Plaxis Advanced Course New Delhi 2014 (Print)

PLAXIS Advanced Course
New Delhi, India.
Venue DepartmentofCivilEngineering
IITDelhi,NewDelhi,India
Date 2931October2014
Lecturers
ProfessorAntonioGens TechnicalUniv.ofCatalonia
ProfessorKRajagopal IITMadras
ProfessorHelmutSchweiger GrazUniv.ofTechnology
Prof.G.V.Ramana IITNewDelhi
DrWilliamCheang PlaxisAsiaPac,Singapore
Plaxis Advanced Course, New Delhi, India (29-31 October 2014) 1/448
CONTENTS
ADVANCEDCOMPUTATIONGEOTECHNICS,NEWDELHI2014 PAGE
Session1:GeotechnicalFiniteElementAnalysis&SoilBehaviour1
CG1 GeotechnicalFiniteElementAnalysis 4
CG2 ElastoplasticityandMohrCoulomb 22
CG3 Exercise1:FoundationonElastoplasticSoils 44
Session2:SoilBehaviour2
CG4 CriticalStateSoilMechanicsandSoftSoilModel 69
CG5 HardeningSoilandHSsmallModels 88
CG6 Exercise2:Triaxial&OedometerSimulation 119
Session3:ModellingofDeepExcavations
CG7 ModellingofDeepExcavations 143
CG8 StructuralElementsinPLAXIS 169
CG9 Exercise3:ModellingofanAnchoredRetainingWallin2D 194
Session4:ModellingofGroundwater,UndrainedConditions&Consolidation
CG10 ModellingofGroundwaterinPLAXIS 216
CG11 DrainedandUndrainedAnalysis 251
CG12 ConsolidationAnalysis 272
CG13 Exercise4:ModellingofanStruttedandEmbeddedExcavationin3D 289
Session5:InitialStresses,SlopeStability&UnsaturatedSoils
CG14 UnsaturatedSoilsandBarcelonaBasicModel 306
CG15 InitialStressesandSlopeStabilityAnalysis 337
CG16 Exercise5:SlopeStabilityAnalysis 375
Session6:ModellingofTunnelsinRock
CG17 HoekBrownandRockJointedModels 392
CG18 ModellingofTunnelsin2D 416
CG19 Exercise6:TunnellinginRock 437
PLAXISADVANCEDCOURSE
Time Wednesday29October2014
Session1:GeotechnicalFiniteElementAnalysis&SoilBehaviour1
9:00 10:00 CG1 GeotechnicalFiniteElementAnalysis Gens
10:00 11:00 CG2 ElastoPlasticityandMohrCoulomb Rajagopal
11:00 11:15 Break
11:15 1:00 CG3 Exercise1:FoundationonElastoPlasticSoils Cheang
1:00 2:00 Lunch
Session2:SoilBehaviour2
2:00 3:00 CG4 CriticalStateSoilMechanicsandSoftSoilModel Schweiger
3:00 4:00 CG5 HardeningSoilandHSsmallModel Schweiger
4:00 4:15 Break
4:15 5:30 CG6 Exercise2:TriaxialandOedometer Cheang
Time Thursday30October2014
Session3:ModellingofDeepExcavations
9:00 9:45 CG7 ModellingofDeepExcavations Gens
9:45 10:30 CG8 StructuralElementsinPlaxis Rajagopal
10:30 10:45 Break
10:45 12:00 CG9 Exercise3:SimulationofanAnchoredRetainingWall Cheang
12:00 1:00 Lunch
Session4:ModellingofGroundwater,UndrainedConditionsandConsolidation
1:00 1:45 CG10 ModellingofGroundwaterinPlaxis Cheang
1:45 2:30 CG11 DrainedandUndrainedAnalysis Gens
2:30 3:15 Break
3:15 3:30 CG12 ConsolidationAnalysis Gens
3:30 5:00 CG13 Exercise4:DewateringinExcavation Siva
Time Friday31October2014
Session5:InitialStresses,SlopeStabilityandUnsaturatedSoils
9:00 10:30 CG14 UnsaturatedSoilsandBarcelonaBasicModel Gens
10:30 11:30 CG15 InitialStressesandSlopeStabilityAnalysis Schweiger
11:30 11:45 Break
11:45 1:30 CG16 Exercise5:SlopeStabilisedbySoilNails Siva
1:30 2:30 Lunch
Session6:ModellingofTunnelsinRock
2:30 3:30 CG17 HoekBrownandRockJointedModels Schweiger
3:30 4:00 CG18 ModellingofTunnelsin2D Schweiger
4:00 4:15 Break
4:15 5:30 CG19 Exercise6:TunnellinginRock Cheang
2931October2014,NewDelhi,India.
New Delhi Advanced 2014
Computation Geotechnics 1
Geotechnical Finite Element Analysis
Professor Antonio Gens
CG1: GEOTECHNICAL FINITE ELEMENT ANALYSIS
Antonio Gens
Technical University of Catalunya, Barcelona
some of the slides were originally created by:

Andrew Abbo (University of Newcastle)
Cino Viggiani (Laboratoire 3S, Grenoble, France)
Dennis Waterman (Plaxis)
Outline
Introduction
Design requirements in geotechnical engineering
Geotechnical methods of analysis
Geotechnical finite element analysis: some remarks
The Finite Element Method

Introduction and general overview
Domain discretization
Element formulation
Constitutive law
Element stiffness matrix
Global equations: assembly and solution
Compute secondary variables
Final remarks
design requirements in geotechnical engineering
Stability (local and general)
Admissible deformation and displacements
design requirements in geotechnical engineering
Flow problems
Sometimes flow and stability/deformation problems are solved together

See tomorrows lecture on consolidation (CG12)
geotechnical analysis: basic solution requirements
Equilibrium (3 equations)
Unknowns: 15
Compatibility (6 equations)
(6 stresses, 6 strains, 3 displacements)
Constitutive equation (6 equations)
Potts & Zdravkovic

(1999)
geotechnical methods of numerical analysis
methods for numerical analysis

Finite difference method
Boundary element method (BEM)
Discrete element method (DEM)
Finite element method (FEM)
Others (meshless methods, material point method, particle methods)
While the FEM has been used in many fields of engineering practice for
over 40 years, it is only recently that it has begun to be widely used for
analyzing geotechnical problems. This is probably because there are many
complex issues which are specific to geotechnical engineering and which
have been resolved relatively recently.
when properly used, this method can produce realistic results which are
of value to practical soil engineering problems
A good analysis, which simulates real behaviour, allows the engineer to

understand problems better. While an important part of the design
process, analysis only provides the engineer with a tool to quantify effects
once material properties and loading conditions have been set
geotechnical finite element analysis
Objectives of the numerical (finite element) analysis

Selection of design alternatives
Quantitative predictions
Backcalculations
Understanding!
Identification of critical mechanisms
Identification of key parameters
Advantages of numerical (finite element) analysis

Simulation of complete construction history
Interaction with water can be considered rigorously
Complex geometries (2D-3D) can be modeled
Structural elements can be introduced
No failure mechanism needs to be postulated (it is an outcome of the
analysis)
(Nearly) unavoidable uncertainties

Ground profile
Initial conditions (initial stresses, pore water pressure)
Boundary conditions (mechanical, hydraulic)
Appropriate model for soil behaviour
Model parameters
Some requirements for successful numerical modelling

Construction of an adequate conceptual model that includes the basic
features of the model. The model should be as simple as possible but
not simpler
Selection of an appropriate constitutive model. It depends on:
type of soil or rock
goal of the analysis
quality and quantity of available information
Pay attention to patterns of behaviour and mechanisms rather than
just to quantitative predictions
Perform sensitivity analyses. Check robustness of solution
Model calibration (using field results) should be a priority, especially of
quantitative predictions are sought
Check against alternative computations if available (even if simplified)
three final remarks

1. geotechnical engineering is complex. It is not because youre
using the FEM that it becomes simpler
2. the quality of a tool is important, yet the quality of a result

also (mainly) depends on the users understanding of both
the problem and the tool
3. the design process involves considerably more than analysis
Borrowed from C. Viggiani, with thanks
The Finite Element Method: introduction and overview
the FEM is a computational procedure that may be used to obtain an

approximate solution to a boundary value problem
the governing mathematical equations are approximated by a series of

algebraic equations involving quantities that are evaluated at discrete points
within the region of interest. The FE equations are formulated and solved in
such a way as to minimize the error in the approximate solution
x xy xz
Governing mathematical equation: bx (equilibrium)
x y z
Algebraic equation: a11x1 a12 x2 a1n xn c1
The FEM is a computational procedure that may be used to obtain an

approximate solution to a boundary value problem
What kind of problem?
Apply load obtain displacements

stiffness matrix
Apply head obtain flow

permeability matrix
Though we would like to know our solution at any coordinates in our

project, we will only calculate them in a certain amount of discrete points
(nodes) and estimate our solution anywhere else
this lecture presents only a basic outline of the method

attention is focused on the first problem using the "displacement based" FE
approach Plaxis Advanced Course, New Delhi, India (29-31 October 2014) 10/448
The FEM involves the following steps (1/2)

Elements discretization
This is the process of modeling the geometry of the problem under
investigation by an assemblage of small regions, termed finite
elements. These elements have nodes defined on the element
boundaries, or within the elements
Primary variable approximation

A primary variable must be selected (e.g., displacements) and rules
as how it should vary over a finite element established. This
variation is expressed in terms of nodal values
A polynomial form is assumed, where the order of the polynomial

depends on the number of nodes in the element
The higher the number of nodes (the order of the polynomial), the
more accurate are the results (the longer takes the computation!)
The FEM involves the following steps (2/2)

Element equations
Derive element equations:
where is the element stiffness matrix, is the vector of nodal

displacements and is the vector of nodal forces
Global equations
Combine (assemble) element equations to form global equations
Boundary conditions
Formulate boundary conditions and modify global equations. Loads
affect P, while displacements affect U
Solve the global equations
to obtain the displacements at the nodes
Compute additional (secondary) variables
From nodal displacements secondary quantities (stresses, strain) are
evaluated

stiffness matrix
Ku F
For soil we dont have a direct relation between load and displacement,
we have a relation between stress and strain.
Displacements Strains Stresses Loads

Differentiate Material Integrate
model
Bu D F dV
Combine these steps: K BT DBdV
The first stage in any FE analysis Footing

width = B
is to generate a FE mesh
A mesh consists of elements

connected together at nodes
Node
We will calculate our solution in
the nodes, and use some sort of
mathematical equation to
estimate the solution inside the Gauss point
elements.
examples: embankment
examples: multi-anchored diaphragm wall
There is a whole zoo of different finite elements available!

Element formulation
displacement interpolation
two-dimensional analysis of continua is generally based on the use of

either triangular or quadrilateral elements
the most used elements are based on an iso-parametric approach
Element formulation
Displacement interpolation
primary unknowns: values of the nodal displacements
displacement within the element: expressed in terms of the nodal values
using polynomial interpolation
n
u( ) Ni ( ) ui , N i shapefunction of node i
i 1
Element formulation
Shape function of node i

Is a function that has value 1 in node i
and value 0 in all other n-1 nodes of the element
Shape functions for 3-node line element

1 1
N1 (1 ) , N 2 (1 )(1 ) , N 3 (1 )
2 2
Shape functions for 5-node line element
Element formulation
Illustration for the six-noded triangular element
3 y quadratic interpolation
6
v
5
x u( x, y ) a0 a1 x a2 y a3 x 2 a4 xy a5 y 2
u v( x, y ) b0 b1 x b2 y b3 x 2 b4 xy b5 y 2
1 2
4
12 coefficients, depending on the values of the

12 nodal displacements
u NU e
Element formulation
Illustration for the six-noded triangular element

Strains may be derived within the element using the standard definitions
u
xx a1 2a3 x a4 y
x
Lu
v
yy b2 b4 x 2b5 y
y
u v
xy (b1 a2 ) (a4 2b3 ) x (2a5 b4 ) y
y x
Lu LNU e BU e BU e
Constitutive law
Constitutive relation (elasticity)

Elasticity: one-to-one relationship between stress and strain
in a FE context, stresses and strains are written in vector form
the stress-strain relationship is then expressed as: = D
linear isotropic elasticity in plane material stiffness matrix

strain

1 v v 0
v 1 v 0
E
D
(1 2v)(1 v) 1 2v
0 0
2
in this case the coefficients of the matrix are constants, which means
that (for linear kinematics) the resulting F.E. equations are linear
Constitutive law
What happens with inelastic constitutive relations?

Advantage with elasticity: the coefficients of the matrix are constants,
the resulting F.E. equations are linear, hence the problem may be solved
by applying all of the external loads in a single calculation step
soils usually do not behave elastically
D
with D depending on the current and past stress history
It is necessary to apply the external load in separate increments

and to adopt a suitable non-linear solution scheme

body forces and surface tractions applied to the element may be
generalized into a set of forces acting at the nodes (vector of nodal forces)
nodal forces may be related P1x

to the nodal displacements by: 3 1y
P

P2 x
K e Ue Pe 6 5 P2 y
P e
Ke element stiffness matrix P1x 1
2

4

K e T
B DBdv P1y
P6 x
P
6y
D material stiffness matrix
recall
B matrix relating nodal displacements to strains

stiffness matrix
Ku F
For soil we dont have a direct relation between load and displacement,
we have a relation between stress and strain.
Displacements Strains Stresses Loads

Differentiate Material Integrate
model
Bu D F dV
Combine these steps: K BT DBdV
Gauss points
Ke
B T DBdv
To evaluate Ke, integration must be performed for each element
A numerical integration scheme must be employed (Gaussian integration)
Essentially, the integral of a function is replaced by a weighted sum

of the function evaluated at a number of integration points
Global stiffness matrix (1)

The stiffness matrix for the complete mesh is evaluated by combining
the individual element stiffness matrixes (assembly)
This produces a square matrix K of dimension equal to the number of

degrees-of-freedom in the mesh
in 2D number of d.o.f = 2 x number of nodes
in 3D number of d.o.f = 3 x number of nodes
The global vector of nodal forces P is obtained in a similar way by

assembling the element nodal force vectors
The assembled stiffness matrix and force vector are related by:
KU P
where vector U contains the displacements at all the nodes in the mesh
Global stiffness matrix (2)

if D is symmetric (elasticity), then Ke and hence K will be symmetric
The global stiffness matrix generally contains many terms that are zero
if the node numbering scheme is efficient then all of the non-zero
terms are clustered in a band along the leading diagonal
assembly
schemes for storage
solution
take into account its sym and

banded structure
number of dofs
Solution of the global stiffness equations

Once the global stiffness equations have been established
(and the boundary conditions added), they mathematically
form a large system of symultaneous (algebraic) equations
KU P
These have to be solved to give values for the nodal displacements
It is advantageous to adopt special techniques to reduce

computation time (e.g. bandwidth and frontal techniques)
Detailed discussion of such techniques is beyond the scope of

this lecture
Computation of secondary variables
Compute additional (secondary) values

once the nodal displacements have been obtained from the inversion
of the matrix K e
KU P
The complete displacement field can be obtained:
n
u ( x, y) Ni ( x, y) ui , Ni shape function of node i
i 1
Strains and stresses are computed at the Gauss points:
BU e
= D
Final remarks
some practical issues
1. A good finite element mesh is important. A poor mesh will give

a poor (inaccurate) solution.
2. Post processing Stress are computed at Guass points only.

Contour plots of stresses involve further processing of the results.
3. Do the results make sense?
4. FEA can be very time consuming!
Elasto-plasticity and Mohr-Coulomb
Professor K.Rajagopal
CG2: ELASTO-PLASTICTY AND MOHR COULOMB
Professor K Rajagopal
IIT Madras

Antonio Gens (UPC, Spain)
S.W. Lee (GCG Asia Golder Associates)
Helmut Schweiger (Technical University of Graz, Austria)
K. Rajagopal (IIT Madras) additional slides September 2013
Contents
A quick reminder of (linear isotropic) Elasticity

Motivations for plasticity (elasticity vs. plasticity)
Basic ingredients of any elastoplastic model
elastic properties (how much recoverable deformation?)
yield surface (is plastic deformation occurring?)
plastic potential (direction of plastic strain increment?)
consistency condition (magnitude of plastic strain increment?)
hardening rule (changes of yield surface?)
Element tests: (drained) simple shear & triaxial tests
Tips and tricks
Advantages and limitations
Constitutive models
Constitutive models provide us with a relationship with stresses and

strains expressed as:
= D
Elasticity
Linear-elastic Non-linear elastic

= D
Hookes law
C
xx 1 0 0 0 xx
1 0 0 0
yy yy
zz 1 1 0 0 0 zz

E 0 0 0 2 2 0 0
xy xy
yz 0 0 0 0 2 2 0 yz

zx 0 0 0 0 0 2 2 zx
Model parameters in Hookes law:
d1
Two parameters:
- d1
- Youngs modulus E
- Poissons ratio
d3
- 1
Meaning (axial compr.):

E
d1
E 1
d1
- 1
d
3 1
d1
E 0 ; -1 0 .5 3
Alternative parameters in Hookes law:

In spherical and deviatoric stress / strain components:
v 1/ K 0 p
p 1
3 1 2 3

0 1/ 3G q
s 1
q ( 1 2 ) 2 ( 2 3 ) 2 ( 3 1 ) 2
2
dxy
Shear modulus:
d xy E dxy
G
d xy 21
Bulk modulus: dp
dp E
K dv
d v 31 2
9KG 3 K 2G
E v
G 3K 6 K 2G
Hookes law
= D
Inverse:
xx 1 0 0 0
xx
1 0 0
0
yy yy
zz E 1 0 0
0 zz

(1 )(1 2 ) 0
0 0 1 0 0
xy 2 xy
yz 0 0 0 0 1
2

0 yz
1
zx 0 0 0 0 0 2 zx
4G 2G 2G
K K K 0 0 0
xx 3 3 3 xx

K 2G 4G 2G
yy K K 0 0 0 yy
3 3 3
zz zz
K 2G 2G 4G
K K 0 0 0
xy 3 3 3 xy
yz 0 0 0 G 0 0 yz

zx 0 0 0 0 G 0 zx
0 0 0 0 0 G

Hookes law
Plane strain = D
4G 2G
K K 0
1 0 3 3

2G 4G
0 K 0
E
D 1 K
(1 )(1 2 ) 3 3
1 2
0 0 0 0 G
2

Axisymmetry
4G 2G 2G
K K K 0
1 0 3 3 3

1 0 2G 4G 2G
E K K K 0
D 1 0 3 3 3
(1 )(1 2 )
0 1 2 K 2 G K
2G
K
4G
0
0 0 3 3 3
2
October
0 2014) 0 026/448 G

Plaxis Advanced Course, New Delhi, India (29-31
Elasticity vs. Plasticity (1)
In elasticity, there is a one-to-one relationship between stress and

strain. Such a relationship may be linear or non-linear. An essential
feature is that the application and removal of a stress leaves the
material in its original condition
Elasticity vs. Plasticity (2)
for elastic materials, the mechanism of deformation depends on the

stress increment
for plastic materials which are yielding, the mechanism of (plastic)

deformation depends on the stress
reversible = elastic irreversible = plastic
Plasticity: some definitions (1)
LINEAR ELASTIC - PERFECTLY PLASTIC

One-dimensional
Y0 = yield stress
IMPORTANT: yield stress = failure stress for perfect plasticity
e p
General three-dimensional stress state e p
LINEAR ELASTIC PLASTIC HARDENING

One-dimensional
Y0 = yield stress
YF = failure stress
IMPORTANT: yield stress failure stress
e p
e p
(29-31
Plaxis Advanced Course, New Delhi, India
General three-dimensional stress state 2014)
October 28/448
LINEAR ELASTIC - PLASTIC WITH SOFTENING
One-dimensional
Y0 = yield stress
YF = failure stress
yield function (1)
when building up an elastic-plastic model,

the first ingredient that we need is a yield surface
(is plastic deformation occurring?)
yield function (2)
F = 0 represents surface in stress space
f f 1 , 2 , 3
f 0 stress state is elastic
f 0 stress state is plastic
f 0 stress state not admissible
The yield surface bounds all elastically attainable states

(a generalized preconsolidation pressure)
yield function (5)
Basically:
changes of stress which remain inside the yield surface are
associated with stiff response and recoverable deformations,
whereas on the yield surface a less stiff response is obtained and
irrecoverable deformations are developed
Where do we get this function f ?
The dominant effect leading to irrecoverable changes in particle

arrangement is the stress ratio, or mobilized friction
The mean normal effective stress p is of primary importance.

The range of values of q for stiff elastic response is markedly
dependent on p
Tresca & Von Mises yield functions are not appropriate
Mohr-Coulomb Model, yield function
To most engineers the phrase strength of soils conjures

up images of Mohr-Coulomb failure criteria
frictional
resistance
independent of
normal stress
Classical notions of Mohr-Coulomb failure can be

reconciled with the patterns of response that we are
modeling here as elasto-plastic behavior
1 and 3 : major and minor principal stresses
The Mohr-Coulomb failure criterion
MC criterion: t*c cos - s* sin
t* = (3 - 1)
s* = (3+1)
1
2 '3 '1 c' cos ' 12 '3 '1 sin '
2c' cos ' 1 sin '

'1 '3
1 sin ' 1 sin '
Note: Compression is negative, and 1: major,

2: intermediate, 3: minor principal stress
19
MOHR COULOMB IN 3D STRESS SPACE
f 1
2
'1 '3 12 '1 '3 sin ' c ' cos '
-1
f > 0 Not acceptable

f = 0 Plasticity
f < 0 Elasticity
-3
-2
plastic potential (1)
Summing up:
Plastic strain increment arises if:

1) the stress state is located on the yield surface (f = 0)
AND
2) the stress state remains on the yield surface after a stress increment
knowledge of function f tells us whether plastic strain is occurring or not
But, this is only one part of the story:

We would also like to know direction and magnitude of plastic strain
will we get plastic volume changes?
and plastic distortion?
for that, we need another concept (another function: g)
plastic potential (2)
flow rule
Recall: plastic deformations depend
on the stress state at which yielding
is occurring, rather than on the route
by which that stress is reached
we have now two functions, f and g

the question is: where do we get g ?
associated and non associated flow rules
it would be clearly a great advantage if, for a given material, yield locus
and plastic potential could be assumed to be the same
f = g only 1 function has to be generated to describe plastic response
also advantageous for FE computations:
the solution of the equations that emerge in the analyses is faster
the validity of the numerical predictions can be more easily guaranteed
is f = g a reasonable assumption?
for metals, it turns out that YES, it is

for geomaterials, NOT
Where is the problem? The assumption of normality of plastic strain
vectors to the yield locus would result in much greater plastic volumetric
dilation than actually observed
Mohr-Coulomb Model, plastic potential
dilatancy angle
plastic dilatancy
how to understand dilatancy

i.e., why do we get volume changes when applying shear stresses?
= + i
the apparent externally mobilized angle of friction on horizontal planes () is
larger than the angle of friction resisting sliding on the inclined planes (i)
strength = friction + dilatancy
consistency condition
Parameters of MC model
E Youngs modulus [kN/m2]

Poissons ratio [-]
c (effective) cohesion
[kN/m2]
(effective) friction angle []
Dilatancy angle []
MC model for element tests
yy
tan
xy
max

sin cos

n
1 sin sin
limitations of MC model (1)
limitations of MC model (2)
warning for dense sands
Tension cut-off
Tension cut-off: if c>o, MC model allows tensile stresses to be

developed
Tension Compression
zone zone
Tension cut-off
Simple Shear Test
Given: E' , ' , c' , ' , , xy , yy o
Initial conditions: xy 0 0, ' xx o K o ' yy o

'1
'3
Boundary conditions: yy 0, xx 0

45
2
Implicit stress increment : '3
'1
'new 'old
At failure for
D ep simple shear test
With Dep evaluated at the beginning of each step (using old)
Dilatancy
Plastic volumetric
xx yy
p p p
strain v
sin
Pla stic shear distortion
p xx
p
yy p 2
xy
p 2
For simple shear

Since, xx xx xx 0
e p
yy
yy
p
and xx 0
e
sin
yy xx 0
p 2 2 p
p p
yy xy
yy
p
xy p tan
xy
p
is positive volume increase in shear - Dilatancy

is negative volume decrease in shear - Contractancy
Results of undrained simple shear test
Non-associated plasticity, f g
Parameter Value
Bulk unit weight of soil, 0 kN/m3
Effective angle of internal friction, 43o
Effective cohesion, c 0 kPa
Poissons ratio, 0.20
Youngs modulus, E 45000 kN/m2
Angle of dilation , -3 o ,0o ,15 o
Bulkmodulusofsoil,Ke =Kw/n 1.86*106 kN/m2
Bulk unit weight of gravity elements, 20 kN/m3
Normal stress in y-direction, yy 100kN/m2
Shear strain, xy 0.025
Ko 0.25
Comparison between drained and undrained results

