
be below the bottom strut and should be computed using other methods such as
those given in NAVFAC DM 7.2 (1982).
7. Use the maximum moment to compute required EI of wall so that allowable
bending stress is not exceeded. Compute I=M*y/σallowable where y is taken as ½
the thickness of the wall. Select thickness of wall based on required EI and wall
type that contractor wants to use. I is moment of inertia of the wall.
8. Use Figure 2 with ΔHall/H and FOSBH to estimate the required system stiffness,
(EI/γt* h4avg).
9. With I from Step 8, compute average vertical support spacing, h. Compare this
spacing with that in Step 2 and adjust the spacing where constructability
considerations allow or increase I to achieve the required system stiffness.
10. Use resulting EI per unit length of wall as input stiffness for the wall.
Prepare the FEA Model
1. Establish the finite element model giving consideration to the soil layering, the
location of lateral supports, depth of the wall, the depth of each excavation stage,
and the locations of any external loads.
2. Input information from above and check that all is correct.
3. Input the construction sequence for the FEA. Include steps to calculate FOS for
the full excavation and for other levels where stability might be a concern.
Do the Finite Element Analysis
1. Make the finite element run.
2. Examine contour plots of stresses, strains and displacements for discrepancies,
anomalies and unusual patterns. If these are present, examine the input data and
results for each excavation step to locate the cause of the anomaly. Challenge any
result that contradicts engineering judgment. Correct errors and rerun the analysis
before proceeding to the next step.
3. Compare FEA forces in the supports to those calculated with TPM and the
allowable stresses. The important part of this step is that the total loads are
comparable and that any differences are understandable and explainable.
4. Compare the maximum moment in the wall to the value computed with TPM and
the allowable moment. The FEA value will typically be similar to the value from
TPM if the wall embedment is less than h. It may be much larger if the wall
embedment exceeds h.

93
Copyright ASCE 2010 2010 Earth Retention Conference (ER2010)
Earth Retention Conference 3
 94 EARTH RETENTION CONFERENCE
Downloaded from ascelibrary.org by Hong Kong Polytechnic University on 02/11/24. Copyright ASCE. For personal use only; all rights reserved.

5. Compare maximum horizontal displacement of the wall, maximum vertical

displacement of ground surface outside the excavation, horizontal strain at the
ground surface outside the excavation and angular rotation of the ground surface
to the values established in section “Allowable Displacements.”

Revise the Trial Design

1. Multiply the FEA lateral support loads by 1.3 to account for variability that TPM
indicates for support design. Adjust the size of lateral support members to meet
structural design requirements for these loads. Resize the wall dimensions to
satisfy structural code using the maximum moment from the FEA analysis.
2. Apply preload of 50% to 100% to supports, if further reduction in the computed
displacements is required.
3. If lateral displacement of the wall exceeds ΔHmax by more than a factor of 2,
increase the support stiffness by a factor of 5. If FEA maximum settlement
outside the wall is considerably more than the allowable value and more than the
FEA maximum horizontal displacement of the wall then, consider increasing the
embedment of the wall.
4. Repeat the steps under “Do the Finite Element Analysis.” Multiple adjustments
and repeats may be required to find the optimal design. A point may be reached
where further increases in strut or wall stiffness will not be very effective and the
support loads will increase significantly. If the displacements remain larger than
the allowable values, the value of h may have to be decreased.
These steps do not consider the effects of uncertainties in the soil profile and
properties. The ease-of-use of modern FEA programs allow multiple parametric
studies and reliability assessments. The computed deflections, loads, stresses,
moments and FOS from these analyses must be less than the allowable values.
Structural Design
Following TPM, the axial capacity of the lateral supports computed in the FEA
should be increased by 33% to account for individual variability that can occur in any
one support. The envelope of maximum values computed with the FEA can be used
as the design moments and shear forces for design of the wall.
Some agencies require an evaluation of the removal of any one lateral support
element. This can be simulated in a two-dimensional FEA by removing the lateral
support one level at a time and examining the effects on the moments and shear in the
wall and forces in the lateral supports. This is a conservative approach. The FEA
forces in the adjacent lateral supports need not be increased by 33% for design since
this is a check on an extreme load condition.
Temperature changes can greatly increase strut loads in hot conditions and
decrease them with possible additional movements in the cold. These need to be
considered in the structural design or steps be taken to reduce temperature changes.
Freezing of ground behind the all can also create large forces in the lateral support
system and wall. This condition cannot be reliably analyzed and should be avoided.
Settlement and bearing capacity failure of the wall base can cause additional
movement when lateral supports are inclined or vertical load is added to the wall.
The FEA will detect this condition and automatically consider it in the analysis.

94
Copyright ASCE 2010 2010 Earth Retention Conference (ER2010)
Earth Retention Conference 3
 EARTH RETENTION CONFERENCE 95
Downloaded from ascelibrary.org by Hong Kong Polytechnic University on 02/11/24. Copyright ASCE. For personal use only; all rights reserved.

DEMONSTRATION OF THE DISPLACEMENT-BASED APPROACH

Figure 7 shows a project where the Owner wants to construct a four-level parking
garage. He would be elated with five levels. The site conditions are typical of some
locations in the US but have been generalized to avoid any similarity to actual
projects. The four-level garage requires excavation to 42.5 ft depth and the five-level
garage requires a 52.5 ft deep excavation. Older design methods usually limited
excavations in these conditions to two or three levels. An evaluation of nearby
structures using the approach described above results in allowable displacement
values of ΔHmax < 1 inch; ΔVmax < ¾ inch; Δαmax < 0.001 and Δεmax < 0.001. Below
the excavation is a normally consolidated soft clay with shear strength at a depth of
45 feet of 620 psf increasing to 1260 psf at 100 ft deep. It is not clear what shear
strength to use to compute FOSBH because the method considers constant shear
strength with depth. FOSBH for the example was computed using the strength at the
mid-point of the soft soil below the bottom of the excavation.
Analyses with PLAXIS using the procedures outlined in this paper gave the
results in Table 3. Case 1a is for 42.5 ft deep excavation without preload and Case 1b
is with 100% preload. Cases 2a and 2b
are for the 52.5 ft deep excavation
without and with preload. The preload
reduces movements but the values are
well above the limit values.
Table 3 summarizes calculations of
strut forces and maximum bending
moments using Terzaghi et al (1996) for
both depths. Also included are the
values determined from FEA. Cases 1a,
1b, 2a and 2b are structurally stable but
the displacements are much more than
the allowable values. The sum of the
strut loads computed by PLAXIS
Figure 7: Hypothetical Project
are 3 to 26% higher than computed
by the TPM method. The loads are considerably higher than the TPM conclusion that
actual earth pressures should, on average, be 25% less than loads computed with their
method. The maximum moments computed by PLAXIS are considerably higher than
those by TPM method, which contradicts the TPM conclusion that generally the
actual moments should be less than those computed from the TPM method. These
higher forces and moments result from increasing the wall and strut stiffness to
reduce deflection, something the TPM method does not take into account.
Another interesting result in Table 3 is the FOS information. From calculations
on many different types of facilities, it has been determined that the PLAXIS FOS is
equal to or better than other methods because it finds the most critical failure surface
of any shape and computes a value equal to or slightly lower than that obtained with
limit equilibrium using Spencer’s method. This capability is particularly useful for
design of excavations where strength varies with depth in ways that other solutions
cannot consider.

95
Copyright ASCE 2010 2010 Earth Retention Conference (ER2010)
Earth Retention Conference 3
 96 EARTH RETENTION CONFERENCE
Downloaded from ascelibrary.org by Hong Kong Polytechnic University on 02/11/24. Copyright ASCE. For personal use only; all rights reserved.

Figure 8 illustrates the critical surfaces determined by PLAXIS for a case where it
reaches to the bottom of the soft soil and another case where it passes through the
middle of the soft clay. Note that FOSBH from FEA is similar to that calculated with
the basal heave equation using the average undrained strength below the excavation.
For Case 3, the PLAXIS FOS is considerably more than the basal heave value
because of the added embedment and increased stiffness of the wall that the basal
heave method cannot consider. Also note the differences in failure surface. Case 1b
develops a shallower failure surface than Case 2b. Basal heave assumes failure at the
bottom of the soft soil.
Table 3: Comparison of PLAXIS Results to Results from Peck Diagrams
(forces in kips/ft, moments in kip-ft/ft)
Case Total Max. FOS Pre- Total Max. M FOS Hmax Vmax εHOR αMAX
Strut M Basal load Strut PLAXIS PLAXIS (in) (in) (%)
Load TPM Heave Load
TPM Eq. PLAXIS
1a 108 80.6 1.14 0 125 138 1.10 8.36 4.85 0.97 0.003
1b 108 80.6 1.14 100 136 189 1.10 6.49 3.76 0.82 0.003
2a 189 84.7 0.93 0 195 413 1.03 18.2 10.02 1.5 0.007
2b 189 84.7 0.93 100 214 434 1.03 13.2 7.78 0.95 0.005
3 108 80.6 0.93 100 209 1040 1.77 2.03 0.86 0.10 0.0005
4 189 84.7 N/A 100 220 1670 N/A 0.99 0.54 0.10 0.0003