Property Value
Bulk unit weight of soil, 0 kN/m3
Effective angle of internal
35o
friction,
Effective cohesion, c 0 kPa
Poissons ratio, 0.30
Youngs modulus, E 26000 kPa
Angle of dilation , -3o, 0o, 5o

Bulkmodulusofsoil,Ke = 1.85*106
Kw/n kN/m2
Bulk unit weight of gravity
20 kN/m3
elements,
Normal stress in y-
100kN/m2
direction, yy
Shear strain, xy 0.025
Influence of dilation angle on behaviour of circular footing
Pressure (kPa)
0 500 1000 1500 2000 2500 3000 3500
0
-0.05
-0.1
-0.15
settlement (m)
-0.2
psi =0
psi=3
-0.25 psi=6
psi=10
-0.3
-0.35
-0.4
E = 35000 kPa
= 0.3
-0.45
c = 1 kPa
=35
Diameter=4m
Possibilities and limitations of the Linear Elastic- Perfectly

Plastic (LEPP) Mohr-Coulomb model
Possibilities and advantages
Simple and clear model

First order approach of soil behaviour in
general 1
Suitable for a good number of practical
applications (not for deep excavations and
tunnels)
Limited number and clear parameters
Good representation of failure behaviour 3
2
(drained)
DilatancyPlaxis
can be included
Advanced Course, New Delhi, India (29-31 October 2014) 42/448
40
Possibilities and limitations of the Linear Elastic- Perfectly Plastic (LEPP) Mohr-
Coulomb model
Limitations and disadvantages
Isotropic and homogeneous behaviour

Until failure linear elastic behaviour
No stress/stress-path/strain-dependent 1
stiffness
No distinction between primary loading and
unloading or reloading
Dilatancy continues for ever (no critical state)
3
Be careful with undrained behaviour 2
No time-dependency (creep)
41
Exercise 1: Shallow Foundation on Elasto-plastic Soil
Dr William Cheang
Elastoplastic analysis of a footing
ELASTOPLASTIC ANALYSIS OF A
FOOTING
Computational Geotechnics 1
INTRODUCTION
One of the simplest forms of a foundation is the shallow foundation. In this exercise we will
model such a shallow foundation with a width of 2 meters and a length that is sufficiently long
in order to assume the model to be a plane strain model. The foundation is put on top of a 4m
thick clay layer. The clay layer has a saturated weight of 18 kN/m3 and an angle of internal
friction of 200 .
Figure 1: Geometry of the shallow foundation.
The foundation carries a small building that is being modelled with a vertical point force.
Additionally a horizontal point force is introduced in order to simulate any horizontal loads
acting on the building, for instance wind loads. Taking into account that in future additional
floors may be added to the building the maximum vertical load (failure load) is assessed. For
the determination of the failure load of a strip footing analytical solutions are available from for
instance Vesic, Brinch Hansen and Meyerhof:
Qf
B
= c Nc + 12 0 B N
0
Nq = e tan tan2 (45 + 12 0 )
0
q 1) cot
Nc = (N
0
2(Nq + 1) tan
(V esic)
N = 1.5(Nq 1) tan 0 (Brinch Hansen)

(Nq 1) tan(1.4 0 ) (M eyerhof )

This leads to a failure load of 117 kN/ m2 (Vesic), 98 kN/m2 (Brinch Hansen) or 97 kN/m2
(Meyerhof) respectively.
2 Computational Geotechnics
SCHEME OF OPERATIONS
This exercise illustrates the basic idea of a finite element deformation analysis. In order to
keep the problem as simple as possible, only elastic perfectly-plastic behaviour is considered.
Besides the procedure to generate the finite element mesh, attention is paid to the input of
boundary conditions, material properties, the actual calculation and inspection of some output
results.
Aims
Input
Start new project

Soil mode
* Create soil layers

* Create and assign soil material sets
Structures mode
* Create footing
* Create load
Mesh mode
* Generate mesh
Staged construction mode
* Determine initial situation

* Calculation of vertical load representing the building weight
* Calculation of vertical and horizontal load representing building weight and wind
force
* Calculation of vertical failure load.
Output
Inspect deformations
Inspect failure mechanism
Inspect load-displacement curve
INPUT
Introduction
Start PLAXIS 2D by double-clicking the icon of the PLAXIS 2D Input program. The Quick
select dialog box will appear in which you can select to start an new project or open an
existing one. Choose Start a new project (see Figure 2). Now the Project properties window
appears, consisting of the two tabsheets Project and Model (see Figure 3 and Figure 4).
Figure 2: Quick select dialog
Project properties
The first step in every analysis is to set the basic parameters of the finite element model.
This is done in the Project properties window. These settings include the description of the
problem, the type of analysis, the basic type of elements, the basic units and the size of the
drawing area.
Project tabsheet
In the Project tabsheet, enter Exercise 1 in the Title box and type Elasto-plastic analysis of
a drained footing or any other text in the Comments box.
Model tabsheet
In the Model tabsheet several model specific parameters can be specified
In the Type box the type of the analysis (Model) and the basic element type (Elements)
are specified. As this exercise concerns a strip footing, choose Plane strain from the
Model combo box. Select 15-node from the Elements combo box.
Figure 3: Project tabsheet of the Project Properties window
Figure 4: Model tabsheet of the Project properties window
The Units box defines the units for length, force and time that have to be used in this
project. There is a choice for several units, both metric and emperial.
For this project use the default units (Length = m; Force = kN; Time = day).
In the Contour box the size of the considered geometry must be entered. The values
entered here determine the size of subsoil input window. PLAXIS will automatically add
a small margin so that the geometry will fit well within the draw area. Enter xmin =0.00,
xmax =14.00, ymin =0.00 and ymax =4.25, see figure 4
Click on the OK button below the tabsheets to close the Project properties window.
Hint: In the case of a mistake or for any other reason that the project properties
should be changed, you can access the Project properties window by
selecting the Project properties option from the File menu.
Soil Mode
The program is now in Soil mode in which the subsoil should be created. As shown in figure
1 the subsoil consist of a single 4m thick clay layer and creating this layer is done in 2 steps:
first the soil layer is defined through the definition of a borehole, after which the material set
representing the clay is defined and assigned to the appropriate layer.
Create soil layer

Select the button Create borehole ( ) and click in the drawing area on the origin to
indicate a borehole should be created there. The Modify soil layers window opens, see
figure 5. Intially this window is empty as no boreholes have been defined yet for this
project.
Figure 5: The initial Modify soil layers window
Now click the Add button in order to add a layer to the borehole.
On the Soil layers tabsheet the different soil layers present in the borehole must be
defined. In this exercise there is only 1 soil layer with the Top at 4.0m and the Bottom at
0.0m, see figure 6.
On the left side of the Modify soil layers window there is a graphical representation of
the borehole. Note that the soil layer does not have a soil material assigned yet.
Above the borehole the Head option specifies the position of the global water level in
this borehole. In this exercise it is assumed that the phreatic level is at groundlevel, so
the Head must be set equal to 4.0 to indicate that the phreatic level is at ground level.
Now press the <OK> button to close the Modify soil layers window. The drawing area
now shows a grey rectangular subsoil.
Figure 6: The Modify soil layers window with 1 borehole containing 1 soil layer
Create and assign material sets

In this exercise 2 material sets will be used: one material set for the clay layer, and the second
material set will be used to model the concrete footing. To create the material sets, follow
these steps:
Select the Materials button ( ) - the Material sets window will open. The list of
material sets available for this project is still empty.
Click on the New button at the lower side of the Material Sets window. A new dialog box
will appear with five tabsheets: General, Parameters, Flow parameters, Interfaces and
Initial (see figure 7).
In the Material Set box of the General tabsheet, write Clay in the Identification box.
Select Mohr-Coulomb from the Material model combo box and Drained from the Material
type combo box.
Enter the proper values for the weights in the General properties box according to the
material properties listed in table 1
Click on either the Next button or click on the Parameters tabsheet to proceed with
the input of model parameters. The parameters appearing on the Parameters tabsheet
depend on the selected material model (in this case the Mohr-Coulomb model).
Enter the model parameters of table 1 in the corresponding edit boxes of the Parameters
tabsheet. The parameters in the Alternatives and Velocities group are automatically
calculated from the parameters entered earlier.
See also figure 8. In this figure the Advanced parameters part has been collapsed.
Figure 7: General tabsheet of the soil and interface data set window for Clay
Since the geometry model does not include groundwater flow or interfaces, the third and
fourth tabsheet can be skipped. Click on the OK button to confirm the input of the current
material data set.
Now the created data set will appear in the tree view of the Material Sets window.
Table 1: Material properties of the clay layer and the concrete footing.
Parameter Symbol Clay Concrete Unit
Material model Model Mohr-Coulomb Linear elastic
Type of behaviour Type Drained Non-porous
Weight above phreatic level unsat 16.0 24.0 kN/m3
Weight below phreatic level sat 18.0 kN/m3
Youngs modulus E0 5.0103 2.0107 kN/m2
Poissons ratio 0 0.35 0.15
0
Cohesion cref 5.0 kN/m2
Friction angle 0 20
Dilatancy angle 0
For the concrete of the footing repeat the procedure, but choose a Linear Elastic material
behaviour and enter the properties for concrete as shown in table 1 (see also figures 9 and
10).
Figure 8: Parameters tabsheet of the soil and interface data set window for Clay
Figure 9: General tabsheet of the soil and interface data set window for Concrete
Figure 10: Parameters tabsheet of the soil and interface data set window for Concrete
Now from the Material sets window drag the Clay material set with the mouse over the
grey subsoil and drop it. The subsoil should now get the colour of the material set, see
figure 11.
Figure 11: Subsoil before (left) and after (right) assigning the Clay material set
This ends the creation of the subsoil in Soil mode. By clicking on the Structures tabsheet now
move to Structures mode.
Structures mode
Introduction
In Structures mode the footing as well as the point load acting on the footing will be created.
However, first an adjustment to the snapping interval must be made in order to be able to draw
the 0.25m thick footing. By default, the snapping interval is set to 1m.
From the vertical toolbar select the Snapping options button ( ). The Snapping window
now opens.
Make sure the options Enable snapping and Show grid are selected
Leave the Spacing to 1 m
Set the Number of snap intervals to 4. This means that every spacing of 1 meter is
divided in 4, hence the snapping distance will be 0.25m.
Click the <OK> button to confirm the new settings and close the window.
Create footing
1. Select the Create soil button ( ) and from the drop-down list that opens now select
the Create soil rectangle button ( ).
2. Move the mouse cursor to the coordinates (x y) = (6 4) and single-click the left mouse
button
3. Now move the mouse cursor to the coordinates (x y) = (8 4.25) and single-click the left
mouse button again. We have now created the footing.
4. Select the Show materials button ( ), the Material sets window will open.
5. Drag-and-drop the Concrete material set onto the footing.
Create load
1. Select the Create load button ( ) and from the drop-down list that opens select the
Create point load option ( ).
2. Move the mouse cursor to the coordinates (x y) = (7 4.25) and single-click the left mouse
button to insert the point load.
This concludes the creation of the footing and loads. By clicking on the Mesh tabsheet now
move to Mesh mode.
Mesh mode
In Mesh mode the user can specify necessary mesh refinements and generate the mesh. In
this exercises no additional mesh refinement will be used.
Select the Generate mesh button ( ). The Mesh options window will open.
Leave the Element distribution to Medium and press <OK> to start mesh generation
If mesh generation finished succesfully this will be confirmed in the Command explorer
with the message "Generated XX elements, YY nodes" where XX and YY stand for the
amount of elements and nodes respectively.
Select the View mesh button in order to view the generated mesh, see figure 12.
Figure 12: Generated mesh
Close the mesh window by selecting the green <Close> button. This ends the Mesh mode.
As no water levels will be used in this exercise, the Water levels mode can be skipped and we
can move directly to Staged construction mode to define the calculation phases.

In Staged construction mode all calculation phases will be defined. In this exercise we will use
5 calculation phases, which includes the initial phase.
Initial phase
The initial phase represents the field conditions that exist at the moment our project starts.
This means that only the subsoil exists in the initial conditions whereas the footing should be
deactivated.
Right-click on the footing. The footing will become red (indicating it is selected) and a
drop-down menu appears.
From the drop-down menu select the option Deactivate in order to deactivate the footing.
Phase 1: Construction of the footing
In the Phase explorer select the Add phase button ( ) so that a new phase will be
added.
Right-click on the footing and from the drop-down menu that appears select the option
Activate to activate the footing, see figure 13.
Figure 13: Geometry configuration for the initial phase (left) and phase 1 (right)
Phase 2: Apply vertical load
added.
Click on the point on which the load acts so that it becomes red. On the left side the data
of the load now appears in the Selection explorer.
Activate the point load and set the value of the vertical component, Fy,ref = -50 kN (=
downwards), see figure 14.
Phase 3: Add horizontal load

added.
of the load again appears in the Selection explorer.
Set the value of the vertical component of the point load, Fx,ref = 20 kN, see figure 14.
Figure 14: Activating and changing the point load through the Selection explorer in phase 2
(left) and phase 3 (right)
Phase 4: Vertical failure load

In this phase we will calculate the vertical failure load as if no horizontal load has been applied.
This means that phase 4 must be a continuation of applying the vertical load in phase 2.
In the Phase explorer select phase 2 so that it will show in bold letter type
Now select the Add phase button ( ) so that a new phase will be added that follows on
phase 2 rather than on phase 3.
of the load again appears in the Selection explorer.
Set the value of the vertical component of the point load, Fy,ref = -500 kN. Note that
Fx,ref should remain 0 (zero).
This finishes the definition of the calculation phases for this project.
Calculation
Load-displacement curves
As a calculation result we would like to draw a load-settlement curve for the footing. In order
to do so, the user must select one or more points for which Plaxis has to gather data during
the calculation:
Select the Select points for curves button ( ). The output program now opens, showing
the mesh with all nodes.
Select the node in the middle underneath the footing, hence at or very close to (x y) = (7
4). The node will appear in the Select points list, see figure 15.
Close Plaxis Output by clicking the green <Update> button.
Figure 15: Selecting points for node displacement curves
Calculate
Press the Calculate button ( ) to start the calculation.

Note that the last calculation phase fails: the intended vertical load of 500 kN cannot be fully
applied due to failure of the subsoil underneath the footing.
RESULTS
Output
After the calculation finishes, click the View calculation results button ( ). Plaxis Output
will open, showing the calculation results of the last calculation phase.
By default Plaxis Output will show the Defomed mesh, see figure 16. If this is not the case the
Deformed mesh can be shown by choosing the menu Deformations Deformed mesh |u|.
Figure 16: Deformed mesh after phase 4
Now choose the menu option Deformations Incremental displacements|u|, see figure
17.
The incremental displacements is the change in displacements in the current calculation
step (here that is the last calculation step of the phase 4). Under working conditions the
change of displacement per calculation step is quite small, but in case of failure, the change of
displacements can be large inside the failure zone. Therefore the Incremental displacements
graph can be very suitable for detecting whether failure occurs and what the failure zone may
look like. Figure 17 shows the typical Prandtl-like failure zone.
Figure 17: Incremental displacements for the final calculation step of phase 4
Finally, we will inspect the load-settlement curve and determine the failure load. To do so,
follow these steps:
From the button bar select the Curves manager button ( ). The Curves manager will
open.
In the Curves manager select the <New> button in order to generate a new curve. Now
the Curve generation window opens.
In the Curve generation window, select for the x-axis data from point A (instead of Project
data) from the drop down list.
Now in the tree below, select Deformations Total displacements |u|
For the y-axis we will plot a Project value, and that is the Multiplier M stage.
Press <OK>. A curve as can be seen in figure 18 will show.
Figure 18: Load-settlement curve
In a Plaxis calculation any change made in a construction phase leads to a so-called unbalance,
that is a disturbance between the total of the internal stresses and the external load. This
unbalance is gradually solved using the M stage multiplier. The M stage multiplier indicates
how much of the unbalance has been solved, where M stage = 0 indicates that no unbalance
was solved and M stage = 1 that the full unbalance has been solved.
In the curve shown in figure 18 the lines at the left indicate the variation of M stage for the first
3 calculation phases, where as the long curved line shows the variation of M stage during
the final phase.
It shows that at failure occurs when M stage = 0.38, hence 38% of the unbalance was solved.
In this case the unbalance applied was the increase of the vertical load from 50 kN/m to 500
kN/m. Hence, at failure the total load applied is the load at the beginning of the phase (50
kN/m) plus 38% of the change of load that could be applied: Fmax = 50 + 0.38 (500 50) = 221
kPa
The exact value of the M stage multiplier can be inspected by moving the mouse cursor over
the plotted line. A tooltip box will show up with the data of the current location.
Comparison
In addition to the mesh used in this exercise calculations were performed using a very coarse
mesh with a local refinement at the bottom of the footing and a very fine mesh. Fine meshes
will normally give more accurate results than coarse meshes. Instead of refining the whole
mesh, it is generally better to refine the most important parts of the mesh, in order to reduce
computing time. Here we see that the differences are small (when considering 15-noded
elements), which means that we are close to the exact solution. The accuracy of the 15-
noded element is superior to the 6-noded element, especially for the calculation of failure
loads.
Hint: In plane strain calculations, but even more significant in axi-symmetric
calculations, for failure loads, the use of 15-noded elements is recommended.
The 6-noded elements are known to overestimate the failure load, but are ok
for deformations at serviceability states.
Table 2: Results for the maximum load reached on a strip footing on the drained sub-soil for
different 2D meshes
Mesh size Element Nr. of Max. Failure

type elements load load
[kN/m] [kN/m2 ]
Medium mesh 15-noded 212 221 117
Very coarse mesh 6-noded 84 281 147
Medium mesh 6-noded 212 246 129
Very fine mesh 6-noded 626 245 129
Very coarse mesh 15-noded 84 224 118
Very fine mesh 15-noded 626 221 117
Analytical solutions of:
- Vesic 117
- Brinch Hansen 98
- Meyerhof 97
In this table the failure load has been calculated as:
Qu M aximum f orce M aximum f orce

B
= B
+ concrete d = 2
+6
From the above results it is clear that fine FE meshes give more accurate results. On the other
hand the performance of the 15-noded elements is superior over the performance of the lower
order 6-noded elements. Needless to say that computation times are also influenced by the
number and type of elements.
ADDITIONAL EXERCISE:
UNDRAINED FOOTING
INTRODUCTION
When saturated soils are loaded rapidly, the soil body will behave in an undrained manner, i.e.
excess pore pressures are being generated. In this exercise the special PLAXIS feature for
the treatment of undrained soils is demonstrated.
SCHEME OF OPERATIONS
In PLAXIS, one generally enters effective soil properties and this is retained in an undrained
analysis. In order to make the behaviour undrained one has to select undrained A as the type
of drainage. Please note that this is a special PLAXIS option as most other FE-codes require
the input of undrained parameters e.g. Eu and u .
Aims
The understanding and application of undrained soil behaviour
How to deal with excess pore pressures.
Use previous input file and ave as new data file
Soil mode
Change material properties, undrained behaviour for clay
Mesh mode
Mesh generation, global mesh refinement B)
Re-run existing calculation phases
Output
Inspect excess pore pressures
Soil mode
INPUT
Use previous input file
If PLAXIS Input is no longer open, start PLAXIS by clicking on the icon of the Input program
and select the existing project file from the last exercise (drained footing). From the File menu
select Save As and save the existing project under a new file name (e.g. exercise 1b).
Change material properties
Change material properties by selecting the Show materials button ( ). Please note
that this button is only available in Soil mode, Structures mode and Staged construction
mode.
From the Material sets window, select the Clay and click on the <Edit> button.
In the Soil window that opened on the first tab sheet (General) change the Drainage
type to "Undrained A" and close the data set.
Mesh generation
The mesh generator in PLAXIS allows for several degrees of refinement. In this example
we will globally refine the mesh, resulting in an increased number of finite elements to be
distributed along the geometry lines:
Go to the Mesh mode
Select the Generate mesh button ( ) and in the Mesh settings window choose Fine
for the Elements distribution.
Calculation
Go to the Staged construction mode. All phases are indicated by (blue arrows)
After mesh (re)generation, staged construction settings remain and phase information is rewritten
automatically for the newly generated mesh. However, this is not the case for points for load
displacement curves due to the new numbering of the mesh nodes.
Click on the Select points for curves button ( ) in the toolbar. Reselect the node
located in the centre directly underneath the footing
Click on the Calculate button ( ) to recalculate the analysis. Due to undrained

behaviour of the soil there will be failure in the 3rd and 4th calculation phase.
OUTPUT
As mentioned in the introduction of this example, the compressibility of water is taken into
account by assigning undrained behaviour to the clay layer. This normally results, after
loading, in excess pore pressures. The excess pore pressures may be viewed in the output
window by selecting:
Select in the Phases explorer the phase for which you would like to see output results.
Start the output program by clicking the View calculation results button ( ).
In PLAXIS Output, select from the Stresses menu the option Pore pressures and then
pexcess , this results in figure 19.
The excess pore pressures may be viewed as contour lines ( ), shadings ( ), stress
crosses ( ) or as tabulated output ( ). If, in general, stresses are tensile stresses the
principal directions are drawn with arrow points. It can be seen that after phase 3 on the
left side of the footing there are excess pore tensions due to the horizontal movement of the
footing. The total pore pressures are visualised using the option of active pore pressures.
These are the sum of the steady state pore pressures as generated from the phreatic level
and the excess pore pressures as generated from undrained loading.
Figure 19: Excess pore pressures at the end of the 3rd phase
Select from the Stresses menu the option Pore pressures and then pactive . The results
are given in figure .
From the load displacement curve it can be seen that the failure load in the last phase is
considerably lower for this undrained case compared to the drained situation, as expected.
For the undrained case the failure load is just under 70 kPa.
Figure 20: Active pore pressures at the end of the 3rd phase
APPENDIX A: BEARING CAPACITY

CALCULATION
Given the formula for bearing capacity of a strip footing:
Qf
B
= c Nc + 12 0 B N
0
Nq = e tan tan2 (45 + 12 0 )
0
q 1) cot
Nc = (N
0
2(Nq + 1) tan
(V esic)
N = 1.5(Nq 1) tan 0 (Brinch Hansen)

(Nq 1) tan(1.4 0 ) (M eyerhof )

Filling in given soil data:
Nq = e tan(20) tan2 (55) = 6.4

1) cot(20) = 14.84
Nc = (6.4
2(6.4 + 1) tan(20) = 5.39
(V esic)
N = 1.5(6.4 1) tan(20) = 2.95 (Brinch Hansen)

(6.4 1) tan(28) = 2.97 (M eyerhof )

The effective weight of the soil:
0 = w 10 kN/m3 = 18 10 = 8 kN/m3
For a strip foundation this gives:

1 2
5 14.83 + 2 8 2 5.39 117 kN/m
(V esic)
Qf
B
= c Nc + 12 0 B N = 5 14.83 + 12 8 2 2.95 98 kN/m2 (Brinch Hansen)

5 14.83 + 21 8 2 2.87 97 kN/m2 (M eyerhof )

Critical State Soil Mechanics and Soft Soil Model
Professor Helmut Schweiger
S C I E N C E P A S S I O N T E C H N O L O G Y
CG 04
CRITICAL STATE SOIL MECHANICS
SOFT SOIL MODEL
Helmut F. Schweiger
Computational Geotechnics Group
Institute for Soil Mechanics and Foundation Engineering
Graz University of Technology
Advanced Course on Computational Geotechnics, New Delhi, India, 29 - 31 October 2014
Critical State / Plaxis Soft Soil model

2
Direct Shear Test | Triaxial Test | Critical State Line | Modified Cam Clay | Stress Paths | Plaxis Soft Soil | Possible Enhancements
CONTENTS
Direct shear test

Triaxial tests
Critical state line
Modified Cam Clay model (MCC)
Drained and undrained triaxial stress paths (NC / OC)
Plaxis Soft Soil model
Possible enhancements of Critical State Models
1
3
Mayne, 2006
DIRECT
Slow DirectSHEAR
Shear Tests TEST
on Triassic Clay,NC Slow Direct Shear Tests on Triassic Clay, Raleigh, NC
140 140
n ' Strength Parameters:
(kPa)
Shear Stress, (kPa)
120
Peak
(kPa)=
120
c' = 0; ' = 26.1 o
214.5
100 100

80
Shear Stress,
80 Peak
60 135.0 60
0.491 = tan '
40 40
Peak
20 45.1 20
0 0
0 1 2 3 4 5 6 7 8 9 10 0 50 100 150 200 250
Displacement, (mm) Effective Normal Stress, n' (kPa)
v v

Direct Shear Box (DSB) Direct Simple Shear (DSS)


4
Mayne, 2006
CC
Void Ratio, e
NC NC
CSL CSL
Log v' CSL

tan'
Shear stress
CSSM Premise:
All stress paths fail on the
critical state line (CSL)

c=0
Effective stress v'
2
5
Mayne, 2006
CC
Void Ratio, e
e0
e NC
NC
ef
CSL CSL
Log v' vo
CSL
STRESS PATH No.1 max = c + tan tan'
Shear stress
NC Drained Soil
Given: e0, vo, NC (OCR=1)
Drained Path: u = 0
Volume Change is Contractive:
c=0
vol = e/(1+e0) < 0 Effective stress v' vo

6
Mayne, 2006
CC
Void Ratio, e
e0
NC
NC
CSL
CSL
vf vo Log v' CSL
tan'
STRESS PATH No.2
Shear stress
NC Undrained Soil
max = cu = su u
Given: e0, vo, NC (OCR=1)
Undrained Path: V/V0 = 0
+u = Positive Excess Porewater vf
Pressures vo
Effective stress v'
3
7
Mayne, 2006
CC
Void Ratio, e
CS OC
NC
NC
CSL
CSL
Log v'
p' CSL
Shear stress
tan'
Overconsolidated States:
e0, vo, and OCR = p/vo
where p = vmax = Pc =
preconsolidation stress;
OCR = overconsolidation ratio
Effective stress v' p'

8
Mayne, 2006
CC
Void Ratio, e
e0
OC
CS NC NC
CSL CSL
vo' vf' Log v'

CSL
Stress Path No. 3
Shear stress
tan'
Undrained OC Soil: u
e0, vo, and OCR
Stress Path: V/V0 = 0

Negative Excess u
vo' Effective stress v'
4
9
Mayne, 2006
CC
Void Ratio, e
e0
OC
CS NC
NC
CSL CSL
vo'
Log v'
CSL
Stress Path No. 4 tan'
Drained OC Soil:
e0, vo, and OCR
Stress Path: u = 0
Dilatancy: V/V0 > 0

vo'
Effective stress v'

10
a) b)
Typical results from drained (a) and undrained (b) triaxial tests on normally consolidated soils
(from Atkinson & Bransby, 1978)
5
11
a)
b)
Typical results from drained (a) and undrained (b) triaxial tests on overconsolidated soils
(from Atkinson & Bransby, 1978)

12
DRAINED TRIAXIAL TEST (NC)
1+e
6
13
UNDRAINED TRIAXIAL TEST (NC)

14
7
15
NCL AND CSL IN p-q-v - SPACE

16
UNDRAINED PLANES
8
17
DRAINED PLANES

18
STATE BOUNDARY SURFACE
9
19
OVERCONSOLIDATION
OCR is very important for

soil behaviour

20
M
1
CSL f=0
Right from the M-line (wet side): q < M p d vp 0, d sp 0 (contraction, hardening)

Left from the M-line (dry side): q > M p d vp 0, d sp 0 (dilatancy, softening)
On the ellipse top: q = M p

d vp 0, d sp Failure!
10
21
Stiffness: primary loading
B-C-D-E-F: slope of yield locus becomes flatter

ratio distortional/volumetric strain becomes larger
normally consolidated
drained compression
D. Muir Wood, 1990

22
Stiffness:
- unloading / reloading
- primary loading
lightly overconsolidated
drained compression D. Muir Wood, 1990
11
23
heavily overconsolidated
drained compression
D. Muir Wood, 1990


24
due to plastic soil behaviour
due to change of stress
normally consolidated
undrained compression
D. Muir Wood, 1990

12
25
due
duetotoplastic
plasticsoil
soilbehaviour
behaviour
due
duetotochange
changeofofstress
stress
lightly overconsolidated
D. Muir Wood, 1990


26
due to change of stress
due to plastic soil behaviour
heavily overconsolidated
D. Muir Wood, 1990

13
27
Elastic deformation is generated according to:

p'
e e e0 e ln 0 unloading/reloading
p
Total deformation is generated according to:

p'
e e0 ln primary compression
p0
e = void ratio
= swelling index
= compression index

28
Generally we prefer notation in strains:
p'
ve ve 0 * ln 0 , *
p 1 e
p'
vp vp 0 ( * * ) ln 0
, *

p 1 e
v = volumetric strain
* = modified swelling index
* = modified compression index
14
29
PLAXIS SOFT SOIL MODEL
M
q 1 MC-line
K0NC

p
pp
Soft Soil model:

Mohr-Coulomb failure surface for strength
M-line for determining K0NC
(no longer acts as CSL, only determines shape of cap)

30
MODIFIED CAM CLAY WITH MOHR COULOMB
15
31
p
Input Parameters:
*
* = / 1+e .. Modified compression index
* = / 1+e .. Modified swelling index *
c Cohesion
Friction angle
Dilatancy angle
ur .. Poisson's ratio for unloading
K0nc . Coefficient of lateral earth pressure in normal consolidation

M . K0nc parameter

32
SS vs MC MODEL - OEDOMETER TEST

Chart 1
vertical strain
0
SS
MC
-0.1
-0.2
-0.3
-0.4
-0.5
0 -100 -200 -300 -400 -500 -600
vertical stress [kN/m2]
vertical stress vs vertical strain

16
33
SS vs MC MODEL - OEDOMETER TEST Chart 1

-600
SS
MC
-500
-400
-300
-200
-100
0
0 -50 -100 -150 -200 -250 -300 -350
horizontal stress [kN/m2]
horizontal stress vs vertical stress


34
PRIMARY LOADING - UNLOADING / RELOADING
elastic
Stiffness: primary loading
region
Stiffness: unloading / reloading
current yield surface
Stiffness: unloading / reloading
17
35
MCC-MODEL - FURTHER DEVELOPMENTS
"Bubble models" with

kinematic hardening
e.g. 3-SKH Model
(Baudet & Stallebrass, 2004)
Anisotropic models based on

Modified Cam Clay (rotated yield
surfaces)
e.g. Wheeler, Ntnen, Karstunen
& Lojander (2003)

36
MCC-MODEL - FURTHER DEVELOPMENTS
Atkinson & Sallfors (1991)

e.g. Leroueil & Vaughan (1990)
18
Hardening Soil and HS-small Models
CG5
HARDENING SOIL SMALL MODEL
Helmut F. Schweiger
CG5 - Hardening Soil Small Model

2
Introduction | Description of HS-Model | Parameters | Comparison with Experiments | Influence of Parameters | HS-small | Summary
CONTENTS
Introduction (why advanced model?)

Short description of Hardening Soil Model
Parameters of Hardening Soil Model
Comparison with experimental data
Influence of important parameters
Extension to account for small strain stiffness (HS-Small)
Summary
Advanced Course on Computational Geotechnics, New Delhi, India, 29 -31 October 2014
1
3
Soil behaviour includes:
difference in behaviour for primary loading reloading/unloading

nonlinear behaviour well below failure conditions
stress dependent stiffness
plastic deformation for isotropic or K0-stress paths
dilatancy is not constant
small strain stiffness (at very low strains and upon stress reversal)
influence of density on strength and stiffness
cannot be accounted for with simple

elastic-perfectly plastic constitutive models

4
oedometer test
1-

HS MC
2
5
q [kN/m2]
250
Mohr Coulomb Model

200
HS-Model
150
100
50
0
0 0.01 0.02 0.03 0.04 0.05
eps_axial

6
smax
Model
[mm]
LE 33
MC 36
HS 60
distance [m ]
0 3 6 9 12 15
-0,2
0,2
s / smax [-]
0,4
Linear Elastic
0,6
Mohr Coulomb
All models calculate settlements
0,8 Hardening Soil Differences in shape of trough
1 and maximum values
3
7
Example for vertical displacements behind a retaining wall
Typical vertical displacements behind a retaining wall

(sheet pile wall in clay)
120
Mohr Coulomb
100
vertical displacements [mm]
80
Hard. Soil
60
40
20
0
0 5 10 15 20
-20
-40
distance from wall [m]
> Hardening Soil Model calculates Settlements

> Mohr-Coulomb Model calculates Heave

8
TRIAXIAL TEST
Applied stress path and results for standard drained triaxial test
1
1
3= constant
3
1
isotropic loading
q 1 3 1
dense soil
3
loose dense
qf qr
1 1
q f = failure value vol loose
q r = residual value vol volumetric 1 2 3

4
9
HYPERBOLIC APPROXIMATION OF STANDARD DRAINED TEST
q 1 3
E50
1
50%
Hyperbola
50%
1
qa q
3 a
m 1
E50 Eref
50
2 E50 qa q
pref a
Eref
50 = reference modulus for primary loading at 50% of strength
msand 0.5 ; mclay 1

10
q 1 3
Asymptote
qf
qa R f 0.9
Hyperbola qf
Rf
hyperbolic for q < qf otherwise q = qf
2 sin
qf 3 a a c cot MC failure criterion
1 sin
5
11
q 1 3
Asymptote
qa
Hyperbola
qa q
1
2 E 50 qa q
3 3 qa q
shear strain 1 3 1
2 4 E 50 qa q

12
SHEAR STRAIN CONTOURS IN P-Q-PLANE

c 0

q q
0.05
0.01
p p
sands : m 0.5 clays : m 1

curved lines straight lines
m
3 qa q 3 a 2 sin a
E50 Eref qa (3 a)
50 1 sin a
4E50 q qa pref a
6
13
LINES OF EQUAL SHEAR STRAINS IN TRIAXIAL TEST
Ref. : Ishihara, Tatsuoka and Yasuda (1975). Undrained deformation and liquefaction of sand under
cyclic stresses. Soils and Foundations, Vol. 15, No. 1.

14
2c ' cos ' 2 '3 sin '

R f qa Failure according to MC criterion
1 sin '
m
c ' cos ' '3 sin '
E50 E ref
50 c ' cos ' p ref sin '

m
c ' cos ' '3 sin '
Eur E ref
ur c 'cos ' p ref sin '

Note: Stress-dependent stiffness based on 3
7
15
DEFINITION OF E50
E50, reference pressure = 200 kPa
E50, reference pressure = 100 kPa

1 3 kPa
3 600 kPa
1500
300
500
100
0 0.1 0.2 0.3

16
STIFFNESS IN UNLOADING-RELOADING
Triaxial tests:
Unloading is purely
elastic in HS model
8
17
LINES OF EQUAL VOLUMETRIC STRAINS IN TRIAXIAL TEST
Biarez, J. & Hicher, P.-Y. (1994), Elementary Mechanics of Soil Behaviour, Balkema - Publishers.

18
LINES OF EQUAL VOLUMETRIC STRAINS IN TRIAXIAL TEST

q [MN/m]
vol 0.25
vol 0.20
0.14
0.07
p [MN/m]
9
19
DENSITY HARDENING IN THE HS MODEL
q MC failure line
Cap
pp
pp p
1 m
c ' cot ' p p is determined by K0nc
p
1 m c ' cot ' p ref

v
is determined by Eoedref

20
DEFINITION OF Eoed
m
c cot 1
Eoed E ref

oed c cot p holds strictly for K0-stress paths only
ref
10
21
ELASTIC REGION
f < 0 + fc < 0
q
fc
f
1
pc
p
q = 1 - 3
REGION 1
p = (1 + 2 + 3) / 3 no yield surface active > elastic

22
SHEAR HARDENING
f > 0 + fc < 0
q
2
fc
f
pc
p
REGION 2
shear hardening surface active
11
23
SHEAR AND VOLUMETRIC HARDENING
f > 0 + fc > 0
q
3
fc
f
pc
p
REGION 3
shear hardening and volumetric hardening surfaces active

24
VOLUMETRIC HARDENING
f < 0 + fc > 0
q
f 4
pc p
REGION 4
volumetric hardening surface active
12
25
HARDENING SOIL MODEL IN PRINCIPAL STRESS SPACE

26
FLOW RULE
Volumetric behaviour
"stress dilatancy theory" (Rowe, 1962)
dilatancy angle > non-associated flow rule
vp sin m p
sin m sin cv
sin m
1 sin m sin cv
1' 3'
sin m
1' 3' 2c cot
sin sin
sin cv
1 sin sin
13
27
FLOW RULE
mobilized dilatancy angle for = 35
Note: flow rule in HS-small
35
30 =0
model is slightly different >
25 =5 dilation undrained shear strength
mobilized dilatancy angle []
20 = 20
15 = 35 predicted is different (lower)
10
5
0
-5
-10
-15
-20
-25 contraction
-30 cv cv
-35
0 5 10 15 20 25 30 35
mobilized friction angle []
negative values of are cut-off in Plaxis q

cv
mob
p'

28
FLOW RULE
Volumetric behaviour sin m sin cv
"stress dilatancy theory" (Rowe, 1962) sin m
1 sin m sin cv
plastic potential Q
m > cv = positive > dilation
q
x cv
x m = cv = 0
x
m < cv = negative > contraction
p'
14
29
OEDOMETER TEST - COMPARISON MC / SS / HS

-600
-550
-500
-450 Mohr-Coulomb model:

ratio 3/1 determined by

-400
-350
-300
-250
Hardening (Soft) Soil model:
Hardening Soil Model ratio 3/1 determined
Soft Soil Model
-200 by K0nc
Mohr Coulomb Model
-150
Unloading: ur
-100
-50
0
0 -50 -100 -150 -200 -250 -300 -350
horizontal stress [kN/m2]

30
OEDOMETER TEST - COMPARISON MC / SS / HS

0.0
-0.1
vertical strain
-0.2
Hardening Soil Model
Soft Soil Model
Mohr Coulomb Model
-0.3
-0.4
-0.5
0 -100 -200 -300 -400 -500 -600
vertikal stress [kN/m2]
15
31
OVERCONSOLIDATION IN HS-MODEL
Initial pre-consolidation pressure Cap position based on
pp0 relates to initial p0 previous stress history
(p*, q*)
q
Calculation of p0 based on:
OCR (Over-Consolidation Ratio)
pp
POP (Pre-Overburden Pressure)
(p*,q*)
pP = Isotropic pre-cons. pressure
p = Vertical pre-cons. pressure
'0yy , p p
'0xx , p '0zz , p K0nc p

pp p
p * = 13 '0xx , p '0yy , p '0zz , p
q* | '0xx , p '0yy , p | Stress point due to current initial
stress state
2
q*
p p ( p*)2


32
OVERCONSOLIDATION
Calculation of p0 based on OCR: Calculation of p0 based on POP:
p 0 OCR ' yy 0 p 0 ' yy 0 POP
p0
OCR POP
'yy0
'yy0 p0 'yy0 p0
16
33
PLASTIC POINTS
-1
Cap point
Mohr-Coulomb point
f<0
Cap & Hardening point
Hardening point
-3
-1
Tension point
-3
Tension cut-off: Principal tensile stress is set to zero

34
PLASTIC POINTS
elastic-plastic
elastic
double hardening
17
35
PLASTICITY IN HS MODEL
q
1. Elasticity
(unloading reloading)
5 4 2. Plasticity
3
(compression)
3. Plasticity
(shear)
1 1 2
4. Plasticity
1 1
(compression + shear)
5. Plasticity
(failure criterion)
p

36
PARAMETERS OF HARDENING SOIL MODEL
friction angle
c cohesion
dilatancy angle
ref
E 50 secant modulus from triaxial test (controls deviatoric hardening)
ref tangential modulus from oedometer test

E oed (controls volumetric hardening)
ref
E ur unloading / reloading modulus
ref ref
default: E ur = 3 E 50
m power for stress dependency of stiffness
18
37
ur Poisson ratio for unloading / reloading (default ur = 0.2)
pref reference stress

(default pref = 100 stress units)
K0nc K0-value for normal consolidation (default = 1-sin)

(controls volumetric hardening)
Rf = qf / qa (default Rf = 0.9)

38
SANDS
Correllation for pref=100 kPa (Lengkeek)

ref
E oed I D 60 MPa ID = relative density
19
39
Stiffness of sand
ref
E50 Eoed
ref How can this be true?
p ref
-1= p ref
-3 = p ref
-3 -1
Note: always plot Eoed, Eur distribution for initial stress state when
using HS-model

40
Loose sands: E50ref 15 MPa

Dense sands: E50ref 50 MPa
m for clays:
approx. 0.9 1.0
20
41
For normally consolidated clays (m=1):
ref
Eoed 12 E50
ref
Order of magnitude (very rough)
50000 kPa
ref
Eoed Correlation with Ip for pref =100 kPa
Ip
500 kPa
ref
Eoed Correlation by Vermeer
wL 0.1
ref
Eoed p ref * Relation with Soft Soil model

42
drained
Ohde / Janbu:
m
1
Eoed E ref
oed ref
p
21
43
Parameter limitations
HS model has internal parameters that are computed from our

engineering input parameters > not all combinations of input
parameters can be used. For very soft soils this could be a problem in
certain cases.
E50
pc
Eur Eoed
e.g. E50 / Eoed > 2 difficult to input
p
double hardening

44
TRIAXIAL TESTS ON LOOSE SAND
22
45
TRIAXIAL TESTS ON DENSE SAND

46
TRIAXIAL TESTS ON DENSE SAND
23
47
OEDOMETER TESTS ON LOOSE SAND
-yy [kPa]
400
Hardening soil model
test data
300
200
100
0
0 0.005 0.01 0.015 0.02 0.025
-yy

48
OEDOMETER TESTS ON DENSE SAND
-yy [kPa]
400
Hardening soil model

300
test data
200
100
0
0 0.005 0.01 0.015 0.02 0.025
-yy
24
49
INFLUENCE E50 / Eoed

deviatoric
Stress [kN/m2] stress Triaxial Compression (drained)
250
Reference values
200
150
Eoed increased
100
50
E50 reduced
0
0 -5.00E-03 -0.010 -0.015 -0.020 -0.025 -0.030
Displacement [m]
axial strain

50
INFLUENCE E50 / Eoed
volumetric
Strain strain Triaxial Compression (drained)
0.000
-2.00E-03
-4.00E-03
Eoed increased
-6.00E-03
-8.00E-03
Reference values
-0.010
0 -5.00E-03 -0.010 -0.015 -0.020 -0.025 -0.030 -0.035
axial strain
Strain
25
51
SMALL STRAIN STIFFNESS
Typical curve of degradation of stiffness with strain

52
THE 1-DIMENSIONAL MODEL BY HARDIN & DRNEVICH

1.0
0.8
In its modified form, the reduction curve is specified
by the two parameters
0.6 G0 (or E0) and 0.7
G/G0[]
Experimental results after Santos & Correia

0.4 G 1

G0 1 a( / 0.7 )
0.2
a = 3/7
0.0
1e3 1e2 1e1 1e0 1e1 1e2 1e3
normalized shear strain / 0.7 [-] /0.7[]
NormalisierteScherdehnung
J.A. Santos, A.G. Correia. Reference threshold shear strain of soils. Its application to obtain a unique strain-dependent
shear modulus curve for soils. Proc. 15th ICSMGE, Istanbul (2001), Vol.1, pp. 267-270.
26
53
If the small strain stiffness model after Hardin & Drnevich predicts a stiffness lower than
Gurref (Eurref respectively) the model switches to hardening plasticity of the standard
Hardening Soil model.
IMPORTANT NOTE: flow rule for deviatoric yield surface (volumetric behaviour) is
different in HSS model.
additional input parameters for HS-small
G0ref
40000 40000
TangentmodulusG[kN/m]
SecantmodulusG[kN/m]
30000 0.7 30000
20000 20000
Gurref
10000 10000
HSSmall
Hardin&Drnevich
0 0
1E5 0.0001 0.001 0.01 1E5 0.0001 0.001 0.01
Shearstrain[] Shearstrain[]
see also:
Thomas Benz, Small-Strain Stiffness of Soils and its Numerical Consequences,
Mitteilung 55 des Instituts fr Geotechnik, Universitt Stuttgart, 2007.

54
DRAINED TRIAXIAL TEST / HS VS. HS-SMALL MODEL
160
140
120
HSsmall E0
HS
q [kN/m]
100
80
60
E0ref
40 G0ref
2(1 ur )
20
Et Eur
0
0 -0.002 -0.004 -0.006 -0.008 -0.01 -0.012 -0.014
yy
27
55
DATA FOR SMALL STRAIN STIFFNESS

Hsu & Vucetic Vucetic & Dobry
Shear strain Shear strain
C-C. Hsu, M. Vucetic. Dynamic and cyclic behavior of soils over a wide range of shear strains in
NGI-type simple shear testing device, UCLA Report ENG-02-228, 2002.
M. Vucetic, R. Dobry. Effect of soil plasticity on cyclic response, Journal of Geotechnical

Engineering, ASCE 117 (1991), No. 1, 89-107.

56

Typical values for G0 (in MPa) for reference stress of 100 kPa
Jamiolkowski et al. 1991
28
57
(2.97 e)2
G0ref 33 [MPa] Hardin & Black (1969)
1 e
G0ref RD 70MPa 60MPa Lengkeek
0.385
0.7 2c(1 cos(2 )) 1(1 K0 )sin(2 ) Benz (2007)
4G0
Order of magnitude:
G0ref (2.5 to 10)Gurref
E urref
0.7 (1 to 2) 104 where Gurref
2(1 ur )

58
EXAMPLE DEEP EXCAVATION
Region of small
Region of larger
strains: G > Gur
strains: G ~ Gur
0
3.0
20
40
Gm=Gref /Gurref
Gm []
2.0
60
80
1.0
100
20 0 20 40 60 80 100 120
see also: Thomas Benz, Small-Strain Stiffness of Soils and its Numerical Consequences,
29
59
Mohr-Coulomb Hardening Soil Hardening Soil

Model Model Small Model
Stress dependent stiffness* NO YES YES
Distinction in stiffness for

primary loading and unloading / NO YES YES
reloading
Plastic strains for stress states

below MC - failure line
(deviatoric and volumetric NO YES YES
hardening)
Failure according to
YES YES YES
Mohr-Coulomb
Small strain stiffness NO NO YES

* (not only dependent on 0, this is possible also with MC-Model)
30
Exercise 2: Simulation of Triaxial & Oedometer Tests
(Calibration of HS Model)
Dr William Cheang
Simulation of laboratory tests
SIMULATION OF LABORATORY TESTS
INTRODUCTION
In daily engineering practice soil parameters are obtained from one or more laboratory tests. In order to perform
the best possible Plaxis calculation these soil parameters have to be translated into input parameters for the
constitutive model used, taking into account the possibilities and limitations of the constitutive model. Most
parameters for the constitutive models used in Plaxis can be determined directly from standard laboratory tests
as triaxial tests and oedometer tests. However, due to the complexity of the models it is recommended to
not simply accept the parameters determined from those tests, but to actually model the tests and see if the
parameters found actually give a proper representation of the real laboratory test results within the limits of the
constitutive models. For this purpose the SoilTest module is available in Plaxis with which in a simple manner
laboratory tests can be simulated without the need for making a finite element model.
In this exercise the SoilTest tool will be used for the simulation of both oedometer and triaxial tests on sand and
clay.
CONTENT
Laboratory tests on Sand

Laboratory tests on Clay
Appendix A: Parameter determination
Appendix B: Introduction to the SoilTest tool
How to model an oedometer test

How to model a triaxial test
SIMULATION OF LABORATORY TESTS

In this exercise results from oedometer and triaxial tests are presented for two different materials and the aim is
to determine the parameters for the Hardening Soil model such that a simulation of the tests within Plaxis gives
the best possible results compared to the original laboratory tests. In short:
1. Determine soil parameters based on given real laboratory tests results
2. Perform the laboratory tests using SoilTest with the parameters found
3. Match SoilTest results with the original laboratory results to find the best matching model parameters for
the Hardening Soil model.
Exercise 1: Laboratory tests on sand
Parameter determination
On a sample of dense sand both oedometer tests and triaxial tests have been performed. The results of those
tests are given in the figures below. Use these figures to determine the parameters for the Hardening Soil model
and collect the parameters in Table 1 (see below the figures). Note that it is possible that some parameters
cannot be determined with the given laboratory results, in which case these parameters have to be estimated.
Figure 1: Oedometer test results on sand
Figure 2: Development of horizontal and vertical stress in oedometer test
Figure 3: Triaxial test unloading-reloading (cell pressure = 100 kPa)
Figure 4: Axial vs. volume strain in drained triaxial test
Collect the soil parameters in table 1:
Table 1: Hardening Soil Parameters of the sand

Parameter Unit Value
ref
E50 [kPa]
ref
Eoed [kPa]
ref
Eur [kPa]
pref [kPa]
ur [-]
c [kPa]
0 [o ]
[o ]
m [-]
K0N C [-]
With these data perform a triaxial test in the SoilTest program.
Exercise 2: Laboratory tests on clay
Figure 5: Oedometer test on Clay
Figure 6: Undrained triaxial (CU) tests at cell pressures of 100 kPa and 400 kPa
Figure 7: Undrained triaxial (CU) test at cell pressure of 100 kPa
Collect the soil parameters in table 2:
Table 2: Hardening Soil Parameters of the clay

ref
E50 [kPa]
ref
Eoed [kPa]
ref
Eur [kPa]
pref [kPa]
ur [-]
c [kPa]
0 [o ]
[o ]
m [-]
K0N C [-]
With these data perform an oedeometer test in the SoilTest program.
APPENDIX A: PARAMETER DETERMINATION

SAND
First we determine parameters from the triaxial test data.
Figure 8: Determine stiffness parameters from drained triaxial test
Cohesion and friction angle
For a cell pressure 30 = 100 kPa a maximum value of approximately |10 30 | = 400 kPa is reached at failure.
The Mohr-Coulomb failure criterium is:
1 0
2 |1 30 | + 12 (10 + 30 ) sin c cos = 0
Considering it is sand we assume that the cohesion is zero and so the Mohr-Coulomb failure criterium reduces
to:
|10 30 |
(10 +30 ) = sin
Filling in 30 = 100 kPa and 10 = 500 kPa as obtained from the test we find for the
friction angle0 = 420
Reference stiffness from triaxial test
The triaxial test stiffness E50 is the secant stiffness over the first 50% of the failure value for | 10 30 |. This is
indicated in red in the triaxial test graph of figure 8.
0 =100 kP a 400
E503 = 0.013 = 30800 kP a
The triaxial test stiffness ,E 50 , is within the Hardening Soil model defined as:
m m
c cos30 sin 30

ref ref
E50 = E50 c cos+pref sin , c = 0 E50 = E50 pref
The reference stress pref is chosen equal to the cell pressure of this triaxial test then
ref 0 =100 kP a
E50 = E503 30000 kPa
Reference unloading-reloading stiffness
Similar to the determination of the reference stiffness for triaxial testing the reference unloading-reloading stiffness
can be determined. In the triaxial test results an unloading-reloading cycle is done for this. The Hardening Soil
model does not have unloading-reloading behaviour with hysteresis but simple non-linear elastic unloading-
reloading behaviour. Therefore a secant value is taken for the unloading-reloading behaviour, as given with the
green line in the triaxial test results.
0 =100 kP a 400
Eur3 = 0.0260.021 = 80000 kPa
Under the same assumptions as for the stiffness in triaxial testing counts:
ref 0 =100 kP a
Eur = Eur3
But this is a bit low value for the unloading reloading stiffness and so
ref
Eur = 90000 kPa
is chosen
Dilatancy angle
From the plot of axial strain versus volume strain the dilatancy angle can be determined according to
v
sin = 21 +v
See figure 9 for details.