Several options can be examined with FEA to determine how to reduce

movements. Additional runs were made on the hypothetical case to examine strut
stiffness, strut spacing, wall stiffness, and wall embedment. The final variation that
keeps the movement of a building on the surface close to the allowable values for the
42.5 ft deep excavation is to increase wall stiffness 210 times, strut stiffness by 6 to
13 times, embedment of wall from 2.5 ft to 22.5 and use a 100% preload. The results
are summarized in Table 3 as Case 3. Similarly for the 52.5 ft deep excavation the
wall stiffness was increased 200 times, strut stiffness by 5 to 10 times with 100%
preload and embedment all the way to a firm base at 100 ft depth for Case 4. Note
that for Case 3, the maximum horizontal displacement exceeds the 1 inch limit but
this occurs at the bottom of the stiff, extended wall. Maximum settlement, horizontal
strain and angular distortion are within the allowable limits for the hypothetical case.

Figure 8: Shear Strain Contours at Failure for Cases 1b and 2b

There are some other interesting results in Table 3. Very large struts with high
preloads and a very stiff wall are required to minimize horizontal displacement and
settlement; but, a stable solution is found. Increasing the stiffness of the wall and
struts greatly increases the maximum moment in the wall as shown for Cases 3 and 4.

96
Copyright ASCE 2010 2010 Earth Retention Conference (ER2010)
Earth Retention Conference 3
 EARTH RETENTION CONFERENCE 97
Downloaded from ascelibrary.org by Hong Kong Polytechnic University on 02/11/24. Copyright ASCE. For personal use only; all rights reserved.

It may not be economical to build these solutions but the FEA results show that
constructing the excavation to 52.5 ft and meeting the tight displacement limits is
possible, even with the soft soil conditions at the bottom of the excavation.
With additional parametric studies other alternatives might be found that would
be less expensive to construct. A considerable number of variations were considered
to find the results for Cases 3 and 4. The results computed from PLAXIS are plotted
onto Figure 2 to compare with Clough et al’s 1989 results. There is general
agreement with the values for cases where struts are not preloaded in the FEA.
Preloading in general reduces the predicted movements relative to those one would
calculate from Figure 2. The good agreement in Figure 2 shows that the PLAXIS
calculations give results comparable to what Clough et al (1989) obtained and hence,
comparable to compiled field cases. The value of performing site-specific FEA
rather than relying on the Clough et al (1989) chart is that variable soil profiles,
variable soil properties, site specific conditions and ESS details can be included in the
analysis. When designing for displacement control, these differences may cause
substantial differences in computed wall movement and wall moment. A site-specific
FEA may give a solution with less displacement than determined with the Clough et
al (1989) chart. Additionally as shown in Table 3, the FEA predicts settlement,
horizontal strain and angular rotation outside the wall, which Boscardin and Cording
(1989) showed must be considered to limit damage to buildings close to the wall.
A-site-specific analysis also avoids having to resort to rules of thumb or limited
empirical data to estimate movements. One generality that is commonly used is that
the maximum settlement of the ground behind the excavation is usually between ½ to
1 times the maximum horizontal displacement of the wall. Figure 9 shows values
obtained with PLAXIS for the considerable number of parametric studies used in this
work. Most of the points compare
with the generality, but there are a
number of points where the settlement
is much larger. These points are cases
with large preloaded struts and limited
wall embedment. The horizontal
displacements of the wall were
reduced but plastic flow of soil
beneath the wall allowed settlement of
the ground surface and angular

distortions that would damage many Figure 9: Computed Values of Maximum

buildings. It is not obvious which Horizontal and Vertical Displacements
conditions will violate this generality.
These analyses assume careful control over construction operations to ensure that
the wall is sound in all respects, the lateral supports and connections are installed as
tightly as possible with preloads locked off at the required values and over excavation
does not occur prior to installing the next level of struts. Failure to do any of these
can increase the movements considerably.
One useful outcome of the displacement-based design approach is that a number
of alternatives can be simulated with a consistent approach to determine which have

97
Copyright ASCE 2010 2010 Earth Retention Conference (ER2010)
Earth Retention Conference 3
 98 EARTH RETENTION CONFERENCE
Downloaded from ascelibrary.org by Hong Kong Polytechnic University on 02/11/24. Copyright ASCE. For personal use only; all rights reserved.

the best performance and which may produce problems. This information can be
combined with construction cost estimates to help determine the optimal design.