With v = 0.048-0.004 = 0.044 and 1 = -0.09-(-0.03) = -0.06 the dilatancy can be calculated as =16o
Note: The Poissons ratio needed for the Hardening Soil model cannot be determined from this graph as this
graph represents an oedometer test in primary loading and the Poissons ratio needed is an unloading-reloading
Poissons ratio.
An acceptable value for the unloading-reloading Poissons ratio is ur = 0.2.
Figure 9: Determination of diltancy angle from drained triaxial test
Oedometer stiffness and power of stress dependent stiffness
From the oedeometer test results we determine the stiffness Eoed for vertical stresses y0 = 100 kPa en y0 =
200 kPa, see figure 10. Note that Eoed is a tangent stiffness. Make sure to use the primary loading part of the
oedometer test results.
0 =100 kP a
y 3200
Eoed = 1.4%0.33% = 29900 kPa
y0 =200 kP a 4000
Eoed = 1.4%0.47% = 43000 kPa
Within the Hardening Soil model the stress dependent oedometer stiffness is defined as:
m m
c cosy0 sin y0

ref ref
Eoed = Eoed c cos+pref sin , c = 0 Eoed = Eoed pref
Choosing the reference pressure pref = 100 kPa gives
ref 0 =100 kP a
Eoed = Eoed
3
30000 kPa
The power m for stress dependent stiffness can now be determined as:
0 =200 kP a m
y
y0

Eoed 43000 200 m

ref
Eoed
= pref 30000 = 100 m = 0.5
Figure 10: Determination of oedometer stiffness and power of stress dependency
K0 value for normal consolidation
The K0 value for normal consolidation (K0N C )can only be obtained if measurements for horizontal stresses have
been performed during the oedometer test. If so, results as given in figure 11 may be obtained. From the primary
loading line can be obtained that
0
x 30 100
K0N C = y0 = 10 = 300 = 0.33
Alternatively one can use Jakis formula
K0N C 1 sin = 1 sin(42o ) = 0.33
Figure 11: Horizontal/vertical stress ratio during oedometer test
Note on unloading-reloading stiffness
If no triaxial test with unloading-reloading is available the unloading-reloading stiffness can also be determined
from an oedometer test with unloading. However, the unloading-reloading stiffness required for the Hardening
Soil model is stress dependent on 3 while the oedometer test results presented in figure 10 give the strain vs
the vertical stress y (= 1 voor oedometer testing).
0 =100 kP a 0 =100/K0N C kP a 0 =300 kP a 400

Eur3 = Eur1 = Eur1 = 1.28%0.91% = 108000 kPa
With pref = 100 kPa (pref refers to 30 !) it follows that
ref 0 =100 kP a
Eur = Eur3 110000 kPa
ref
This is a bit high and so a value of Eur = 90000 kPa is chosen.
Table 3: Summary of Hardening Soil Parameters for the sand

ref
E50 [kPa] 30,000
ref
Eoed [kPa] 30,000
ref
Eur [kPa] 90,000
pref [kPa] 100
ur [-] 0.2
c [kPa] 0
0 [o ] 42
[o ] 16
m [-] 0.5
K0N C [-] 0.33
CLAY
Cohesion and friction angle
We start with the determination of the strength parameters based on the CU triaxial tests.
Figure 12: Determination of soil strength parameters for clay
The black dotted lines is the Mohr-Coulomb failure criterium in the p-q plane. In principal stresses the Mohr-
Coulomb failure criterium is defined as:
|1 3 | 1 +3

2 + 2 sin c cos = 0
With p0 = (10 + 230 )/3 and q = 10 30 under triaxial test conditions this can be rewritten as:
2p0 + 13 q

q 6sin 0 6c cos
2 = 2 sin c cos = 0 q = 3sin p + 3sin
Hence, the slope M of the Mohr-Coulomb line in p-q plane is defined as:
6sin 195
M= 3sin = 200 = 250
From the intersection between Mohr-Coulomb line and the vertical axis where p=0 the cohesion can be determined:
6c0 cos
q= 3sin = 0 c = 0 kPa
Reference oedometer and unloading-reloading stiffness
From the results of the oedometer test the oedometer stiffness as well as the unloading-reloading stiffness can
be determined. As the graph is given on logarithmic scale one cannot simply draw a tangent line as was done
for the oedometer test on sand.
Figure 13: Determination of oedometer and unloading/reloading stiffness
Considering that both primary loading and unloading/reloading paths are straight lines in the log(p)-v graph,
hence they have a relation of the form:
y = v = A log(y0 )
2 1 0.3700.270
A= log(2 )log(1 ) = log(120)log(30) =0.166
In order to determine the stiffness we calculate the derivative of the strain over the stress and change to natural
logarithm:
ln(y0 )
y = v = A ln(10)
dy dy0 ln(10)
dy0 =A 1
ln(10) 1
y0 E= dy = A y0
The E modulus found is the oedometer stiffness can be rewrittens as:
y0

ln(10)
E = Eoed = A pref pref
In the Hardening Soil model the oedometer stiffness is defined as (assuming c = 0) :
m
y0

ref
Eoed = Eoed pref
Hence:
ref ln(10)
Eoed = A pref and m=1
If we choose pref = 100 kPa and with the previously determined A = 0.166 we get:
ref ln(10) 2.3

Eoed = A pref = 0.166 100 = 1.4 MPa.
The determination of the unloading-reloading stiffness follows the same method:
y = v = B log(y0 )
2 1 0.4270418
B= log(2 )log(1 ) = log(120)log(30) =0.0149
dy0 ln(10)
Eur = dy = B y0
However, the Eur in the Hardening Soil model is dependent on the smallest principal stress, which is x0 in an
oedeometer test and not y0 .
During the unloading process there is no linear relation between horizontal and vertical stress, as in the beginning
of unloading y0 > x0 where as after much unloading y0 < x0 . Therefore the assumption is made that during
unloading on average x0 = y0 .
0

ln(10) ln(10) ln(10) x
Eur = B y0 = B x0 = B pref pref
With the definition of Eur in the Hardening Soil model of
0
m
ref x
Eur = Eur pref
ref
Follows, in a similar way as for the Eoed , that
ref ln(10) 2.3

Eur = B pref = 0.0149 100 =15 MPa and m = 1
Stiffness from triaxial test
As only undrained triaxial test data is available it is only possible to determine an undrained E50 and not an
effective E50 . Therefore the only solution is to estimate the E50 with several runs of the SoilTest program using
different input values for the reference E50 until the best fit for the undrained triaxial test data is found. Typically
for normally consolidated clays the effective reference E50 is in the range of 2-5 times the effective reference
ref
Eoed , hence this can be used as a start value for the estimation procedure. By doing so a value E50 3.5 MPa
of is found.
K0 value for normal consolidation
The K0-value for normal consolidation can only be obtained if measurements for horizontal stresses have been
performed during the oedometer test. As this is not the case here we can only use the estimation according to
Jakys rule:
K0N C 1 sin = 1 sin(250 ) =0.58
Poissons ratio
The Poissons ratio for unloading and reloading is again estimated as ur = 0.2
Table 4: Summary of Hardening Soil Parameters for the clay

ref
E50 [kPa] 3,500
ref
Eoed [kPa] 1,400
ref
Eur [kPa] 15,000
pref [kPa] 100
ur [-] 0.2
c [kPa] 0
0 [o ] 25
[o ] 0
m [-] 1.0
K0N C [-] 0.58
APPENDIX B: INTRODUCTION TO THE SOILTEST TOOL

For the simulation of laboratory tests Plaxis offers the SoilTest tool based on a single stress point calculation
that makes it possible to do fast simulations without the need for a finite element mesh. The SoilTest tool can be
called from within the material sets database or from within the definition of a material set. (see figure ).
Figure 14: The SoilTest tool
In the following paragraphs a step-by-step description is given on how to model both an oedometer test and a
triaxial test with the help of many screen shots of the SoilTest tool. Please note that any parameters given on
those screen shots have no relation with the actual exercise and are solely for illustrating the possibilities of the
SoilTest tool.
How to model an oedometer test
In order to model an oedometer test first the material data set has to be created. After doing so, press the
<SoilTest> button to start the SoilTest tool. The window that opens is show in figure .
Figure 15: Main window of the SoilTest tool
In the main window select the Oedometer tabsheet and set the parameters as indicated in Figure .
Figure 16: Setting the oedometer test parameters
After the the oedometer test has been calculating graphs with results appear at the bottom of the SoilTest window.
The user can double-click these graphs to view them in separate windows. Furthermore, custom charts can be
added, see figure 4.
Figure 17: Inspect oedometer test results
How to model a triaxial test

From the material database or the material set definition window press the <SoilTest> button to start the SoilTest
tool. In the main window choose the tabsheet Triaxial and set the type of test as well as the test parameters as
shown in figure
Figure 18: Defining a triaxial test
After the triaxial test has been calculated graphs with results appear at the bottom of the SoilTest window. As
described above for the oedometer test, the user can double-click this graphs to view them in separate windows
as well as add custom charts.
Modelling a triaxial test with unloading/reloading
The standard functionality in SoilTest for simulation of a triaxial test does not allow for an intermediate unloading-
reloading path. However, the SoilTest functionality contains a General option with which soil test can be defined
in terms of boundary stresses or strains on all sides of a soil test cube. Hereafter it will be shown how this can
be used for the simulation of a triaxial test with unloading/reloading path.
After opening the SoilTest option from the material set definition window the tabsheet General should be chosen.
On this tabsheet a list of calculation phases can be defined where stress or strain increments can be applied.
Initial phase
First of all we have to specify whether stresses or strains will be applied on the boundaries during the test. For
this exercise stresses will be applied. Now the values of the initial stresses on the soil sample have to specified.
For a triaxial test the initial stresses are the cell pressures acting on the soil, hence for xx , yy and zz the cell
pressure has to entered. The cell pressure is a water pressure and so there will be no shear stress acting on the
soil: xy = 0. See figure for details.
Figure 19: General option for simulation of laboratory tests used for triaxial test
Phase 1
Apply a stress increment in vertical direction (yy ) until the stress level where the unloading path should start.
Note that the horizontal stresses (xx and zz ) remain the same as they represent the cell pressure. Hence,
the horizontal stress increments are zero in this phase.
Phase 2
Press the Add button to add another phase to the phase list. This phase represents the unloading phase. See
figure for details.
Phase 3
Press the Add button once more in order to add the 3rd phase. This phase represents the reloading of the soil
as well as the continuation of primary loading until either failure or a higher stress level from where for instance
another unloading/reloading cycle is going to be made.
Figure 20: Unloading/reloading cycle in a triaxial test using the General option
Modelling of Deep Excavations
CG7: MODELLING OF DEEP EXCAVATIONS
Antonio Gens

Helmut Schweiger (Technical University of Graz, Austria) (many!)
Outline
Introduction
Fundamentals of excavation modelling
Excavation
Hydraulic conditions
Thin wall vs. thick wall
Wall support
Interfaces
Simplified example: effect of constitutive model

A note on factors of safety
Introduction
Analysis of a deep excavation involves modelling of
various excavation stages

interaction wall / soil > interface elements
changes in groundwater level
struts or anchors (including load transfer to soil)
adjacent structures (buildings, tunnels, piles, )
Requires advanced constitutive model because
stress paths in soil are not monotonic (significant change in stress path
direction)
primary loading and unloading / reloading occurs in different parts of the domain
analysed
some areas will experience large strains with significant plastic deformations,
others will be in the very small strain range
> simple elastic - perfectly plastic models not suffcient
Introduction
2D model
sufficient
Introduction
3D model required

Typical excavation sequence:
I 1st excavation
II Pre-stressing anchor
III Final excavation
I 1st excavation
The domain excavated is removed from the mesh

In the nodes on the excavated boundary, out-of-balance forces appear:
they are exactly equivalent to the effect of excavation

II Pre-stressing anchor
The anchor element is installed (created)

The pre-stress load is applied
III Final excavation
The excavated domain is removed and the corresponding out of balance

forces appear
Attention to hydraulic conditions!
Wet excavation
Simply click on the cluster or clusters to be excavated
The water remains in the excavated domain, water pressures automatically
appear
Hydraulic conditions do not change
Dry excavation
Click on the cluster or clusters to be excavated
Use cluster dry option or cluster phreatic line
Set new hydraulic conditions
Setting of new hydraulic conditions
Consider each individual case carefully
Select the relevant situation: undrained, drained or consolidation
it may be different for different materials
Specifying pore pressure distributions

General phreatic level
Applies to all clusters that have not been separately defined
Cluster phreatic level
Applies to one specific cluster
Cluster dry
Makes a specific cluster dry
Interpolate
Interpolates pore pressures between clusters above and below
User-defined pore pressure
Specify pore pressure at reference level and the rate of increase in the y direction
Groundwater flow calculation
Gives the steady-state solution
Groundwater flow calculation: example
Equal pore pressure values on both

sides at the foot of the wall!
Groundwater flow calculations

Plate elements are permeable
Interface elements are impermeable (unless
deactivated in the groundwater flow calculation)
Dewatering
Z-shape phreatic level gives wrong results:
1 2 3
General
General
6
General
7
No equilibrium in horizontal water pressures:
Local peak stresses
5
Local peak strains
Non-physical horizontal displacements
8
Non-physical excess pore pressures

0 4
Possible incorrect water pressure acting on

wall
Walls thin wall vs. thick wall
Thin wall
Wall thickness << wall length
Shearing not important
No end-bearing, only friction
Plate element suffices
Thick wall
Wall thickness significant
Shearing important
End-bearing capacity needed
Use soil elements with material set representing wall material
In order to obtain structural forces a plate may be inserted

Walls thick wall
Soil elements with material set representing wall material
Difficult to obtain structural forces from soil elements, therefore introduce
plate:
No influence on deformation: low stiffness, no weight
Located in on the neutral line (usually the middle)
Tight bonding to the concrete elements: no interfaces d
Soil elements: Esoil=Ewall, I = 1/12*d3 , d = wall thickness
Plate element: EI = EsoilI / x, choose x large (e.g. 10000)
uplate = usoil Mwall = x*Mplate, Qwall = x*Qplate
NOTE: Nwall = Nplate+y,0 , y,0 = initial vertical stress in

soil elements
Wall support
Lagging
Use short additional section of plate
perpendicular to the wall
Create short section with increased
stiffness using multiple chains
Struts
Full excavation: node-to-node anchor
Half (symmetric) excavation: fixed-end
anchor
Anchors
Grout anchor: node-to-node anchor +
geotextile for grout body
Ground anchor: node-to-node anchor +
perpendicular plate element
Wall support
Ground anchors
Axial forces in ground anchors:
Input geometry
real distribution of axial forces in

ground anchor
Generated mesh
axial forces in geotextile element
Nrod <> Ngrout due to shared node between anchor, geotextile and soil
Interfaces
Soil-structure interaction
Wall friction
Slip and gapping between soil and structure
Interface material properties
Taken from soil using reduction factor Rinter
Cinter = Rinter * Csoil
tan(inter) = Rinter * tan(soil)
inter = 0 for Rinter < 1, else inter = soil
Ginter = (Rinter)2 * Gsoil
Some building codes prescribe soil-wall friction angle :
Rinter = tan()/ tan(soil)
Individual material set for interface
Interfaces
Suggestions for Rinter:

Interaction sand/steel = Rinter 0.6 0.7
Interaction clay/steel = Rinter 0.5
Interaction sand/concrete = Rinter 1.0 0.8
Interaction clay/concrete = Rinter 1.0 0.7
Interaction soil/geogrid = Rinter 1.0
(interface may not be required)
Interaction soil/geotextile = Rinter 0.9 0.5 (foil, textile)
Stability should not be critical on Rinter !
Goal of study
assess influence of relative stiffness between wall and soil on results
assess influence of wall length
show influence of constitutive model
> show trends and qualitative behaviour rather than
quantitative comparison
Constitutive models
Mohr-Coulomb
Hardening Soil
HS_small
Soft Soil (excavation in clay only)
Modified Cam Clay (excavation in clay only)
Results
wall deflection
bending moments / strut forces
earth pressure distribution (active - passive)
vertical displacements behind wall

MODELS COMPARED
Mohr-Coulomb Model
q 1 3
perfectly plastic
elastic
Note: In this study stiffness for variations of MC analyses is assumed as:

MC_a > EMC = (E50 + Eur)HS at level of base of wall
MC_b > EMC = Eur, HS at level of base of wall
MC_c > EMC = E50, HS at level of base of wall
MC = 0.3
MODELS COMPARED
Hardening Soil Model
q
plastic
elastic

MODELS COMPARED
Hardening Soil Small Model
additional input parameters for HS-small
G0ref
40000 40000
TangentmodulusG[kN/m]
SecantmodulusG[kN/m]
30000 0.7 30000
20000 20000
Gurref
10000 10000
HSSmall
Hardin&Drnevich
0 0
1E5 0.0001 0.001 0.01 1E5 0.0001 0.001 0.01
Shearstrain[] Shearstrain[]
Thomas Benz, Small-Strain Stiffness of Soils and its Numerical Consequences,

MODELS COMPARED
Soft Soil Model

PROBLEM SPECIFICATION
Wall length:
Dense sand and marl: 9 m
Medium dense sand: 10 m
Clay: 11 m
DENSE SAND
Parameters for Hardening Soil model (HS)
Parameter Meaning Value
[kN/m] Unit weight (unsaturated) 18
sat [kN/m] Unit weight (saturated) 20
[] Friction angle (Mohr-Coulomb) 41
c [kPa] Cohesion (Mohr-Coulomb) 0
[] Angle of dilatancy 15
ur [-] Poissons ratio unloading-reloading 0.20
E50ref [kPa] Secant modulus for primary triaxial loading 30 000
Eoed ref
[kPa] Tangent modulus for oedometric loading 30 000
Eurref [kPa] Secant modulus for un- and reloading 90 000
m [-] Exponent of the Ohde/Janbu law 0.55
pref [kPa] Reference stress for the stiffness parameters 100
K0nc [-] Coefficient of earth pressure at rest (NC) 1-sin()
Tension [kPa] Tensile strength 0
Additional parameters for Hardening Soil Small model (HSS)

G0 [kPa] Small-strain Shear modulus 112 500
0,7 [-] Reference shear strain where Gsec=0.7G0 0.0002

DENSE SAND
horizontal wall displacement [mm] bending moments [kNm/m]

15 12 9 6 3 0 -3 -6 -80 -60 -40 -20 0 20 40
0 0
HS HS
HSS HSS
MC_a 1 MC_a 1
MC_b MC_b
MC_c MC_c
2 2
depth below surface [m]
3 3
4 4
5 5
6 6
7 7
8 8
9 9
DENSE SAND 2
earth pressure [kN/m ]
-120 -100 -80 -60 -40 -20 0 20 40
0
HS
HSS
MC_a 1
0 4 8 12 16 20 24
14
surface displacement [mm]
2
12
10 HS
HSS
8

MC_a 3
6 MC_b
4 MC_c
2 4
0
-2
-4 5
-6
-8
6
Strut forces: 7
HS: -102 kN/m
HSS: -107 kN/m 8
MC_a: -78 kN/m
MC_b: -83 kN/m
9
MC_c: -72 kN/m

LOOSE TO MEDIUM DENSE SAND
r [kN/m] Unit weight (saturated) 20
Eoedref [kPa] Tangent modulus for oedometric loading 16 000
Rf [-] Failure ratio 0.90


35 30 25 20 15 10 5 0 -5 -10 -140 -120 -100 -80 -60 -40 -20 0 20 40
0 0
HS HS
HSS HSS
MC_a 1 MC_a 1
MC_b MC_b
MC_c MC_c
2 2
3 3

4 4
5 5
6 6
7 7
8 8
9 9
10 10


0 5 10 15 20 25
25
20
15
10
5
0
-5
HS
-10 HSS
-15 MC_a
-20 MC_b
-25 MC_c
-30
Strut forces:
HS: -151 kN/m
HSS: -154 kN/m
MC_a: -119 kN/m
MC_b: -127 kN/m
MC_c: -105 kN/m
Ratio G/Gur at final stage

CLAY

CLAY
Parameters for Soft Soil model (SS)
ur [-] Poissons ratio 0.20
* [-] Modified swelling index 0.0125
* [-] Modified compression index 0.0556

CLAY

120 110 100 90 80 70 60 50 40 30 20 10 0 -10 -150 -120 -90 -60 -30 0 30
0 0
HS HS
HSS HSS
1 1
MC_a MC_a
MC_b MC_b
MC_c MC_c
SS 2 2
SS
3 3
4 4
5 5
6 6
7 7
8 8
9 9
10 10
11 11
CLAY

0 10 20 30 40 50 60 70
120
100 HS
80 HSS
MC_a
60 MC_b
40 MC_c
SS
20
0
-20
-40
-60
Strut forces:
HS: -115 kN/m
HSS: -120 kN/m
MC_a: -96 kN/m
MC_b: -93 kN/m
MC_c: -103 kN/m
SS -75 kN/m

MARL
highly overconsolidated > POP = 2 000 kPa, K0 = 2.0

MARL
Parameters for Soft Soil model (SS)

ur [-] Poissons ratio 0.20
* [-] Modified swelling index 0.0018
* [-] Modified compression index 0.0067

MARL

12 10 8 6 4 2 0 -2 -80 -60 -40 -20 0 20 40 60 80 100
0 0
HS HS
HSS HSS
MC_a 1 MC_a 1
MC_b MC_b
MC_c MC_c
SS SS
2 2
3 3
4 4
5 5
6 6
7 7
8 8
9 9
MARL

0 10 20 30 40 50 60 70
12
10 HS
HSS
8 MC_a
MC_b
6
MC_c
4 SS
-2
-4
Strut forces:
HS: -254 kN/m
HSS: -208 kN/m
MC_a: -212 kN/m
MC_b: -192 kN/m
MC_c: -239 kN/m
SS -195 kN/m

MARL
Ratio G/Gur at final stage
Material behaviour
Mohr-Coulomb: unrealistic deformations

Overestimation over bottom heave
Often heave of soil behind the wall
Hardening Soil model: qualitative realistic deformations

Better bottom heave, but increases with model depth
Settlement trough behind wall, but often too shallow and too wide
HS-small model: qualitative and quantitative realistic deformations

Good bottom heave independent of model depth
More realistic settlement trough behind the wall (narrower and
deeper)

NOTE ON / c - REDUCTION
Material parameters soil layer:
sheet pile
type AZ-18 = 35
1.0 m c = 0.1 kN/m

1.5 m
= 17 kN/m
= 0
6.5 m Properties sheet pile wall:
EA = 3.008E6 kN/m
2
EI = 6.84E4 kNm /m
4.0 m
Mpl = 505 kNm/m
Properties strut:
2
E = 3.0E7 kN/m
A = 0.24 m2
Horizontal strut distance: 1 m
wall elastic
Msf = 1.95

wall elastic - perfectly plastic

Msf = 1.73
INFLUENCE OF DISCRETISATION
without interfaces with interfaces

mesh type number of number of nodes number of number of nodes
elements 6 noded 15 noded elements 6 noded 15 noded
1 116 267 997 137 330 1207
2 224 493 1881 260 589 2217
3 430 923 3565 464 1021 3897
4 935 1958 7655 1010 2149 8337
5 1785 3678 14495 2060 4301 16841
6 4876 9893 39289 4558 9405 37696

INFLUENCE OF DISCRETISATION
Influence of discretisation and interface

elements on calculated factor of safety
Final remarks
Influence of constitutive model on results of numerical
analysis of deep excavations has been addressed
Simple constitutive models have severe limitations in

predicting realistic displacements for deep excavations, in
particular surface settlements behind the wall
For highly overconsolidated soils differences in results for

different constitutive models is less pronounced (behaviour is
predominantly elastic) but small strain stiffness remains
important
Effect of structural elements on safety factor has been

discussed
Structural Elements in PLAXIS
Professor K.Rajagopal
Structural Elements in PLAXIS 2D
Dr William Cheang Plaxis AsiaPac, Plaxis Academy
Contribution:
Dr William Cheang
Ir Dennis Waterman
Prof. K Rajagopal
Dr Ronald Brinkgreve
Contents
1. StructuralelementsavailableinPlaxis
2. UsageofstructuralelementsinFEmodelling
3. Plate elements(BeamandShellelement)
4. Anchor elements(Springelement)
5. Geotextile elements(Membraneelement)
6. Interface elements(Zerothicknesselement)
7. EmbeddedPileRowelements
1.Structural elements in Plaxis
Section 3.42 & 14.5

1. Plateelement
Section 3.45, 3.46 & 14.1

2. Anchorelement
Section 3.43 & 14.3
3. Geogridselement
Section 3.44 & 14.1
4. Interfaceelement
5. Embeddedpilerowelement
wall strip footing tunnel
geotextile wall ground anchor cofferdam
strut anchored wall Incremental wall 4

3.1 Plate Element
1. 3or5nodedlineelements(for6nodedor15nodedelementmesh)
2. 3degreesoffreedompernode
3. Plateshave:
o Axialforces
o Shearforces
o Bendingmoments
o Hoopforces(axisymmetry)
4. Elasticorelastoplasticbehaviour
5. Formodellingwalls,floors,tunnels
3.2 Plate Element
Plates elasticparameters
h3 b
EI E (b = 1 m)
12
EA E h b (b = 1 m)
EI (Equivalent rectangular
d h 12
EA plate thickness)
h h
b
b = 1 m in plane strain
b = 1 meter in axisymmetry
b
Plaxis Advanced Course, New Delhi, India (29-31 October 2014) 172/448 6
PLATES:elastoplasticbehaviour
Np
M
Mp
3.4 Plate Element 1200

Envelope
Elastoplastic plate
1000
Elasticplate
800
600
N
400
200
0
200 0 200 400 600 800 1000 1200
M
10090 80 70 60 50 40 30 20 10 0 10 20 30 40 50 60 70 80 90 100
0
Elastoplastic plate
10 Elasticplate
N
15
20
25
M
Effect on Global FOS by c/phi Reduction
CBPElastic,Failure CBP Elasto-Plastic

withnoPlastic Failure with Plastic
Hinge, Hinge, FOS=1.40
FOS=1.75
1. Elastic wall excludes possibility of wall plastic hinge; and over-estimate FOS=1.75
2. Allowing for wall plastic hinge (Elasto-plastic wall) gave lower FOS=1.40 and smaller soil yielded
zone behind the wall
9
Plates connections Spring data:

Stiffness
6 8 Min/Max moment
Rotation
spring
5 7
Hinged connection
Rigid connection
(default)
Illustration: Connection.P2D
Walls thinwallvs.thickwall
A. Thinwall
1. Wallthickness<<walllength
2. Nomuchendbearing,onlyfriction
Plateelementsuffices
A. Thickwall
1. Wallthicknesssignificant
2. Endbearingcapacityneeded
Usesoilelementswithmaterialsetrepresen ngwallmaterial
Inordertoobtainstructuralforcesaplatewithfictitiousproperties
maybeinserted
Walls thick wall
1. Soilelementswithmaterialsetrepresentingwallmaterial
2. Difficulttoobtainstructuralforcesfromsoilelements,
thereforeintroduceveryflexibleplatewithinthesolidwall
elements:
Noinfluenceondeformation:lowstiffness,noweight
Locatedinontheneutralline(usuallythemiddle)
Tightbondingtotheconcreteelements:nointerfaces
(Illustration: Beam.P2D): d
Solid elements: Esoil=Ewall, I = 1/12*d3 , d = wall

thickness
Plate element: EI = EsoilI / x, choose x large (e.g. 106)
uplate = usoil Mwall = x*Mplate, Qwall = x*Qplate
Placing a Hinge for free rotation of plates
Can be placed at joint between plate elements allows for free rotation 13
Full-height panel support to reinforced soil walls rigid Incremental panel support to reinforced soil walls
body rotation of facing bulging type deformation at front end
2. Anchor Element
Anchors:Nodetofixed
Tomodelsupports,anchorsandstruts
a) Elastoplasticspringelement
b) Oneendfixedtopointinthegeometry,otherendisfullyfixedfor
displacement
c) Positioningatanyangle
d) Prestressingoption
Anchors nodetonode
Tomodelanchors,columns,strutsandrods
a) Elastoplasticspringelement
b) Connectstwogeometrypointsinthegeometry
c) Nointeractionwiththemeshalongtheanchor
rod
d) Prestressingoption
15
Anchors materialproperties
Axialstiffness,EA (foroneanchor) [kN]
Spacing,Ls (outofplanedistancebetweenanchors) [m]
Maximumanchorforceforcompressionandtension,|Fmax,comp|and|Fmax,tens|[kN]
Ls
Anchors prestressing
A. DefinedinStagedconstructionphase
B. Bothtension(groutanchor)orcompression(strut)
possible
Tension = positive
17
Geogrids
1. 3or5nodedlineelement
2. Elasticorelastoplasticbehaviour
3. Noflexuralrigidity(EI),onlyaxialstiffness(EA)
4. Onlyallowsfortension,notforcompression
+ Geogrid Element
Anchor + Geogrid Element
Groundanchors
1. Combinationofnodetonodeanchorandgeogrid
2. Nodetonodeanchorrepresentsanchorrod(freelength)
(nointeractionwithsurroundingsoil)
3. Geogridrepresentsgroutedpart(fullinteractionwithsurroundingsoil)
4. Nointerfacearoundgroutedpart;interfacewouldcreateunrealisticfailuresurface
5. Workingloadconditionsonly nopullout
6. Ifpulloutforceisknownthiscanbeusedbylimitinganchorrodforce
19
Modeling of soil nails using

geogrid elements
Axial force distribution along fixed length (modelled using geogrid)
Nrod <>Ngrout duetosharednode

betweenanchor,geotextile
andsoil
Probableactualdistributionofaxial
forcesingroundanchor
axialforcesingeotextileelement
Input geometry Generated mesh

21
GroundAnchors:Influenceofnodenumbersalong
structuralelements
22
InterfaceElement
1. Soilstructureinteraction
1. Wallfriction
2. Slipandgappingbetweensoilandstructure
2. Soilmaterialproperties
A. TakenfromsoilusingreductionfactorRinter
3. Individualmaterialsetforinterfacepossible
23
Interfaces soil structure interaction

A. Doublingofnodesto(partially)uncouplesoilandstructuralelement
B. Springconnectionbetweensoilnodesandstructuralnodes
1. Normalspring
2. Shearspring
C. Allowsformodellingofslip,gappingandclosingbetweensoilandstructure
D. Canalsobeusedbetweentosoilmaterials
Output:
Normal stresses
Shear stresses
Displacements
Interfaces material properties
Soilmaterialproperties
TakenfromsoilusingreductionfactorRinter
Cinter =Rinter *Csoil
tan(inter) =Rinter *tan(soil)
inter =0forRinter <1
=soil Rinter =1
t,inter =Rinter *t,soil
Ginter 2
= (Rinter) *Gsoil
1. ResidualreductionfactorRinter,res onlyaffectsstrength,
notstiffness
2. Individualmaterialsetforinterfacepossible
Interfaces material properties

1. Residualstrengthafterreachingmaximumshearstrength
2. Gapclosure
Foundation Pressure Gap opening Gap closing Foundation

pressure reduction (soil shrinkage) (soil swelling) pressure
(soil shrinkage) (soil swelling)
6.2 Interface Element
Interfaces reductionfactor
SuggestionsforRinter:
1. Interactionsand/steel =Rinter 0.6 0.7
2. Interactionclay/steel =Rinter 0.5
3. Interactionsand/concrete =Rinter 1.0 0.8
4. Interactionclay/concrete =Rinter 1.0 0.7
5. Interactionsoil/geogrid(groutedbody) =Rinter1.0
(interfacemaynotberequired)
6. Interactionsoil/geotextile =Rinter0.9 0.5(foil,textile)
With reference to BS8002:
27
Interfaces
Trytoomitstressoscillationsatcornersofstiffstructures
Inflexible
corner points,
may cause bad
stress results
Flexible corner
points with
improved stress
results
Embedded pile row
A. 3or5nodedlineelement
B. Interactionwithsoilthroughaninterfaceallowingforskinfriction
(linearskinresistancedistribution)
C. Endbearingcapacitythroughspringconnection
D. Soilcanflowinbetweenpiles
E. AvailablefromPlaxis2D2012
Embeddedpilerow parameters
Material data
Pile shape
Rayleigh damping (dynamics)
Skin resistance
Base resistance
Interface stiffness factors
Embedded pile row parameters
1. Materialdata(stiffnessE andmaterialweight)
2. Pileshape(circularorsquare,hollowormassive,userdefined)
3. Crosssectionaldata(areaA,momentofinertiaI)
4. Outofplanespacing
5. Skinresistanceattopandbottomofthepilerow(Ttop,max andTbottom,max)
6. Baseresistance(Fmax)
7. Interfacestiffnessfactors
Determinedbycurvefittingonpredefinedloaddisplacementcurves
DefaultvaluesobtainedfromtheloaddisplacementcurveforapileinDutch
soilconditions(boredpileinsand,submerged)accordingtothenational
annexofEurocode7.
Modelling pile rows in 2D

Structures are predominantly 2D, however with local 3D elements
How to model in 2D?
32 Plaxis Advanced Course, New Delhi, India (29-31 October 2014) Embedded 185/448
pile row
Current options: plates and n2n anchors
Limitations:
Plates+interfaces:
discontinuous mesh,
possibly unrealistic shear
planes
N2n anchor: no soil-
structure interaction, no
bending stiffness
33 Embedded pile row

New option: embedded pile row
Combines advantages of plate
and n2n anchor:
Continuous mesh (soil can
flow through pile row)
Axial and bending stiffness
Soil-structure interaction due
to line to line interface
No unrealistic shear planes
pile row
Principle of embedded pile row
Note that in essence we
have two coupled springs
(soil and interface)!
Shaft and base also

have sliders to specify
max. capacity
Principle of embedded pile row

In axial loading the 2D soil displacement appears to be equal to
average 3D soil displacement, for all values of soil stiffness and
interface stiffness (Sluis, 2012) => fit interface stiffness to obtain
realistic load-displacement behaviour for pile row
pile row
Fitting axial stiffness
Fit with load-displacement curves of Bored piles from Dutch annex
of EUROCODE (NEN6743-1)
Stiffness based on:
Interface stiffness factor (ISF):
(Sluis, 2012)
Fitting lateral stiffness

Fit with Plaxis 3D embedded pile for various Ls/Deq for:
1. piles loaded by loads
2. piles loaded by soil displacement Lspacing = 1m
Horizontal pile displacement [m]

Disadvantages: -0.005 0 0.005 0.01 0.015
- ad.1 For large Ls/D no fit possible 0
-1
- ad.2 Soil displacements dependent on ISF -2
-3
Depth [m]
-4
-5
However using the same formula -6
for axial shaft gives reasonable fit -7
-8
for Ls/Deq 2-8 -9
-10
2D (ISF_RN = 0.5) 2D (ISF_RN = 1)
2D (ISF_RN = 10) 3D
pile row
Input
Stiffness
Weight
Select pile type (A, I)
Lspacing
Ttop / Tbottom
Fmax, base
ISF values
Stiffness of
spring =
ISF * Gsoil /
Lspacing
Input
Select Top point (= top or
bottom geometry point)
Indicate connection (Rigid,
hinged or free) for Top
point
Possibilities
Foundation piles
Anchors
Soil nails
??
A A A A A
1 4 18 20 24 25 19 2
6
46 43 44
48 50
16 47 38 45
49 51
32 39 36
8 7
12 14 17 23 28 31 41 37 13
40 34
9 11 22 27 30 42 35 10
5
15 33
y 21 26 29
0 x 3
Case description
Bridgedeck Piledabutment
Existingroad/railway Embankment
Soilstructure
interaction?
Softlayers(peat/clay)
Deepsand(foundationlayer)
pile row
3D model with emb. pile Pilerow:massivecircular
pilesD=0.54m
c.t.c.2.4m
6.5m Ls/Deq=4.4
2D model with emb. pile row / plates
pile row
Main results 1st pile row
Q2dplate[kN]
10 10
M2dplate[kNm]
Q2demb[kN]
N2dplate[kN]
M2demb[kNm]
N3D[kN] 5 N2demb[kN] 5
M_23D[kNm] N3D[kN]
Q_133D[kN] M_23D[kNm]
0 0
verticalheight[m]
Q_133D[kN]
verticalheight[m]
600 400 200 0 200 400 600 400 200 0 200 400
5 5
10 10
15 15
20 20
M/Q/N M/Q/N
Conclusions
Embedded pile row gives improved modelling possibilities
for piles in 2D;
Default setting for ISF in Plaxis based on (Sluis, 2012),
derived for a limited number of cases => validate if
default settings are applicable for your situation!
pile row
THANK YOU
Presentation by:
1. Dr William Cheang
Acknowledgment to the contributors and co-workers:

1. Dr Ronald Brinkgreve
2. Dr Lee Siew Wei
3. Prof Rajagopal
Exercise 3: Simulation of an Ground Anchored Embedded
Retaining Wall
Dr William Cheang
Tied-back excavation using HSsmall
TIED-BACK EXCAVATION
Using the HSsmall model
INTRODUCTION
A building pit was constructed in the south of the Netherlands. The pit is 15 m deep and 30 m
wide. A diaphragm wall is constructed using 100 cm diameter bored piles; the wall is anchored
by two rows of pre-stressed ground anchors. In this exercise the construction of this building
pit is simulated and the deformation and bending moments of the wall are evaluated.
The upper 40 m of the subsoil consists of a more or less homogeneous layer of medium dense
fine sand with a unit weight of 18 kN/m3 . Triaxial test data of a representative soil sample is
given in figure 2. Underneath this layer there is very stiff layer of gravel, which is not to be
included in the model. The groundwater table is very deep and does not play a role in this
analysis.
Figure 1: Geometry for tied-back excavation
Material parameters
Determination of stiffness & strength properties (sand)
In this exercise the HSsmall model is used and the model parameters for the sand layer have
been extracted from the triaxial test data (see figure 2). The HSsmall model takes into account
the stress-dependency of soil stiffness, elasto-plastic behaviour under both compression loading
and shear loading and increased stiffness in areas with very low strain levels. The soil
parameters can be found in table 1, while the determination of the soil parameters can be
found in appendix A.
Figure 2: Triaxial test data for the sand layer
Secant wall
The secant wall consists of 100cm diameter bored piles with an intermediate distance of 80cm,
hence there is a 20cm overlap of the piles. This configuration is taken this into account for the
determination of the cross sectional area (A) and moment of inertia (I) per meter out-of-plane
(see Appendix B). The concrete stiffness is Ec =2.7107 kN/m2 with a specific weight =16
kN/m3, which leads to the material parameters as given in Table 2. The determination of the
stiffness parameters can be found in Appendix A.
Table 1: Soil parameters for the HSsmall model
Parameter Symbol Sand Unit

Material model Model HSsmall
Type of behaviour Type Drained
Unsaturated weight unsat 18.0 kN/m3
Saturated weight sat 18.0 kN/m3
Drained triaxial test stiffness E ref
50 2.0104 kN/m2
Drained primary oedometer E ref
oed 2.0104 kN/m2
stiffness
Unloading/reloading stiffness E ref
ur 8.0104 kN/m2
Power for stress-dependent m 0.5
stiffness
Cohesion c0ref 1.0 kN/m2
Friction angle 0 35
Dilatancy angle 5
Small-strain shear modulus Gref
0 10.0104 kN/m2
Threshold shear strain 0.7 1.5104
0
Unloading/reloading Poissons ur default
ratio
Reference stress pref default kN/m2
Coefficient for lateral stress KN
0
C
default
under primary loading
Interface strength reduction Rinter Manual : 0.6
Coefficient for lateral initial K0 automatic
stress
Ground anchors
The anchors are made of 32mm diameter steel bars at an intermediate distance of 1m. The
steel bars have a stiffness of Es =210106 kN/m2 . The anchors have an representative capacity
of 605 kN per anchor. In combination with a secant wall the anchors may be prestressed to
a maximum level of 80% of the design capacity, which is the representative capacity divided
by a partial safety factor that has been determined at 1.5 This leads to a maximuim prestress
force of 322 kN per anchor. The maximum compression force of the anchor is not important
as the anchors will not be loaded under compression. The grout body that forms the bonded
length of the anchor behaves relatively weak under tension compared to the steel bar inside.
Therefore it is assumed that both stiffness and strength of the bonded part of the anchor are
fully determined by the steel bar. This leads to the material properties for both the anchor rod
(free length) and grout body (bonded length) as given in tables 3 and 4. Note that it is chosen
to use the representative capacity of the anchors as maximum anchor force.
It is assumed that the anchor rod is present inside the full length of the grout body, hence the
properties of the embedded pile row are based on the composite properties of anchor rod and
grout, which are considered to be mainly determine the stiffness and strength properties of
the anchor rod.
Table 2: Properties of the secant wall (plate)
Parameter Symbol Secant wall Unit

Material behaviour Material type Elastic
Isotropic Yes
End bearing No
Axial stiffness EA 20*106 kN/m
Flexural stiffness EI 1.67*106 kN/m2 /m
Weight w 15.0 kN/m/m
Poissons ratio 0.15
Furthermore, it is assumed that the shear resistance between grout body and soil is sufficient
and so the skin resistance of the grout body is set to a very high value.
Table 3: Properties of the anchor rods (node-to-node anchors)
Parameter Symbol Anchor rod Unit

Material behaviour Material type Elastoplastic
Axial stiffness EA 170*103 kN
Spacing Lspacing 1.0 m
Max. tension force |F max,tens | 605 kN
Max. compression force |F max,comp | 605 kN
Table 4: Properties of the grout bodies (embedded pile rows)
Parameter Symbol Grout body Unit

Modulus of elasticity E 2.1*108 kN/m2
Material weight 60 kN/m3
Pile type Pile type Predefined -
Predefined pile type Predefined pile type Massive circular pile -
Diameter Diameter 0.032 m
Skin resistance Ttop,max , Tbot,max 1000 kN/m
Base resistance Fmax 0 kN
Interface stiffness factor Default values -
INPUT
Project properties
For the dimensions of the project, enter xmin = 0, xmax = 100, ymin = 60 and y max = 0
in the Model tabsheet of the Project properties window, see figure 3.
Accept all other settings as their defaults and press the <OK> button.
Figure 3: Project properties, Model tabsheet
Soil mode
Select the Create borehole button ( ) and insert a borehole at (x y) = (0 0)
In the Modifiy soil layers window define a single layer in the borehole with Top = 0m and
Bottom = -60m.
The phreatic level in the project area is very deep and so the excavation is done completely
above the phreatic line. Set the Head of the borehole to -60m to indicate that the phreatic
level is at the far bottom of the borehole.
Close the Modify soil layers window.
The drawing area now shows the subsoil of 100m wide and 60m deep in grey colour.
Select the Show materials button ( ) so that the Materials window opens.
Create a soil material sets according to the parameters given in table 1. Any parameter
not specified in this table should be left at its default value.
Assign the material set to the single soil layer.
This will lead to the subsoil as given in figure 4.
Figure 4: Subsoil of the model
Structures mode
In Structures mode we will now define the structural elements such as the wall and the anchors
as well as the levels of excavation.
First we will draw the wall:
Select the Create line button ( ) and from the menu that opens select the Create plate
button ( ).
Draw the wall as a plate element from (x y) = (15 0) to (15 -25).
Soil-structure interaction is modelled with an interface element. Since there will be interaction
between soil and wall on both sides of the wall, an interface is required on both sides of the
wall:
From the Create line button menu now select the Create interface button ( ).
Draw the interface for soil-structure interaction from (x y) = (15 0) to (15 -25) and back to
(15 0) in order to get soil-structure interaction on both sides of the wall.
Now the anchors are added. The anchors are made out of two parts: a free length formed by
just the anchor rod and modelled with a node-to-node anchor, and a bonded length which is
the grout body with the anchor rod inside, modelled by an embedded pile row element:
To draw the anchor rods of the grout anchors (hence, the free length sections) select
again the Create line button, now followed by the Create node-to-node anchor button
( ).
Draw the upper anchor from (x y) = (15 -5) to (30 - 15) and right-click to end the drawing.
Now draw the lower anchors from (x y) = (15 -10) to (30 -20) and again right-click to end
the drawing.
Finally, insert the grout bodies (hence, the bonded length sections) using embedded pile
rows. To do so, select the Create embedded pile row button from the Create line button
menu.
Draw the upper grout body from (x y) = (30 -15) to (37.5 -20) and right-click to end
drawing.
Draw the lower grout body from (x y) = (30 -20) to (37.5 -25) and again right-click to end
drawing.
Embedded pile rows typically have a free moving foot and a head for which the connection
to adjacent plate elements can be specified. This connection can either be Free, Hinged or
Rigid. In order to avoid that the head of the embedded pile row is connected to the underlying
soil and thus preventing the head of the grout body to move relative to the soil, the connection
type must be set to Free:
With the <Ctrl> key pressed on the keyboard, select both embedded pile rows in the
drawing area.
In the Selection explorer change the Connection for both embedded pile rows from Rigid
to Free, see figure 5.
Figure 5: Selected embedded pile rows in Selection explorer
Hint: As interfaces can be introduced on both sides of a geometry line, one

should pay attention to the arrows on the cursor. These arrows indicate
where the program will locate the interfaces.
Hint: It is not necessary to create a geometry line before creating plates,

geogrids, embedded pile rows or anchors. When drawing a plate, geogrid
or embedded pile row, a geometry line is automatically added. Anchors do
not create corresponding geometry lines. This is not necessary since
anchors do not interact with the underlying soil.
Finally, the levels of excavation must be defined:
From the Create lines menu select the Create line option
Draw a line from (x y) = (0 -5) to (15 -5) to define the first level of excavation and right-click
to stop drawing.
Now draw the second level of excavation from (x y) = (0 -10) to (15 -10).
Finally, draw the third level of excavation from (x y) = (0 -15) to (15 -15).
Material sets
Open the Material sets window by selecting the Show materials button ( ).
In the Material sets window change the Set type to Plates and create a new plate
material set according to the parameters given in table 2.
Assign the material set to the wall. This can be done either by dragging the material set
over the different parts of the wall, or by selecting the parts of the wall wall and change
the material set in the Selection explorer.
In a similar way, create material sets for the anchors and embedded pile rows according
to the parameters given in tables 3 and 4 and assign them.
Figure 6: Full geometry
Hint: Please note that the interface doesnt extended for a short distance
underneath the beam anymore. In 2D one had to extend the interface
manually. In 2DX it will be extended automatically at the mesh definition to
overcome a singular point at the bottom of the wall.
Mesh mode
Press the Generate Mesh button ( ). In the Mesh options window that appears set
the Element distribution to Medium and press the OK button.
Press View mesh ( ) to inspect the resulting mesh as shown in figure 7.
Figure 7: Medium finite element mesh
The mesh is automatically refined around the wall and anchors and needs no further refinement.