DISPLACEMENT MONITORING TO REDUCE RISK

Even when a careful evaluation and analysis are done to design an excavation
support system, there remain sources of uncertainty that can result in poor
performance and damage the complete works or adjacent structures. There is also the
unknown effect of quality of the construction workmanship and attention to detail.
Contractors are production oriented people who resist actions that slow them down.
Limiting depth of excavation until lateral supports are installed, preloading supports
with lock off at design levels, reducing strut spacing and controlling ingress of water
and soil particles are steps many contractors resist. These mechanisms can increase
lateral wall movements and ground settlements by factors of two or more. Detection
of unexpected behavior during the excavation process can only be done with
performance monitoring from start to finish. An effective monitoring program must
be a part of the design for any excavation where the consequences of larger than
expected movements are significant. Marr (2007) discusses benefits of performance
monitoring programs that are applicable to deep excavations.
CONCLUSIONS
Design of excavations in urban areas is becoming more complex due to increasing
requirements for deeper excavations on poorer sites, tighter limits on allowable
displacements for adjacent structures, and new methods of construction that extend
beyond the experience base used to develop previous design methods. Increasingly,
limiting movements control ESS design, not avoiding collapse. This requires a new
approach to designing ESS, one that focuses on controlling movements.
The step-by-step design approach laid out in this paper will result in a design that
limits movements, provides adequate factors of safety against soil failure and
provides realistic forces and moments to design the structural members. Each site has
a unique set of conditions that may violate generalities and rules of thumb used in the
past. With tighter restrictions and more demanding situations, site-specific finite
element analyses permit more accurate prediction of horizontal deflection of the wall
and settlement, angular rotation and horizontal strain outside the wall.
The application of the proposed method to a hypothetical project shows that
despite very difficult conditions involving excavation into soft clay, ways can be
found that meet strict displacement limits. These involve stiff struts with 100%
preloads, stiff walls and significant wall embedment. The solutions may or may not
be economically feasible but they show the power of using FEA to provide a
consistent and rational method to design excavation support systems that meet
specific displacement criteria.

REFERENCES

AASHTO (1996) Standard Specifications for Highway Bridges,” 16th Edition. Washington,
D.C. and 2002 17th Edition.

98
Copyright ASCE 2010 2010 Earth Retention Conference (ER2010)
Earth Retention Conference 3
 EARTH RETENTION CONFERENCE 99
Downloaded from ascelibrary.org by Hong Kong Polytechnic University on 02/11/24. Copyright ASCE. For personal use only; all rights reserved.