Change to the Staged construction mode to define the construction phases for this project.
The entire construction process consists of five phases. Define the phases, as shown graphically
below.
Initial phase
For the initial phase make sure that:
all soil layers are activated (coloured)
all stuctures are deactivated (grey)
Phase 1: First excavation

In the first calculation phase, the diaphragm wall is activated and the first excavation takes
place. See figure8a.
Click on the Add phase button ( ) in the Phases explorer to add the first calculation
phase.
Select the Activate button ( ) and click on the 4 parts of the wall to activate the complete
wall. Note that this will automatically also activate the interfaces on both sides of the wall.
Also click on the soil representing the first excavation to deactivate the soil.
(a) Phase 1: Wall installation and first excavation (b) Phase 2: Installation of the top anchor
Figure 8: Construction phases 1 & 2
Phase 2: Install first anchor

In the second phase a new option is used: the prestressing of anchors.
Add a new calculation phase in the Phases explorer.
Select the Activate button ( ) and click on the grout body (the embedded pile row) of
the upper anchor to activate it.
Then click on the anchor rod (node-to-node anchor) to also activate the anchor rod.
Now that the grout anchor is active, it needs to be prestressed:
Select the Select button ( ) and select the anchor rod of the upper anchor, hence the
node-to-node anchor.
In the Selection explorer click on the option Adjust prestress that can be found as part
of the node-to-node anchor. A checkbox now appears.
Select the checkbox (so that its checked). An extra input field Fprestress now appears. Fill
in a presstress of 300 kN, see figure9.
In the geometry a black node-to-node anchor indicates that the anchor is activated. The letter
"p" indicates that a prestress force will be active in the anchor.
Figure 9: Adjusting the presstress
Phase 3, 4 and 5
Now define the remaining phases according to figures 10a, 10b and 10c.
In phase 3 excavate the second part of the excavation
In phase 4 activate the lower anchor and prestress it to 300 kN
In phase 5 excavate the remaining 3rd part.
Hint: When processing an anchor in a certain calculation phase the anchor
force will exactly match the prestress force at the end of that phase. In
following calculation phases without prestressing, the anchor force will be
influenced by the excavation process
(a) Phase 3: Second excavation stage (b) Phase 4: Installation of the lower anchor
(c) Phase 5: Final excavation stage
Figure 10: Construction phases 3, 4 and 5
Calculate
When all phases are defined, press the Calculate ( ) button. Ignore the warning to select
nodes for curves.
OUTPUT
After calculation, press the View calculation results button ( ) to inspect the results.
The results of phase 5 are presented in figure 11. After this final stage the calculated excavation
bottom heave is about 5 cm.
Figure 11: Deformed mesh (phase 5)
Select the button Select structures ( ) and double click on a node-to-node anchors.
Plaxis will present a table in which the stress in all anchors may be inspected. Anchor
forces are approximately 340 kN where the lower anchor has a slightly higher anchor
force than the upper anchor.
When double-clicking on one of the embedded pile rows the change of axial forces within
the grout body can be investigated. For both grout bodies the axial forces equals the
anchor force where its connected to the node-to-node anchors. The axial decreases
almost linearly over the length of the grout body.
By double-clicking on the wall the structural forces in the wall can be inspected. The
maximum bending moment should be in the order of 480 kNm/m (figure 12)
When double-clicking on an interface only the results of part of the interface can be
seen. In order to see the results for the whole interface chain, keep <Shift> pressed
on the keyboard while double-clicking on the interface. In figure 13 the left side are the
passive earth pressures and the right side are the active earth pressures. It can be seen
that only a small part of the maximum passive earth pressures has been mobilized at
this stage.
Figure 12: Bending moments in the Figure 13: Effective normal stresses in the
secant wall interface
Geometry size
For any project the geometry has to be made sufficiently large so that the boudary conditions
have no influence on the calculation results. This means in practice that close to the boundaries
(with exception of a axis of symmetry) displacements should be small and stresses should be
undisturbed. When using the HSsmall model there is an interesting plot that can be used to
check this.
From the Stresses menu choose the option State parameters and then G/Gur .
This plot shows the actual shear stiffness divided by the unloading/reloading shear stiffnes
at engineering strain level. For areas with very small deformations the stiffness will be high
(small strain stiffness) and so the value of G/Gur > 1. Hence, the geometry is sufficiently large
if next to the boundaries, with exception of the axis of symmetry, G/Gur > 1, which indeed is
the case.
Hint: State parameters are additional quantities that relate to the state of the
material in the current calculation step, taking into account the stress
history. Examples of state parameters are the isotropic overconsolidation
pressure (pp ) and the hardening parameter p that specifies the maximum
shear strain level reach in the stress history.
Surface settlements
In Plaxis Output it is possible to see calculation results in a user-defined cross section. This
feature will be used to check the surface settlements behind the secant wall.
Click the Cross section button . The Cross section points window appears, see figure
14.
It is possible to draw a cross section by hand and check in the Cross section points window
what the coordinates are of the start and end point of the cross section. However, it is also
possible to position the cross section at a specific location by defining the coordinates of the
start and end point manually.
Move the mouse to the Cross section points window and fill in the coordinates (15, -0.1)
for the first point and (100, -0.1) for the second point and press OK. This will create a
cross section from the right boundary of the model until the secant wall just below the
soil surface. The cross section will open in a new window.
From the Deformations menu select Total displacements and then u y to see the vertical
displacements of the soil surface. The maximum settlement is about 12 mm, see figure 15.
Figure 14: Cross section points window
Figure 15: Vertical displacements behind the secant wall
APPENDIX A: DETERMINATION OF SOIL PARAMETERS FROM

TRIAXIAL TEST
Figure 16: Triaxial test for sand layer
Strength parameters
Fill in 1 and 3 in the Mohr-Coulomb criteria:
1 3 = (1 + 3 ) sin + 2c cos
Since the cohesion will be small, assume c = 0:
1 3
1 +3
= sin
370100
370+100
= sin
= 35o
= 30 = 5o
For reasons of numerical stability, use c = 1 kPa
Stiffness parameters
The triaxial test has a cell pressure 3 = 100 kPa. This corresponds with reference pressure,
so E50 = Eref
50 .
ref v 135
E50 = v
= 0.675%
= 2.0 104 kP a
For Sand it can be assumed that

ref ref
Eoed = E50 = 2.0 104 kP a
ref ref
Eur 4 E50 = 8.0 104 kP a
m = 0.5
Additionally it is assumed that:
Gref
0
ref
= 1.25 Eur = 1 105 kP a
0.7 = 1.5 104
APPENDIX B: MATERIAL PROPERTIES SECANT WALL

For a plane strain model material properties for the secant wall have to specified per meter
length of the wall. In order to do so we first recognize the secant wall as consisting of repetitive
parts at a certain intermediate distance, as shown in Figure 17.
Figure 17: Secant wall as repetitive equal sections
Compared to the original bored piles the repetitive sections have a reduced cross sectional
area. Though it can be analytically derived how much the reduction is, the fastest way to
determine this is to draw the repetitive section on paper with a fine grid based on the original
bored piles with a diameter of 1000mm and an overlap of 200mm and count squares. Using
this method the cross sectional area of the repetitive section is determined as As = 0.74 m2 .
Since the sections are at a distance D apart where D is given as 800mm, the cross sectional
area of the wall per meter is given as:
Awall = ADs = 0.74
0.8
= 0.93 m2 /m
For the moment of inertia is assumed that the influence of the reduced cross sectional area
is negligble as the reduction is close to the axis of bending and symmetric. Therefore the
moment of inertia per meter wall is determined as:
Ipile 4 (0.5)4
Iwall = D
= r
4D
= 40.8
= 61.3 103 m4 /m
With Econcrete = 2.7 107 kN/m2 this gives
EA = (2.7107 )(0.93) = 2.5107 kN/m

EI = (2.7107 )(61.3103 ) = 1.67106 kNm2 /m
And for the weight:
w = A = 16 0.93 = 15 kN/m/m
Modelling of Groundwater in PLAXIS
Dr William Cheang
Modelling of Groundwater in PLAXIS AE
CONTENTS
A. Introduction
1. Groundwater in Geotechnical Engineering
2. Plaxis
B. Definitions Pore Pressures in Plaxis
1. Active
2. Water
3. Steady-state
4. Excess
C. Generation of Porewater Pressures in Plaxis
1. Porewater Pressure due to Hydrostatic Condition (Phreactic Control)
2. Pore Pressures due to Groundwater Flow (Steady or Transient States)
3. Pore Pressures due to Consolidation (Steady State Only plus Excess PWP)
4. Pore Pressures due to Fully Coupled Consolidation (Transient Only)
D. Hydraulic models
1. Fully Saturated Soils
2. Partially Saturated Soils
E. Case Histories
1. Excavations
F. References
22/10/2014 2
Modelling of Groundwater in Plaxis
A. INTRODUCTION
22/10/2014 3
Groundwater Analysis
A. Geotechnical problems are related to groundwater
B. Two extreme conditions of porewater response are normally considered,
they are:
1. Drained
2. Undrained (Method A, B & C)
C. Real soil behaviour is related to time , i.e. transient, with the porewater
pressure being dependent on imposed:
1. Permeability
2. Rate of loading
3. Hydraulic (Flow) boundary conditions
D. The interstitial voids of the soil skeleton can be fully or partially filled with
pore fluid and therefore effective stresses are influenced by this action
E. This lecture will look into the following issues:
1. The setup of pore pressures in Plaxis
2. Input parameters
3. Some examples of groundwater regimes
22/10/2014 4
B. DEFINITIONS OF PORE PRESSURES IN
PLAXIS AE
22/10/2014 5
A. DEFINITIONS AND MODES
A. Definition of Pore Water Pressure Terms in Plaxis
1. Pactive - Active State Pore Pressures (idea of Total Pore Water Pressure)
2. Pwater - Pore Water Pressures (Psteady + Pexcess)
3. Psteady - Steady State Pore Pressures (Background)
4. Pexcess - Excess Pore Pressures
B. Calculation Modes (AE)
1. Ground Water Flow Only
2. Consolidation (based on Pexcess ONLY)
3. Fully Coupled Flow-Deformation Consolidation (based on Total PWP ie Pactive)
22/10/2014 6
A1.ACTIVE PORE PRESSURE Pactive
22/10/2014 7
A2. PORE WATER PRESSURE Pwater
22/10/2014 8
A3.STEADY-STATE AND EXCESS PORE PRESSURE
Plastic Drained or Undrained Analysis
Consolidation process is dissipation of Pactive or Pexcess to Psteady

values in the long-term condition
22/10/2014 9
A4. SATURATION AND SUCTION
Include Steady and Excess components
This is know as the apparent cohesion

Similar in concept to Prof Fredlunds strength model Ca = (ua-uw)*tan b =Seff *tan
22/10/2014 10
A5. TERZAGHI OR BISHOP EFFECTIVE STRESS
Terzaghi Effective Stress (only for Saturated Soil with pressure or suction)
Bishop Effective Stress (only for Saturated/Unsaturated Soil with pressure or suction)
22/10/2014 11
A5. EFFECTS OF IGNORE SUCTION
22/10/2014 12
B. CALCULATION MODES
22/10/2014 13
B1. GROUNDWATER FLOW ONLY MODE
Groundwater Flow mode:

Similar to PlaxFlow but with huge improvements in the kernel (see Galavi, 2010)
All functionalities of PlaxFlow rewritten in PLAXIS code (new)
Steady state groundwater flow
Transient groundwater flow (ONLY POSSIBLE IN THIS MODE)
All types of boundary conditions
New features in wells and drains
Faster calculation (new)
22/10/2014 14
B2. PLASTIC OR CONSOLIDATION (OLD CLASSICAL)
A. Steady-state pore pressures
1. Phreatic lines
2. Steady-state groundwater flow
analysis
B. Excess pore pressures

1. Undrained material type in
combination with Plastic
calculation, combine with
2. Consolidation analysis
INPUT KERNEL
Steady State Excess Porewater Deformation

Note:
Note:
1. Undrained Analysis
1. Hydrostatic
2. Consolidation Analysis
2. GWF calculation
Active Porewater = Steady-state + Excess

22/10/2014 15
B3. FULLY COUPLED FLOW-DEFORMATION MODE

(OLD ADVANCED MODE)
A. Consolidation analysis
B. Transient groundwater flow analysis is
implied in the calculations
KERNEL
Active Porewater Deformation

Note:
1. Undrained Analysis
2. Consolidation Analysis
22/10/2014 16
B. GENERATION OF POREWATER
PRESSURES
22/10/2014 17
GENERATION OF STEADY-STATE PORE PRESSURES
Steady-state (Background) pore pressures can be generated by:
1.Phreatic and Cluster Approach (Hydrostatic)
2.Groundwater Flow Analysis
22/10/2014 18
PHREATIC AND CLUSTER APPROACH
A. Phreatic Level
1. General Phreatic Level
2. Cluster Phreatic Level

B. Cluster Pore Pressure Distribution
1. Interpolation (Adjacent to clusters or phreatic lines)
2. Cluster Dry
3. User-defined Pore Pressure Distribution
22/10/2014 19
C1. PHREATIC LINE
a) Porewater pressures are hydrostatic

b) Calculated based on gamma-water * height of the water column
c) Simple situations (water-table is horizontal)
d) No flow
e) For cases, i.e. simple excavations, foundations or embankments
22/10/2014 Hydrostatic steady-state pressure along interface

20
GENERAL PHREATIC LINE
22/10/2014 21
GENERAL & CLUSTER PHREATIC LINES
22/10/2014 22
Cluster: Dry
22/10/2014 23
Cluster: User-defined Pore Pressure Distribution
22/10/2014 24
Combination: Phreatic and Cluster Options
22/10/2014 25
COMBINATION: PHREATIC & CLUSTER OPTIONS
22/10/2014 26
CLUSTER DRY + INTERPOLATION COMBO
Cluster Dry
Interpolated
Case Histories
22/10/2014 27
GROUNDWATER FLOW ANALYSIS

A. Steady-state Pore Pressure Generation based on Groundwater Flow Calculation
B. Influence by:
1. Soil Permeability
2. Boundary Conditions (External or Internal)
C. Phreatic line is calculated for

1. Confined flow problems
2. Unconfined flow problems
D. Steady-state groundwater flow:

1. No change in flow field with time
2. Position of phreatic is fixed (influence by k and geometry of hydraulic passage)
3. Long-term flow field condition
E. Transient-state groundwater flow (ONLY possible in Groundwater Flow Mode):

1. Flow field influence by time
2. Position of phreatic line changing with time
3. Applicable to problems where pore pressure and hydraulic boundaries are changing with time.
22/10/2014 28
GROUNDWATER FLOW: STEADY STATE
Calculation based on setup of:
1. Boundary conditions:
a) Prescribed water levels (constant)
b) Closed flow boundaries (bottom, axis of symmetry)
c) Wells and drains (constant)
d) Interface elements (on=impermeable, off=permeable)
e) Inflow / outflow (constant)
2. Soil permeabilities
3. Phreatic level in the soil is being calculated for t=
22/10/2014 29
GROUNDWATER FLOW: TRANSIENT-STATE
Transient groundwater flow:

1. Boundary conditions
a) Prescribed water levels (changing with time)
b) Closed flow boundaries (bottom, axis of symmetry)
c) Wells and drains (changing with time)
d) Interface elements (on=impermeable, off=permeable)
e) Inflow / outflow (changing with time)
2. Flow field changes in time:

a) Constantly changing natural water conditions
b) Relatively fast building process, pumping, wells
3. Embankments with river changes, tidal change
4. Reservoir impoundment and drawdown
5. Precipitation problems
22/10/2014 30
Pore pressures steady-state
Steady-state flow:
Water conditions
settings
Create global water level
Closed Flow
RESULTS FLOW FIELD RESULTS GROUNDWATER HEAD
22/10/2014 31
STEADY STATE ARTESIAN CONDITION
WATER CONDITIONS
SETTINGS: Set Cluster Water Level
SAND
CLAY (INTERPOLATE BETWEEN

CLUSTERS PORE PRESSURES)
SAND
Set Cluster Water Level
RESULTS FLOW FIELD CLOSED FLOW
RESULTS GROUNDWATER HEAD
22/10/2014 32
SOME POINTS: FINITE ELEMENT MODELING
1. GWF calculation generally needs finer mesh than deformation analysis

2. GWF calculation generally needs large number of steps than deformation
analysis
3. GWF calculation usually converges, but can be problematic when:
a) Mesh is too coarseness
b) Elements are distorted
c) Large differences in permeabilities
22/10/2014 33
SOME POINTS- FINITE ELEMENT MODELING
1. Qualitative evaluation:
Flow field
Location of phreatic line
2. Quantitative evaluation:
Heads, pore pressures compared to hydrostatic,
Compare with measurements or field experience
22/10/2014 34
D. HYDRAULIC MODELS IN PLAXIS
22/10/2014 35
Groundwater flow flow in unsaturated soil
Water content and permeability in unsaturated zone
k k rel k sat , k rel f h p , S

(h p )
S (h)
n
hp= => pressure head
22/10/2014 36
A. Linear Model
krel
1
hp = -
hp = 0
m
0 hp

1 hp 0 Saturated For numerical stability

k rel 1 m h p 0 hp Partially saturated
hp 1 Ae
Dry
3 N int
22/10/2014 37

Soil Water Characteristic Curve (SWRC or RC)
22/10/2014 38
van Genuchten model
1 g n
S ( h p ) S res ( Ssat Sres ) 1 g a h p

gn ( g )
n
2
g n 1
g n g n

gl

krel S Se 1 1 Se g n 1

with S Sres
Se
Ssat Sres
Ssat,Sres: saturated and residual saturation

ga, gn and gl: curve fitting parameters
22/10/2014 39
Approximate van Genuchten model
1 if hp 0
Linear in Saturation

hp
S hp 1 if h ps h p 0
hps
0 if h p h ps

1 if hp 0
4h p Log-linear in Permeability
h
krel h p 10 pk if h pk h p 0
4
10 if h p h pk

hps: length of partially saturated zone under hydrostatic conditions
hpk: pressure head at krel=10-4

22/10/2014 40
Groundwater flow - material data sets
A. Parameters:
1. Permeabilities (kx, ky)
2. Void ratio (to calculate storage)
3. Elastic storage coefficient
(The volume of water that a unit volume of saturated soil loses due to
a unit reduction in the applied water head)
4. Maximum unsaturated zone height
B. Soil classification
1. Particle fractions
2. Predefined series (Staring, Hypres, USDA) with Van Genuchten and
Approx. van Genuchten parameters.
3. User-defined
22/10/2014 41
Soil classification - Staring

Dutch soil classification system
18 upper soils data sets
18 lower soil data sets
Upper soils:
< 1m below soil surface
Lower soils:
all deeper soils
22/10/2014 42
Soil classification: Hypres
Hydraulic Properties of
European Soils
Particle distribution:
< 2m
2m - 50m
50m 2mm
5 upper soils data sets

5 lower soil data sets
1 organic soil data set
22/10/2014 43
Soil classification: USDA

United States Department
of Agriculture
Particle distribution:
< 2m
2m - 50m
50m 2mm
12 soils data sets
No difference between
upper and lower soils
22/10/2014 44
Soil classification and Van Genuchten parameters
Relative permeability
Degree of saturation
22/10/2014 45

E. EXAMPLES
22/10/2014 46
Modeling of
Ground Water in Excavation Analysis
22/10/2014 47
Effects of GWT on Excavation Analysis

For PLAXIS FEM Program:
Steady State GWT Calculation is a separate program from Excess Pore
Pressure and Consolidation Calculation
This can lead to many different ways to include Effects of GWT on Excavation
Analysis
The GWT or Phreatic Surface can be determined by either
Method A Steady State Flow calculation (Prefered Method)
Method B User Defined Phreatic Surface, ie head is constant on a
vertical section (to model hydrostatic pressure on both sides of excavation)
22/10/2014 48
Possible GWT
Conditions in
Excavations
2ba
uC w
2b a
(2b c)a
uG w
2b c a
49
PLAXIS Model of Full GWT
h=Ha (const)
h=Hb(const)
Ha
Hb
CLOSED FLOW Boundary
22/10/2014 50
PLAXIS Model of GWT Drawdown
Phreatic surface, PP=0
GWT drawdown
h=Ha(const)
h=Hb(const)
Ha
Hb
CLOSED FLOW Boundary
22/10/2014 51
PLAXIS Model of Hydrostatic GWT
h=Ha(const)
Ha
h=Hb(const)
Hb
Hydrostatic both sides but PP not in Equilibrium

This may give problems as there are incorrect
effective stresses in the mesh
22/10/2014 52
Seepage and Excavations
A. GWT lowering by Steady State Seepage
B. GWT lowering by Transient Seepage
22/10/2014 53
GWT lowering SS Seepage
Excavate 5m, k=1e-5 m/s Excavate 10m, k=1e-5 m/s
Lower 1.3m
Lower 3.0m
Excavate 15m, k=1e-5 m/s GWT is nearly proportional to

excavation depth
Lower 5.6m
22/10/2014 54
GWT lowering SS Seepage
Lower 5.6m
Lower 5.6m
For SS case, GWT is not

Excavate 15m, k=1e-9 m/s dependent on k.
Lower 5.6m Pattern of GW heads is

function of geometry only and
soil layer arrangements
22/10/2014 55
GWT and Transient Seepage
Lower 1.3m
Lower 0.8m
Excavate 5m in 30 days.
Excavate 5m, k=1e-9 m/s
Sands, k=1e-5 m/s is like SS
Lower 0.3m case
Clays, k=1e-9 m/s very little
GWT lowered
22/10/2014 56
Lower 3.0m
Lower 1.8m
Excavate next 5m in 30 days.

Lower 0.3m case
GWT lowered
22/10/2014 57
Lower 5.6m
Lower 3.6m
Excavate next 5m in 30 days.

Lower 0.3m case
GWT lowered
22/10/2014 58
Modeling of
Water drawdown in Dam Analysis
22/10/2014 59
Rapid Drawdown Example Fully Coupled Consolidation

and Time Dependent Boundary Conditions
A and B are Head BC drawdown from H=25m to

H=5m in 50 days (SLOW CASE)
C is Free Seepage BC drawdown from H=25m to
H=5m in 50 days (SLOW CASE)
Water Conditions Settings
A
C
B FILL CORE
SUBSOIL
CLOSED FLOW BDY
10/22/2014 60
Rapid Drawdown Example
Time Dependent Boundary Conditions
From H=25m to H=5m in 50 days (Slow Case)
H=25m
H=5m
10/22/2014 61

H=25m
H=5m
10/22/2014 62
Potential Slip Surface by c/phi reduction for the

Case of Slow DD in 50 days FOS=1.477
10/22/2014 63

WL at 25m FOS=1.752
WL at 5m Very Slow DD FOS=1.671
WL at 5m Slow DD in 50 days FOS=1.477
Beware of unwanted suction; better to

switch off suction in design (safer)
WL at 5m rapid DD in 5 days FOS=0.969 (FAILURE!!)
10/22/2014 64
Pore Water Pressure and Saturation
Rapid Case H=25m to 5m in 5 days
Slow Case H=25m to 5m in 50 days
10/22/2014 65
Reservoir Drawdown (Flow Field)

Steady State at High Reservoir
Rapid Drawdown H=25 to 5m in 5 days
Slow Drawdown H=25 to 5m in 50 days
10/22/2014 66
END
22/10/2014 67
Drained and Undrained Analysis
CG11: DRAINED AND UNDRAINED ANALYSIS
Antonio Gens

Ronald Brinkgreve (Plaxis bv / Delft University of Technology)
outline
drained / undrained (conditions and analysis)

drained / undrained soil behaviour
typical results from drained and undrained triaxial tests
strength parameters in effective stresses and total stresses
what is the critical case: drained or undrained?
modelling undrained behaviour with Plaxis
general procedure
three methods
Method A: effective stresses
Method C: total stresses
Method B (effective stress, hybrid method)
undrained shear strength (for Method A)
undrained behaviour with Mohr-Coulomb Model
undrained behaviour with advanced models
influence of dilatancy
summary
drained / undrained (conditions and analysis)
in undrained conditions, no water movement takes place and,

therefore, excess pore pressures are built up
u 0, '
in drained conditions, no excess pore pressures are built up

u = 0, = '
drained analysis appropriate when

permeability is high
rate of loading is low
short term behavior is not of interest for problem considered
undrained analysis appropriate when

permeability is low and rate of loading is high
short term behavior has to be assessed
triaxial test (NC soils) drained / undrained soil behaviour
drained undrained
triaxial test (OC soils) drained / undrained soil behaviour
drained undrained
stress paths in undrained triaxial test NC / OC
1 3
t
2
3
s' 1
2
3
s 1
2
Undrained soil behaviour
Implications of undrained soil behaviour

excess pore pressures are generated
no volume change
in fact small volumetric strains develop because a finite (but high) bulk
modulus of water is introduced in the finite element formulation
undrained shear strength is the critical parameter controlling
stability
Strength parameters
Mohr-Coulomb parameters in terms of effective stress

c tan

in terms of principal effective stresses
1 3 3 c c
1 sin ; t s sin
2 2 tan tan
Strength parameters
Mohr-Coulomb parameters in terms of principal effective stress
tan = sin
t
1 3
2 a = ccos
s'
'1 '3 , s 1 3
2 2
1 3 3 c c
2
1
2

tan sin ; t s
tan sin ; t s sin c cos

Strength parameters
Mohr-Coulomb parameters in terms of total stresses
Only undrained conditions!
c tan total stresses
Cu
1 3 1 3
2
F
2
F cu ,
-Cu
Effective stresses
Soil behaves as if it was cohesive

cu ( su ) : undrained shear strength
cu only changesPlaxis
if drainage occurs (no change if undrained conditions prevail)
Advanced Course, New Delhi, India (29-31 October 2014) 256/448
What is the critical case: drained or undrained?
t
OC
OC
NC
NC
Loading
Unloading
s, s
note that for soils in general:

factor of safety against failure is genarallylower for
short term (undrained) conditions for loading
problems (e.g. embankment)
factor of safety against failure is generally lower
for long term (drained) conditions for unloading
problems (e.g. excavations)
however
What is the critical case: drained or undrained?

t
OC
OC
NC
NC
Loading
Unloading
s, s
For very soft NC soil, factor of safety against failure may

be in some cases lower for short term (undrained)
conditions for unloading problems (e.g. excavations)
For very stiff OC soil, factor of safety against failure may
be in some cases lower for short term (undrained)
conditions for loading problems (e.g. embankment)
FE modeling of undrained behavior
undrained analysis in terms of effective stress (Plaxis 2010)
type of material behaviour: undrained (Plaxis < 2010)
the issue:
constitutive equations are formulated in terms of
' D'
we need to compute D
D
principle of effective stress ' f

with f pw pw pw 0 0 0
T
since the strains are the same in each phase,

' D '
13 03 Kf
f D f D f Ke Ke
03 03 n
pore fluid stiffness, related to the
bulk modulus of pore fluid (water) Kf
We need D D
' f D D ' D f ( D ' D f )
D D' D f
Example: linear elastic model + plane strain
D D' D f K
E
G
E
3(1 2 ) 2(1 )
4 2 2 G G'
K ' 3 G K ' G
3
K ' G 0
3
&xx &xx
e
& K ' 2 G 4 2
K ' G K ' G 0 &eyy
yy 3 3 3 D
&zz &zz
e
K ' 2 G 2
K ' G
4
K ' G 0 &e
&xy 3 3 3 xy
0 0 0 G

4 2 2
K 3 G K G
3
K G 0
3
&xx &xx
e
& K 2 G 4 2
yy K G K G 0 &eyy

&zz
3 3 3
&zz
e D
K 2 G 2
K G
4
K G 0 &e
&xy 3 3 3 xy
0 0 0 G


D D' D f
4 2 2
K ' 3 G K ' G
3
K ' G 0
3
Ke Ke Ke 0
K ' 2 G K ' G
4 2
K ' G 0
K Ke Ke 0
D 3 3 3 e
Ke Ke Ke 0
K ' 2 G 2
K ' G
4
K ' G 0
3 3 3 0 0 0 0
0 0 0 G

4 2 2
K ' 3 G K e K ' G K e
3
K ' G K e
3
0

K ' 2 G K 4
K ' G K e
2
K ' G K e 0
D 3
e
3 3

K ' 2 G K e 2
K ' G K e
4
K ' G K e 0
3 3 3
0 0 0 G

D D' D f
4 2 2
K ' 3 G K e K ' G K e
3
K ' G K e
3
0

K ' 2 G K 4
K ' G K e
2
K ' G K e 0
D D ' De 3
e
3 3

K ' 2 G K e 2
K ' G K e
4
K ' G K e 0
3 3 3
0 0 0 G

4 2 2
K 3 G K G K G 0

3 3
4 4
K 2 G 4
K G
2
K G 0
K ' G K e K G
D 3 3 3 3 3

K 2 G 2
K G
4
K G 0
3 3 3
0
0 0 G
K K ' K e
all the above (which is valid for any soil (or model) for which the
principle of effective stress applies) can be easily combined
with the FEM
instead of specifying the components of D, specify D' and Ke

D D' D f
when calculating stresses,
f K e v
' f
' D'
a value must be set for Ke

real value of Ke = Kw/n ~ 1106 kPa (for n = 0.5)
in fact, the pore-fluid is assigned a bulk modulus that is
substantially larger than that of the soil skeleton (which
ensures that during undrained loading the volumetric strains
are very small)
PLAXIS automatically adds stiffness of water when undrained

material type is chosen using the following approximation:
Kw Eu 2 G 1 u
K total K'
n 31 2 u 31 2 u
E' 1 u
K total assuming u = 0.495
31 2 u 1 '
Note: this procedure gives reasonable results only for ' < 0.35 !
modeling undrained behavior with PLAXIS
method A (analysis in terms of effective stresses):

effective strength parameters (MC: c', ', )
effective stiffness parameters (MC: E50', )
method B (analysis in terms of effective stresses):

total strength parameters c = cu, = 0, = 0
effective stiffness parameters E50', '
method C (analysis in terms of total stresses):

total stiffness parameters Eu, u = 0.495
FE modeling of undrained behavior (method A)
analysis in terms of effective stress
type of material behaviour: undrained method A (version 2010)
undrained (version < 2010)
u changes (excess pore water pressures generated)
' D '
In the case of Mohr Coulomb model:
effective strength parameters c, ,
the total stiffness matrix is computed as: D D' D f
available for all models (including user defined models)!
FE modeling of undrained behavior (method A)

t
uf
u
TSP
ESP
cu
s, s
single set of parameters in terms of effective stress (undrained,

drained, consolidation analysis consistent)
realistic prediction of pore pressures (if model is appropriate)
the undrained analysis can be followed by a consolidation
analysis (correct pore pressures, correct drained parameters
and Cu is updated automatically)
Cu is a consequence of the model, not an input parameter!!



FE modeling of undrained behavior (method C)

analysis in terms of total stress
type of material behaviour: undrained method C (version 2010)
drained (version < 2010) (in spite of
modelling an undrained situation)
u does not change
available for Linear elastic, Mohr Coulomb, NGI-ADP models
FE modeling of undrained behavior (method C)
t
TSP=ESP
cu
s, s
parameters in terms of total stress

no prediction of pore pressures (only total stresses
are obtained)
the undrained analysis can not be followed by a
consolidation analysis
Cu is an input parameter!!



FE modeling of undrained behavior (method B)
analysis in terms of effective stress
type of material behaviour: undrained method B (version 2010)
undrained (version < 2010)
u changes
constitutive equations are formulated in terms of (but strength
in total stresses!)
' D '
D
D D' D f
3 E
Resulting undrained stiffness parameters Eu ; u 0.495
2 1
available for Mohr Coulomb, HSM, HSSM, NGI-ADP models
FE modeling of undrained behavior (method B)

t
ESP TSP
cu
s, s
parameters in terms of total stress and effective stress

prediction of pore pressures (generally unrealistic)
the undrained analysis should not be followed by a
consolidation analysis (pore pressures unrealistic and
Cu is not updated automatically)
Cu is an input parameter!!
Undrained shear strength (for Method A)
In method A, Cu is a consequence of the model, not an input parameter!!
Therefore, it is necessary to know what value of Cu we are actually using
In the case of the Mohr-Coulomb model (in plane strain), it is easy to compute
Cu analytically
For plane strain: the undrained effective stress path rises vertically
1
2 0 '2 ' ( '1 '3 ) ; '2 ' ('1 '3 )
E
Linear Elasticity
p ' 1 1
v 0 p ' 0 p ' '1 '2 '3 '1 '3 (1 ' ) 0
K' 3 3
1
'1 '3 s' 0 tan = sin
2
t
1 3
2
Effective Stress
a = ccos
Path, ESP
A to , so s'
'1 '3 , s 1 3
2 2
Undrained shear strength from the Mohr Coulomb model
Plane strain: effective stress path rises vertically
tan = sin
t
1 3 Effective Stress
2
Path, ESP B
a = ccos
cu
A
to , so
s'
'1 '3 , s 1 3
2 2
1
cu c 'cos ' so 'sin ' c 'cos ' vo ho sin '
2
cu c 'cos ' 1
1 K 0 sin ' , K 0 ho
'v 0 'v 0 2 vo
Undrained shear strength from the Mohr Coulomb model
The Mohr Coulomb model in terms of effective stresses

OVERESTIMATES the undrained shear strength of soft clays!
cuMC
cu real
s, s
Undrained shear strength (for Method A)

Advanced models can provide better values of undrained shear
strength (but check!)
q
ine
el
advanced i lur
f a
models
advanced
models
elastic-perfectly
cu,3
2c u,3 plastic models
cu,2
c2cu,1u,1 2c u,2
pc p
Results from undrained triaxial tests using simple and advanced constitutive models
Undrained shear strength from advanced models
Although it is possible, in a few simple cases, to obtain an analytical

expression for cu, it is advisable to perform a numerical laboratory
test to check the value of undrained shear strength actually supplied
by the model
It is important to perform the numerical laboratory test under the
same condition as in the analysis
Plane strain, triaxial, simple shear
Correct initial stresses
Compression, extension, simple shear
Not all cu values are achievable with a particular model
Soft soil model
Parameters
c ' 0.1 kPa ' 23 K 0NC 1 sin ' 0.609 ur 0.15 * 0.11 * 0.0275
140.00
140.00
120.00
120.00
100.00 100.00
1-3 (kPa)
80.00 80.00
1-3 (kPa)
60.00 60.00
40.00
40.00
20.00
20.00
0.00
0 20 40 60 80 100 120 140 160 0.00
0.00 20.00 40.00 60.00 80.00 100.00 120.00 140.00 160.00
p' (kPa)
p' (kPa)
Triaxial (compression) Triaxial (extension)
cu/v=0.279 cu/v=0.214
Soft soil model
Parameters
c ' 0.1 kPa ' 23 K 0NC 1 sin ' 0.609 ur 0.15 * 0.11 * 0.0275
140.00
120.00
100.00
80.00
1-3 (kPa)
60.00
40.00
20.00
0.00
0 20 40 60 80 100 120 140 160
p' (kPa)
Plane strain (compression) Plane strain (extension)
cu/v=0.279 cu/v=0.277
influence of dilatancy on undrained shear strength
if we set 0 then, negative volumetric plastic deformations

occur at failure:
v ve vp (elastic-plastic behavior)
v 0 (undrained conditions)
vp 0 ve 0 p ' K ' ve 0
At failure: q M p ' q 0
t s sin t 0
result: unlimited increase of q (or t), i.e. infinite strength!!
Therefore, in undrained analysis, dilatancy, , must be set to zero!

influence of constitutive model and dilatancy
simulation of undrained triaxial compression test MC / HS model - q vs 1
300
275
250
225
200
175
q [kN/m ]
2
150
125
100
75
MC non dil
50 MC dil
HS_1 non dil
25 HS_1 dil
0
0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 2.25 2.50 2.75 3.00
1 [%]
influence of constitutive model and dilatancy
simulation of undrained triaxial compression test MC / HS model - q vs p'
300
275 MC non dil

MC dil
250 HS_1 non dil
HS_1 dil
225 total stress path
200
175
q [kN/m ]
2
150
125
100
75
50
25
0
0.00 25.00 50.00 75.00 100.00 125.00 150.00 175.00 200.00 225.00 250.00
2
p' [kN/m ]
summary
FEM analysis of undrained conditions can be performed in effective
stresses and with effective stiffness and strength parameters
(Method A)
Method A must be used:
if consolidation/long term analysis are required
advanced soil models are adopted
undrained shear strength is a result of the constitutive model
care must be taken with the choice of the value for dilatancy angle
Methods B and C provide alternative ways to analyze undrained
problems but:
the constituive model dos not generally represent the true soil
behaviour (before failure)
potentially useful for stability problems in undrained conditions
(specification of undrained shear strength is straightforward)
Consolidation Analysis
CG12: CONSOLIDATION ANALYSIS
Antonio Gens

CONSOLIDATION: OUTLINE
Introduction
Theory of consolidation
Permeability
FEM for consolidation analysis
Calculation modes
Final remarks
TYPES OF ANALYSIS
Drained
Loading/Construction/ excavation: very slow (in relation to the
soil permeability)
Undrained
Loading/Construction/ excavation: very fast (in relation to the
soil permeability)
Intermediate cases: consolidation analysis

Both mechanical and hydraulic (flow) problems interact
More complex computations: coupled analysis
EXAMPLE
Excess pore
water pressure
Consolidation
Initial state Undrained loading Final state
OTHER EXAMPLES
1973 1984
Ekofisk tank
OTHER EXAMPLES
Construction at intermediate rates
Change of hydraulic conditions
What is the net outflow/inflow of water in a elementary volume of soil?
qx q y
div q q
x y
qx q y qz
In three dimensions: div q q
x y z
In a saturated soil the net outflow/inflow of water must be equal to:

change of porosity
volume change due to water compression (very small but
sometimes important)
we assume solid grains incompressible
Considering:
fully saturated soil, incompressible particles
Net outflow/inflow of water per time increment: qt
Must be equal to:

n
1. Change of porosity:
t
n
2. Water compression: pw
Kw
n n n n
i.e. qt pw ; qt pw
t K w t Kw
v = volumetric strain Kw = bulk stiffness of water

pw = pore pressure change q = pore water flow
n = porosity t = time increment
Equation of continuity (with soil deformation and water compressibility)

It expresses the conservation of water mass
n n
v qt pw
t Kw
change of n depends on changes of effective stresses (flow

problem depends on the mechanical problem)
change of pore pressures (effective stresses) is derived from the
flow problem (mechanical problem depends on the flow problem)
mechanical and flow problem interact and must be solved
simultaneously
COUPLED PROBLEM!
Pore water flow:
Darcys law (isotropic permeability):

k k = permeability
q pw w = unit weight of water
w
Total change of volumetric strain in time,
considering homogeneous permeability:
n n k n
v qt pw pw t pw
t Kw w K w
v k n pw
2 pw
t w K w t
v n pw k 2
pw
t K w t w
General 3D case:
' v '
D mT D mT 1,1,1, 0, 0, 0
t t t t
Assuming isotropic elasticity!:
v 1 p ' 1 ( p pw ) 1 p 1 pw

t K ' t K ' t K ' t K ' t
p = mean total stress
1 p 1 n pw k p = effective mean stress
2 pw
K ' t K ' K w t w
E'
where K' = bulk stiffness of soil skeleton
3 1 2 '
kK' pw p
Considering incompressible water: 2 pw
w t t
1D consolidation:
v 1 ' 1 ( pw )
2H
t Eoed t Eoed t
1 1 n pw k 2
pw
Eoed t Eoed K w t w
(1 ') E '
where Eoed (1 ')(1 2 ') = constrained modulus of soil skeleton
k Eoed pw
Considering incompressible water: 2 pw
w t t
Coefficient of consolidation:
k Eoed L F L3 L2
Cv units: (m2 / s)
w 2
T L F T
Note: Cv is mainly controlled by permeability and is less variable with respect to
stress level than individual components (they vary in opposite directions)

1D consolidation, considering a constant total stress : 0
t
pw k Eoed
cv 2 pw where cv = consolidation coefficient =
w
t
cv t
T
H2

1D consolidation, considering a constant total stress : 0
t
pw k Eoed
cv 2 pw where cv = consolidation coefficient =
w
t
cv t
T
H2
PERMEABILITY
Dependence on grain (pore) size
Soil k (cm/s)
Clean gravel >1
Clean sand 1 - 10-2
(coarse)
Sand mixture 10-2 - 5x10-3
Fine sand 5x10-2 -10-3
Silty sand 2x10-3 -10-4
Silt 5x10-3 -10-5
Clay 10-6 and less
Harr (1962)
Permeability often anisotropic

q x k x q y k y
x y
PERMEABILITY
Dependence on void ratio
PERMEABILITY
PLAXIS allows consideration of change of permeability with void ratio
k e
log
k0 ck
Default value for ck is 1015
Use realistic ck only with advanced models
There may be large contrasts of permeability between different

materials in the same problem
Too much permeability contrast may cause numerical difficulties
The ratio between the highest and lowest permeability value
should not exceed 105
To simulate an almost impermeable material (e.g. concrete), a
value lower by a factor 1000 is sufficient
Classical formulation
Consolidation analysis based on excess pore pressure (EPP):
pactive psteady pexcess
Assumptions:
Steady state pore pressure is constant in time (horizontal phreatic level or
steady state pore pressure from groundwater flow calculation)
Excess pore pressure can change in time
Fully saturated soil (above and below phreatic level)
Limitation:
Time dependent hydraulic boundary is not possible (variable phreatic level)
Node
Element
Stress
point
Mesh:
Elements: Interpolation of primary variables
Nodes: Primary variables (displacements, pore pressures)
Stress points: Derived variables (strains, stresses, Darcy velocities)
Same order of interpolation in PLAXIS
19
Calculations:
Stiffness matrix
Coupling matrix
K v L p f Forces Equilibrium
Pore pressures
Displacements
Flow matrix
Displacements
dv dp
H pL S q
T
Net flow
dt dt Continuity
Compressibility of water
Transposed coupling matrix
K L v 0 0 v0 f

LT S p 0 t H p 0 t q System of equations
* *

S t H S q q0 q
* *
Solution: Displacements and (excess) pore pressures

Calculations:
K v L p f Equilibrium
dv dp
H pL S q
T
Continuity
dt dt
K L v 0 0 v0 f

S p 0 t H p 0 t q
LT *

*
System of equations
S t H S q q0 q
* *
H: permeability matrix H ( N )T k ( N ) dV
V
T
S: compressibility matrix S
n
Kw
N N dV
V
K: stiffness matrix T
K B D B dV
V
L: coupling matrix T
L B m N dV
V
Solution: Displacements and (excess) pore pressures
Boundary conditions (mechanical)

Fixities
Loads
Boundary conditions (hydraulic )
Prescribed groundwater head
No flow
Others: drains, wells, infiltration
INITIAL PORE PRESSURE AND EXCESS PORE PRESSURE
Sand
Clay pw
Sand pw0
e.g. t=0 t=1 t = 10 t = 40 t = 100 days
pw pw 0 pw pw 0 steady state pore pressure
Consolidation in classical mode in terms of excess pore pressures

Steady state pore pressures have to be established

Time step
Automatic time stepping is required
Critical time step (smaller steps may cause stress oscillations)
l2
tcritical
Cv
l = element length
= 80 for 15-node triangles
= 40 for 6-node triangles
Calculations:
Consolidation Staged construction > Time interval t
Consolidation Minimum pore pressure > |p-stop|
Consolidation Incremental multipliers > Time increment
Typical: Plastic calculation (staged construction) with undrained materials, followed

by consolidation analysis
Rate loading: Time increment and load increment give loading rate
Output:
Deformations
Stresses
Excess pore pressure
History curves
(e.g. pore pressure as function of time)
Validation: One-dimensional consolidation
Validation: One-dimensional consolidation
Calculation models
New features from PLAXIS 2D 2010

From the 2010 version, PLAXIS offers two types of consolidation
analysis: classical mode and advanced mode
TYPES OF CONSOLIDATION ANALYSIS
PLAXIS 2010 offers two types of consolidation analysis: classical

mode and advanced mode
Classical mode Advanced mode
Consolidation in terms of excess Consolidation in terms of total pp
pp (generated in undrained or Fully coupled analysis between
consolidation analyses) deformations and pore pressures
Coupling between deformation and Change in steady state pp,
excess pp excess pp and deformation
Steady state pp generated prior to influence each other
calculation More boundary conditions for flow
Changes in steady state pp may Soil can be saturated, dry or
change excess pp and unsaturated
deformation but not viceversa
Bishop stress is adopted
(semi-coupled)
Soil fully saturated or dry
Final remarks
Conclusions
FEM is quite suitable for 2D and 3D consolidation analysis
2D or 3D coupled consolidation is different from 1D or uncoupled consolidation
PLAXIS has several options for consolidation based on excess or total pore
pressure
Future perspectives
The classical and advanced calculation modes will be replaced (in version
2013?) by two options: ignore suction or allow suction
Exercise 4: De-watering in Excavation
Mr Siva Subramanian & Dr William Cheang
Excavation and dewatering
EXCAVATION AND DEWATERING
INTRODUCTION
This example involves the dry construction of an excavation. The excavation is supported
by concrete diaphragm walls. The walls are tied back by pre-stressed grout anchors. The
HSsmall model is used to model the soil behaviour. Special attention is focused on the output,
which provides us insight in the soil behaviour and its interaction with structural elements. It
is noted that the dry excavation involves a groundwater flow calculation to generate the water
pressure distribution.
The excavation is 20 m wide and 10 m deep. 15 m long concrete diaphragm walls of 0.35
m thickness are used to retain the surrounding soil. Two rows of ground anchors are used
at each wall to support the walls. The upper anchor has a total length of 14.5 m and an
inclination of 33.7o (2:3). The lower anchor is 10 m long and is installed at an angle of 45o .
The excavation is symmetric so only one half of the problem needs to be modelled.
Figure 1: Excavation supported by tie back walls
The relevant part of the soil consists of three distinct layers. From the ground surface to a
depth of 3 m there is a fill of relatively loose fine sandy soil. Underneath the fill, down to a
minimum depth of 15 m, there is a more or less homogeneous layer consisting of dense well
graded sand. This layer is particular suitable for the installation of the ground anchors. In the
initial situation there is a horizontal phreatic level at 3 m below the ground surface, (i.e. at the
base of the fill layer) Below the sand layer there is a loam layer which extends to large depth.
INPUT
Project properties
Start a new project in Plaxis. The symmetric problem can be modelled with a geometry model
of 60 m width and 40 m depth. Hence, set the model Contour to xmin = 0m, xmax = 60m,
y min = 40m and y max = 0m. Keep all other settings to their defaults.
Soil mode
Define a single borehole taking into account the following soil layers:
1. A fill layer of 3 m thick
2. A sand layer of 14 m thick
3. A loam layer of 23 m thick
Make sure the Head of the borehole is set to -3 m.

For the material properties, use the data given in table 1.
Table 1: Soil and interface properties

Parameter Symbol Fill Sand Loam Unit
Material model Material model HSsmall HSsmall HSsmall
Drainage type Drainage type Drained Drained Drained
Unsaturated soil weight unsat 16.0 17.0 17.0 kN/m3
Saturated soil weight sat 20.0 20.0 19.0 kN/m3
ref 3
Reference secant stiffness E50 20.5 10 38.5 103 20.0 103 kN/m2
from triaxial test
ref
Reference tangent Eoed 20.5 103 35.0 103 20.0 103 kN/m2
stiffness from oedometer
test
ref
Reference Eur 61.5 103 115.5 103 60.0 103 kN/m2
unloading/reloading
stiffness
Power for m 0.5 0.5 0.7
stress-dependent stiffness
Cohesion c0ref 1.0 1.0 8.0 kN/m2
0 o
Friction angle 30.0 34.0 29.0
o
Dilatancy angle 0.0 4.0 0.0
4
Threshold shear strain 0.7 1.0 10 1.0 104 1.5 104
Reference small-strain Gref
0 180.0 103 350.0 103 180.0 103 kN/m2
shear modulus
Advanced parameters Default Default Default
Horizontal permeability kx 1.0 0.5 0.01 m/day
Vertical permeability ky 1.0 0.5 0.01 m/day
Interface strength Rinter 0.65 0.7 Rigid
reduction factor
Coefficient for initial K0 Automatic Automatic Automatic
horizontal stress
Structures mode
The proposed geometry model is given in figure 2, all sizes and coordinates are given in
meters.
A ground anchor can be modelled by a combination of a node-to-node anchor and a embedded
pile row (purple line). The embedded pile row simulates the grout body (bonded length)
whereas the node-to-node anchor simulates the anchor rod (free length). The diaphragm
wall is modelled as a plate. The interfaces around the plate are used to model soil-structure
interaction effects.
Figure 2: Geometry model of building pit
The properties of the concrete diaphragm wall are entered in a material set of the plate type.
The concrete has a Youngs modulus of 35 GPa and the wall is 0.35 m thick. The properties
are listed in table 2.
Table 2: Properties of the diaphragm wall

Parameter Symbol Diaphragm wall Unit
Material type Material type Elastic
Isotropic - Yes
End bearing - Yes
Axial stiffness EA1 ,EA2 1.2 107 kN/m
Flexural stiffness EI 1.2 105 kN/m2 /m
Weight w 8.3 kN/m/m
Poissons ratio 0.15
For the properties of the ground anchors, two material data sets are needed: One of the
anchor type (anchor rod) and one of the embedded pile row (grout body). The anchor data
set contains the properties of the anchor rod and the embedded pile row data set contains the
properties of the grout body. The data are listed in tables 3 and 4.
Table 3: Properties of the anchor rod

Parameter Symbol Anchor rod Unit
Material type Material type Elastic
Axial stiffness EA 2.5 105 kN
Spacing Ls 2.5 m
Table 4: Property of the grout body (embedded pile row)

Mesh mode
Click the Mesh generation button and in the Mesh option window that appears, set the Element
distriibution to Fine. This should give a mesh as shown in figure 3.
Figure 3: Generated finite element mesh
Water conditions / Staged construction mode

The calculation consists of the initial phase and six phases.
In the first phase the wall is constructed.
In the second phase the first 3 m of the excavation are constructed without connection
of anchors to the wall. At this depth the excavation remains dry.
In the third phase the first anchor is installed and prestressed.
The fourth phase involves further excavation to a depth of 7 m, including the de-watering
of the excavation. This involves a groundwater flow analysis to calculate the new pore
water pressure distribution, which is a part of the definition of the third calculation phase.
In the fifth phase the second anchor is installed and prestressed.
The sixth phase is a further excavation (and de-watering) to the final depth of 10 m.
The calculation is done using a so-called semi-coupled analysis. This means that the groundwater
flow field is generated first and used as input to the deformation analysis. In other words, the
groundwater flow will have an effect on the deformations in the soil, but the deformations
in the soil will not change the flow field. This assumption is reasonable if the flow field
will not be disturbed by excess pore pressures resulting from undrained behaviour or by
significant changes in permeability due to large deformations. In this excavation problem
indeed permeabilities are high and undrained behaviour should be of little or no importance.
All calculation phases are defined as Plastic calculations of the Staged construction type
and standard settings for all other parameters. The instructions given below are limited to a
description of how the phases are defined within the Staged construction mode.
Initial phase
In Staged construction mode make sure that all soil is activated and all structural elements
are deactivated.
The initial phreatic line follows from the Head information specified in the borehole, and
should be located at y = -3m. That is, on the separation of the fill and sand layer.
Phase 1: Construction of the diaphragm wall

Construction of the diaphragm wall takes 5 days. Therefore, fill in a Time interval of 5
days in the General section of the Phases window.
In Staged construction mode activate the wall as well as the interfaces along the wall.
Phase 2: First excavation stage

In the Phases window do:
In the General section, set the Timer interval to 7 days

In the Deformation control parameters section select the option Reset displacements
to zero as were not interested in the displacements caused by the installation of the
wall.
In Staged construction mode
deactivate the first excavation part.
Phase 3: Prestress first anchor

Prestressing the first row of anchors will take 1 day, hence in the Phases window enter a 1
day construction time for this phase.
In Stage construction mode
activate the upper embedded pile row representing the grout body of the first anchor.
select the upper node-to-node anchor so that it appears in the Selection explorer.
In the Selection explorer now activate the anchor and also select the option Adjust
prestress. Set a prestress force of 300 kN.
Phase 4: Second excavation stage and dewatering

The second excavation stage includes dewatering and will take 10 days. Dewatering will be
simulated using a steady-state groundwater flow analysis.
In the Phases window
in the General section, set the Time interval to 10 days

also in the General section, set the Pore pressure calculation type to Steady-state
groundwater flow.
deactivate the second excavation part.
In Water conditions mode
No water flow can occur through a axis of symmetry. Therefore the axis of symmetry
must be a closed flow boundary. To do so, do for all lines on the left boundary and
for the bottom boundary:
Right-click the line.