Bjerrum, L. (1963) “Discussion session IV.” Proc, European Conf. on Soil Mech. and Found.
Engr., Wiesbaden, Germany, II, 135-137.
Boscardin, M.D. and Cording, E.J. (Jan. 1989) “Building Response to Excavation Induced
Settlement.” ASCE:JGE, Vol. 115, No. 1.
Burland, J.B. and Wroth, C.P. (1974) “Settlement of buildings and associated damage.” Proc
Conf. on Settlement of Structures, Pentech Press, London, England.
Clough, G.W., Weber,P.R. and Lamont, J. (1972) “Design and Observation of a Tied-Back
Wall,” Proc. Specialty Conf. on Performance of Earth and Earth Supported Structures,
Purdue University, pp. 1367-1390.
Clough, G.W. and Hansen, L.A.(1981) “Clay Anisotropy and Braced Wall Behavior.”
ASCE:JGED Vol. 107, No. CT7, July 1981, pp. 893-914.
Clough, G.W, Smith, E.M. and Sweeney, B.P.(1989) “Movement Control of Excavation Support
Systems by Iterative Design.” Proc. ASCE, Foundation Engineering: Current Principles and
Practices, Vol. 2, pp. 869-884.
Clough, G.W. and O’Rourke, T.D. (1990) “Construction Induced Movements of In-situ Walls.”
Proc, ASCE Conf. on Des. And Perf. of Earth Retaining Struct., Geotech. Spec.Publ. No.25,
ASCE, New York, 439-470.
Duncan, J.M. and Buchignani, A.L. (1976) “An Engineering Manual for Settlement Studies.”
Department of Civil Engineering, University of California, Berkeley.
Goldberg, D.T., Jaworski, W.E. and Gordon, M.D.(1976) “Lateral Support Systems and
Underpinning.” Report FHWA-RD-75-128, Vol. 1, FHWA, Washington, D.C.
Grant, R., J. T. Christian, and E. H. Vanmarcke (1974), "Differential Settlement of Buildings,"
ASCE:JGED, Vol. 100, No. GT9, pp. 973-991.
Hana, A.I. and Clough, G.W.(1981) “Prediction of Movements of Braced Cuts in Clay.”
ASCE:JGD, Vol. 107, No. GT6, pp. 759-778.
Jaworski, W.E. (1973) “An Evaluation of the Performance of a Braced Excavation,” Ph.D.
Thesis, Massachusetts Institute of Technology.
Henkel, D. (1971) The calculation of earth pressure in open cuts in soft clays,” The Arup Journal
London, 6, No 4. pp. 13-15.
Kulhawy, F. H. and Mayne, P.W. (1990) Manual on Estimating Soil Properties for Foundation
Design. EPRI Report No. EL-6800.
Ladd, C.C. and Foote, R. (1974) “A new design procedure for stability of soft clays.”
ASCE:JGED, Vol. 100, No. GT7. pp. 763-786.
Ladd C.C. (1991) “Stability Evaluation During Staged Construction.” The Twenty-Second Karl
Terzaghi Lecture, ASCE:JGE Vol. 117, No. 4, 540-615.
Lambe, T.W. and Marr, W.A. (1979) "Stress Path Method-Second Edition", ASCE:JGED, Vol. 105,
No. GT6, pp. 727-738.
Marr, W. A (2007) “Why Monitor Performance?” ASCE: GSP 117. Keynote paper presented to 7th
International Symposium on Field Measurements in GeoMechanics, Boston.
Meyerhof, G.G. (1953) “Some recent foundation research and its application to design.” Struct Engr.
Vol. 31, pp. 161-167.
Meyerhof, G.G. (1956) “Discussion paper by Skempton et al – Settlement analysis of six structures in
Chicago and London”, Proc. ICE, 5, No.1, p170.
NAVFAC (1982) Design Manual DM-7, Department of Navy, Washington, D.C.
Peck, R. B. (1969) “Deep Excavations and Tunneling in Soft Ground.” State-of-the-Art-Report,
7th ICSMFE, Mexico City, State-of-the-Art Volume, pp. 225-290.
Polshin, D.E. and Tolkar, R.A. (1957) “Maximum Allowable Nonuniform Settlement of
Structures, Proc. 4th ICSMFE, Vol.I, p. 402.
Poulos H.G. (2000) “Foundation Settlement Analysis – Practice Versus Research.” 8th Spencer J.
Buchanan Lecture, Texas A&M University, College Station, Texas
Sabatini, P.J., Bachus, R.C., Mayne, P.W., Schneider, J.A. and Zettler, T.E. (2002) Geotechnical
Engineering Circular No. 5: Evaluation of Soil and Rock Properties. FHWA-IF-02-034.

99
Copyright ASCE 2010 2010 Earth Retention Conference (ER2010)
Earth Retention Conference 3
 100 EARTH RETENTION CONFERENCE
Downloaded from ascelibrary.org by Hong Kong Polytechnic University on 02/11/24. Copyright ASCE. For personal use only; all rights reserved.

Skempton, A.W. (1951) “The bearing capacity of clays,” Proc. British Bldg, Research Congress,
pp. 180-189.
Skempton, A.W. and MacDonald, D.H. (1956) “The allowable settlement of buildings.” Proc.
Inst. of Civ. Engrs., Part III, 5, 727-784.
Sowers, G.F. (1962) Shallow Foundations, Foundation Engineering, G.A. Leonards (ed.)
McGraw-Hill Book Co., New York, NY, pp. 525-632.
Terzaghi, K. (1943) Theoretical Soil Mechanics, John Wiley & Sons, Inc., New York, NY.
Terzaghi, K. Peck, R.B. and Mesri, G. (1996) Soil Mechanics in Engineering Practice: Third
Edition. John Wiley & Sons, Inc.

100
Copyright ASCE 2010 2010 Earth Retention Conference (ER2010)
Earth Retention Conference 3