From the menu that pops up, select the option Activate.
In the Selection explorer set the Behaviour to Closed. The boundary is now
drawn as a thick black line to indicate its an active closed boundary.
The groundwater head boundary conditions needed for the groundwater flow analysis
can be applied in a simple manner by using the global water level:
Select the Create water level button ( )
Draw a water level from (x y) = (-1 -7) to (20 -7), (30 -3) and finally to (61 -3)
Right click the water level that was just created and from the popup menu select
Make global to make sure the new water level is the global water level in this
calculation phase.
Phase 5: Prestress second anchor

Prestressing the second row of anchors will take 1 day, hence in the Phases window enter a
Time interval of 1 day.
In Stage construction mode
activate the lower embedded pile row representing the grout body of the second
anchor.
select the lower node-to-node anchor so that it appears in the Selection explorer.
In the Selection explorer now activate the node-to-node anchor and also select the
option Adjust prestress. Enter a 500 kN prestress force.
Phase 6: Third excavation stage and dewatering

The third excavation stage includes dewatering and will take 7 days. Dewatering will be
simulated using a steady-state groundwater flow analysis.
In the Phases window
in the General section, set the Time interval to 7 days
also in the General section, set the Pore pressure calculation type to Steady-state
groundwater flow.
deactivate the third excavation stage.
In Water conditions mode.
Check that both the axis of symmetry and the bottom of the model are closed
boundaries.
Draw a new general phreatic level from (x y) = (-1 -10) to (20 -10), (30 -3) and (61
-3).
Right-click the newly generated water level and make it the global level.
Phase 7: Safety analysis

Finally, create a Safety phase following phase 6 in order to determine the factor of safety for
the final excavation stage.
The safety analysis needs more than the default 100 additional calculation steps in order to
reach failure and therefore the number of additional steps must be increased:
Open the Phases window
In the section Numerical control parameters deactivate the option Use default iter parameters
and set Max steps to 200.
Nodes for load displacement curves

Select some nodes for load displacement curves, for instance the top of the wall at (x y) =
(10 0) and the middle of the excavation bottom at final depth at (x y) = (0 -10).
Now start the calculation.
ALTERNATIVE: TRANSIENT
GROUNDWATER FLOW
As an alternative calculation a fully coupled analysis will be performed. This analysis couples
transient groundwater flow, consolidation and deformations implying that the groundwater flow
field, the development and dissipation of pore pressures and the deformation are all calculated
simultaneously and all influence each other. This type of analysis should be performed if the
flow field is expected to be varying in time (transient flow) or when significant changes in
permeability due to large deformations are likely to occur. In this excavation problem the main
reason to use this analysis is to take into account that the flow field will not reach a steady-
state during excavations and so a transient flow analysis is required. The addidional effects
of coupling the flow field with undrained behaviour will probably be small as this project deals
with high permeabilities. Note that a fully coupled analysis requires that the calculation type is
Fully coupled flow-deformation.
It is possible to re-use the project made for the calculation using the method of steady-state
analysis:
Save the project under a different name
Change the calculation phases according to the description below.
Initial phase
No changes have to be made
Phase 1: Construction of the diaphragm wall

In the Phases window in the General section, change the calculation type to Consolidation.
Phase 2: First excavation stage

Phase 3: Prestress first anchor

Phase 4: Second excavation stage and dewatering

In the Phases window in the General section, change the calculation type to
Phase 5: Prestress second anchor

Phase 6: Third excavation stage and dewatering

Phase 7: Safety analysis

No changes have to be made
Select all phases for calculation and start the calculation
OUTPUT
Figure 4 gives the total displacements for the final phase for both the calculation with steady-
state groundwater flow and the transient groundwater flow.
The excavation using steady-state flow gives a maximum displacements of about 20 mm while
excavation using transient flow gives a maximum displacement of about 18 mm. Note that the
colours of the graphs are both scaled of 0 to 22 mm in 11 intervals.
Figure 4: Total displacements for the semi coupled analysis (left) and the fully coupled analyis
(right)
Figure 5 shows the vertical displacements for the final phase for both calculations. For the
displacements behind the wall the excavation using steady-state analysis clearly gives more
vertical displacements over a slightly larger distance from the excavation than the excavation
with transient flow.
Figure 5: Vertical displacements for the semi coupled analysis (left) and the fully coupled
analyis (right)
The extreme bending moments are about -155 kNm/m and 75 kNm/m for the excavation using
semi-coupled analysis while the extremen bending moments for the excavation using fully
coupled analysisare about -145 kNm/m and 95 kNm/m. Hence, the transient flow calculation
leads to a slightly smaller field bending moment, but a slightly higher foot bending moment
than the steady-state flow.
Figure 6: Bending moments in the wall for the semi coupled analysis (left) and the fully coupeld
analyis (right)
Figure 7 shows the horizontal displacements of the top of the wall as a function of
construction time for both the excavation using semi-coupled and fully-coupled analysis.
Figure 7: Horizontal wall displacements for the excavation
Finally, looking at the factor of safety for the final situation (see figure 8) it can be seen that the
fully coupled analysis gives a marginally larger factor of safety than the semi coupled analysis.
Figure 8: Strength reduction curve for the determination of the factor of safety
Unsaturated Soils & Barcelona Basic Model
CG14: UNSATURATED SOILS AND BARCELONA
BASIC MODEL
Antonio Gens

Vahid Galavi (Plaxis)
Unsaturated soils: outline
Introduction: characteristic problems in unsaturated soils

Suction in unsaturated soils
Flow and consolidation in unsaturated soils
Stress variables
Mechanical behaviour of unsaturated soils
Shear strength
Volume change
The BBM model for unsaturated soils
Unsaturated soils
Solid
Liquid Gas
V pores Vliquid Vgas Vliquid Vliquid

n Sr 1 Sg
Vtotal Vtotal V pores Vliquid Vgas
Porosity Degree of saturation
Unsaturated soils: failure
Shum Wan Road landslide, Hong Kong Island on August 13th,1995

Photographs from Geotechnical Engineering Office, Hong Kong
Engineering problems involving unsaturated soils: collapse
Fig. 4
Collapse in Via Luigi Settembrini, Naples (15-09-2001)
Unsaturated soils: a bit of history

1950s, 1960s
Relevance of suction recognized
Interpretation in terms of single effective stress
Late1960s, 1970s
Unsaturated soils as difficult soils, `special soils, regional soils
Late1970s, 1980s
Recognition of need for two stress variables
State surface approach
Late 1980s, 1990s onwards
Large expansion of research
Suction control and measurement
Elastoplastic models
Incorporation into mainstream Soil Mechanics

Stress variables
Shear strength
Volume change

New additional variable: suction
Water potential, : work required to transport a unit mass from a
reference pool of pure water to the soil water under consideration
c o g z
Matric Osmotic Gas Gravitational
c ( uw ua ) : Matric (capillary) potential

o cm RT : Osmotic potential
g ( ua uatm ) : Gas pressure potential
z w z : Gravitational potential
Review panel (1965)

Gravitational
potential
Gas pressure
potential
Matric potential
SEMIPERMEABLE
MEMBRANE
ns
os RT
V
SOLUTE
PURE
WATER
Osmotic potential
New additional variable: suction
Water potential, : work required to transport a unit mass from a
reference pool of pure water to the soil water under consideration
c o g z (Review panel, 1965)

Matric Osmotic Gas Gravitational
Total water potential controls water flow
Water potential affects mechanical behaviour. Not all potential
components have, however, the same effect
s w c :Matric suction w o :Osmotic suction

st s : Total suction
Total suction is directly related to relative humidity (psychrometric law)
In most cases, only matric suction is relevant for mechanical behaviour

Stress variables
Shear strength
Volume change
Unsaturated soils: flow and consolidation
Saturated soils: equation of continuity (with soil deformation)
n q x q y n
0
t x y t
Unsaturated soils: equation of continuity (with soil deformation)
(n S r ) q x q y
0
t x y (n S r )
t
n (, s ) S r (, s ) q x q y
Sr n 0
t t x y
Unsaturated soils: equation of continuity (with soil deformation)
(n S r ) q x q y (n S r )
0
t x y t
n (, s ) S (, s ) q q y
Sr r n x 0
t t x y
Constitutive Retention Darcys

law curve law
Unsaturated soils: retention curve
(also called soil water characteristic curve, SWCC)
Sr f pa pw = f s

1. Retention curves exhibit hysteresis effects

2. This value of negative water pressure is called the air entry value
for that soil
3. Negative pore pressures can exist in saturated soils
UNSATURATED
SATURATED
SATURATED
UNIFORM
(Brooks and Corey, 1964)

Retention curve: analytical expressions
Plaxis uses the Van Genuchten expression

1 g n
gn
gn
S ( p ) S residu ( S sat S residu ) 1 g a p
p s ( pa pw ) in m
S sat 1.0 S residu 0.027

g a 2.24 m 1 g n 2.286
An alternative linearized VG model is provided in Plaxis

Parameter values are provided depending on soil classification
Used-defined models are possible
Unsaturated soils: Darcys law
h pw
q k k h y
dy dy w
Relative permeability
k k rel ( S r ) k sat
Fredlund &
Rahardjo (1993)

Relative permeability: analytical expressions
Plaxis uses the Van Genuchten expression
2

g n 1

g n g n
S r S residu

k rel ( S r ) ( S e ) g l 1 1 S e g n 1 S e
S sat S residu

S sat 1.0 S residu 0.027
g a 2.24 m 1 g n 2.286
g l 0.0
An alternative linearized VG model is provided in Plaxis

Hydraulic parameters: data set
Standard (Topsoil part of Hypres)

Hypres (International soil
classification system)
USDA (International soil classification
system)
Staring (Dutch soil classification
system)
User-defined

Hydraulic parameters must be chosen very carefully!

Finite element formulation for coupled consolidation analysis
(advanced model only!)
The same shape functions are used for pore water pressure and
displacement.
i 1 i 1
K Q v 0 0 vi fu
T p 0 t H
i t (G q )
Q S t H w p w p

p
k rel k rel
H ( N )T G p ( N )T k w g ( N ) dV
sat sat
k ( N ) dV
H: permeability matrix V
w V
w
T nS dS
S: compressibility matrix S N
K
n
dp
N dV
V w w
K: stiffness matrix T
K B M B dV
V
Q: coupling matrix Q S B m N dV
T
Calculation procedure
The presented system is highly non-linear. The non-linearity occurs
because of suction dependent degree of saturation (Q and S) and the
suction dependent permeability (H).
A fully implicit scheme of integration is used to solve the fully coupled
flow-deformation analysis.
k rel
H ( N )T
sat
H: permeability matrix k ( N ) dV
V
w
T nS dS
S: compressibility matrix S N n N dV
V K w dp w
T
Q: coupling matrix Q S B m N dV
V

Stress variables
Shear strength
Volume change
A fundamental variable for unsaturated soils: (matric) suction
Matric suction is often associated with capillary phenomena
s ua u w
Intergranular capillary forces
In most cases, only matric suction is relevant for mechanical behaviour
Stress variables for unsaturated soils

It is necessary to use two stress variables to characterize the
behaviour of unsaturated soils
Conventional effective stress concept does not apply!
ua 1 ( s, Sr ) 2 ( s, Sr )
Class I ua (1 0)
Includes net stress and suction (BBM): ua , s
Class II ua 1 ( s)
Class III ua 1 ( s, Sr )
Includes Bishops stress and suction (Plaxis BBM): ua S r s , s
Stress variables for unsaturated soils
Class I
Includes net stress and suction (BBM): ua , s
Easy representation of conventional stress paths (including laboratory test
paths)
Difficulties in the transition saturated-unsaturated
Hysteresis and hydraulic effects difficult to incorporate
Independent function required to model the increase of strength with suction
Class III
Includes Bishops stress and suction (Plaxis BBM): ua S r s , s
Representation of conventional stress paths not straightforward, sometimes
impossible
No difficulties in the transition saturated-unsaturated (it recovers Terzaghis
effective stress on reaching Sr=1)
Hysteresis and hydraulic effects can be naturally incorporated
The increase of strength with suction results from stress variable definition
Stress variables
Isotropic plane using net stress and suction
: Wetting at constant (net) stress

: Loading a saturated soil
: Drying at constant (net) stress
: Loading at constant suction
: Stress path during a swelling pressure test

Stress variables
Shear strength
Volume change
Unsaturated soils: features of behaviour

Shear strength
Suction-controlled direct shear cell (Escario and Sez, 1980).

Behaviour of unsaturated soils: shear strength
Shear strength increases with suction
Moderate suctions
Escario & Sez (1986)

Shear strength increases with suction: a bilinear relationship
f c ' n p a tan ' s tan b
Variation of
apparent cohesion
and friction with
suction
Fredlund & Rahardjo (1985)
Shortcoming: close to saturation should be equal to '

b
Shear strength
s
s Large suctions
s
MATRIC SUCTION, s (kg/cm2) Escario & Juc(1990)
The bilinear relationship is not valid and must be modified

The rate of increase of strength close to saturation must be tan
The increase of strength is not linear but it becomes asymptotic at high
suctions
A single stress variable may account for the strength increase with suction
' pa ( pa pw ) ; ( Sl ) (Bishop stress)

Bishop (1959) proposal for effective stress:
' ua (ua uw ) ; ( Sr )
Bishop (1959) proposal for effective stress: ' ua (ua uw ) ; ( Sr )
Often ( Sr ) Sr leading to Bishop stress: ' ua Sr (ua uw )
Sr
Assuming (ua 0) and Sr leads to the simplified expression:
' ( S r u w )
f c' n tan ' c' ( n S r uw ) tan ' c' n tan ' S r uw tan '
The variation of shear strength with suction depends on the variation of Sr with
suction
Often Sr is replaced by Se (PLAXIS!)

S r S res Se: effective saturation
Se
S sat S res Sr: degree of saturation
Sres: residual saturation (suction )
Ssat: saturation when (suction=0)
Sand (USDA): (Sres=10,5%)
S r pw
S e pw

Hydraulic parameters: data set
Standard (Topsoil part of Hypres)

Hypres (International soil
classification system)
USDA (International soil classification
system)
Staring (Dutch soil classification
system)
User-defined

Hydraulic parameters must be chosen very carefully!

Stress variables
Shear strength
Volume change
Behaviour of unsaturated soils: volume change
Suction-controlled oedometer cell

Behaviour of unsaturated soils: consolidation lines
Suction increases the apparent preconsolidation stress

The soil can sustain a higher void ratio at the same stress value
(Oedometer
tests on a
Brazilian
residual soil;
Fig. 5 Lemos, 1998)

Collapse
Jennings & Burland

(1962)
Two stress variables are required to describe collapse
Volume change behaviour on saturation depends on applied stress level

Triaxial wetting tests on Ko-consolidated samples of Lower Cromer Till
(Maswoswe, 1985)
swelling collapse
After collapse, the saturated void ratio is recovered

Collapse (compression) is observed on wetting at high stresses, but a
(small) swelling is observed upon wetting at low stresses
Specimens of compacted clay at 90% of Normal Proctor energy and two

different water contents (Escario & Sez, 1973)
Sample A
Sample B
During collapse, volume strain may change sign (it can only be observed in
suction controlled tests)

Stress variables
Shear strength
Volume change
Elasto plastic model for unsaturated soils
The Barcelona Basic Model
(BBM)
Alonso, E.E., Gens, A. & Josa, A. (1990). A constitutive model for partially saturated soil,
Gotechnique, 40, No3, 405-430.
For partially saturated soils which are slightly or moderately expansive..
Implemented as UDSM in PLAXIS

by Gonzalez & Gens (2008) using
Bishops stress and suction as
stress variables.
The Barcelona Basic Model (BBM)
Net stress: ( pa ) Suction: s ( pa pw )

Isotropic plane Yield curve
S=0 Yield LC
Elastic
VOID RATIO, e
domain
SUCTION, s
S3
S3
S2 S2
S1
S1
S=0
MEAN NET STRESS, p MEAN NET STRESS, p
Yield curve Yield curve

LC1 LC2
Elastic
domain Loading
SUCTION, s
p o 1
L
S1
C
Collapse
p p
*
o 1
*
o 2 MEAN NET STRESS, p
Isotropic plane

Elastic Yield curve

LC LCB LCC
domain
A B C
SUCTION, s
SUCTION, s
plastic
compression
elastic
swelling
C B A vol
p p p
*
o
*
o B
*
o C compression swelling
MEAN NET STRESS, p

Isotropic plane
sSUCTION,
q
LC
s CSL (s)
Elastic CSL (s=0)

domain
s
s=0
po* po po* po
MEAN NET STRESS, p MEAN NET STRESS, p
Isotropic plane Deviatoric plane

Three-dimensional view
Unsaturated soils
FEATURES OF BEHAVIOUR
Suction increases the apparent preconsolidation pressure (yield
stress) and (often) soil stiffness
Volume change behaviour depends on stress level. Swelling or
compression (collapse) may occur depending on applied load
Collapse behaviour
After collapse soil lies on saturated consolidation line
Volume change reversal may occur during collapse
Volume change behaviour is path independent only for a certain
class of stress paths
Shear strength increases with suction
Elasto plastic model for unsaturated soils: BBM
Yield surface
2
g
f 3 J 2 p ps P0 p
2
g 30
Cam-clay ellipse (other choices are possible!)
g is function of lode angle () and J is the second deviatoric stress tensor.
Cohesion increases linearly with suction (not

really required in Plaxis implementation!):
ps k s S
ks is an input parameter.
Yield surface: LC curve

Preconsolidation stress varies with suction:
*0 *
P * *s *
P0 Pr 0
Pr
Pr : a reference mean stress (fitting LC curve).
0* : modified compression index of saturated soil.
* : modified swelling index of (un)saturated soil.
*s *0 1 r e S r
r & : input parameter (fitting LC curve).
55
Plastic potential
g 2
2
g 3J
2
p ps P0 p
g 30
is used to obtain K0 path for normally consolidated soils.
1
M M 9 M 3 *
1 *
96 M 0
Hardening law:
P0
dP0 d vp
0
* *
Elastic behaviour:
The elastic behaviour of the model is the same as Modified Cam-Clay
model:

d ve , p dp ; G
p
: elastic stiffness due to mean stress (input parameter) .
In addition to that, change in suction may produce volumetric elastic
strain according to (not really required in Plaxis implementation!):
s
d ve , s dS
3 s patm
s : elastic stiffness due to suction (input parameter) .
57
BBM: Triaxial tests
70 60
C3
s=0 kPa s=100 kPa s=200 kPa s=0 kPa s=100 kPa s=200 kPa
60 50
C2
50
40
CSL
q (kPa)
40
q (kPa)
30
30
C1
20
20
10
10
A B2 B3
0 0
0 0.1 0.2 0.3 0.4 0.5 0 20 40 60 80
Axial strain p' (kPa)
250 2.3
s=0 kPa s=100 kPa s=200 kPa
B3 C3 2.25
200
A
2.2
150 s=0 kPa s=100 kPa s=200 kPa
B2
s (kPa)
=1+e
B3
2.15
B2
100 C2
2.1
50 C3
2.05 C2
C1
A C1
0 2
0 10 20 30 40 50 60 70 0 10 20 30 40 50 60 70 80
q (kPa) p' (kPa)
Gonzalez (2008)
BBM: drying-wetting test
125 2.3
YLD A
YLD C 2.2
C NCL 0
100 B
2.1 NCL 1
A
75 2.0 B
s (kPa)
v=1+e
1.9 C
50 1.8
1.7
25
1.6 D
A D 1.5
0
0 50 100 150 200 250 300
0 50 100 150 200 250 300
p' (kPa)
p' (kPa)
Gonzalez (2008)

The Barcelona Basic Model Capable of reproducing
(BBM) main features of
unsaturated soil
behaviour
Provides a consistent
framework for an
integrated understanding
of unsaturated soil
behaviour
Complete/consistent but
based on drastically
simplified assumptions
Compatible with classical
models of saturated soils
Lots of simplifications!
Initial Stresses and Slope Stability Analysis
CG15
INITIAL STRESSES
PHI-C-REDUCTION
SLOPE STABILITY ANALYSIS
Helmut F. Schweiger
Initial Stresses / Phi-c Reduction / Slope Stability Analysis

2
Initial Stresses | Phi-c-Reduction | Comparison with Limit Analysis / Limit Equilibrium Analysis | Example of Failure | Unsaturated Slope
CONTENT
Initial stresses
general
K0-procedure
gravity loading
special cases
Phi-c-reduction
safety factor
safety factor analysis
examples
final advice
Comparison Phi-c-reduction Limit analysis Limit equilibrium analysis
Analysis of a failure
Slope stability analysis considering rainfall infiltration
1
3
Initial stresses represent the equilibrium state of the undisturbed soil

and consist of:
Soil weight
Loading history
In Plaxis two possibilities exist:

K0-procedure
Gravity loading

4
GENERATION OF INITIAL STRESSES AS FIRST STEP IN CALCULATION

PROCEDURE
Needed:
coefficient for lateral earth pressure K0
Disadvantage:
No equilibrium for inclined surface (a nil-step can be used)
Advantage:
No displacements are generated, only stresses, can take
into account initial OCR or POP
2
5
CALCULATION OF INITIAL STRESSES BY GRAVITY LOADING

Disadvantage
Non-physical displacements are created
Difficult to influence K0 value
Difficult to introduce initial OCR or POP

For 1D compression: 'h 'v so K0
1 1
Note: holds for elasticity and Mohr-Coulomb,
otherwise K0nc
Advantage
Equilibrium satisfied in all case
Attention: ignore undrained
> dont produce excess pore water pressures

6
CASES WHERE GRAVITY LOADING SHOULD BE USED INSTEAD OF

K0-PROCEDURE
3
7
SPECIAL CASES
For example a slope in overconsolidated soil or if

(approximate) modelling of geological history is required
Gravity loading needed due to geometry,

but initial OCR or POP required
K0 procedure using Phase 1: Phase 2:

desired K0, OCR, POP Excavate excess soil Reset displacements
Check K0, OCR !

8
SPECIAL CASES
For complex initial situations like inner city building projects it may be needed to
use several calculation phases to model the current situation before starting the
calculation for the actual project.
existing buildings
our
project
our project
initial phase 1 phase 2 our project
reset displacements to 0
4
9
Methods of assessing the factor of stability problems:

Limit equilibrium method > generally used in practical engineering
Displacement finite element method > phi-c-reduction (strength reduction method)
Finite element limit analysis
Limit equilibrium analysis

> Overall form of the failure surface needs
to be determined in advance
> Distribution of the interslice forces is
assumed differently in various methods
> Kinematic admissibility is not ensured
> Need to perform a global search for
identifying the failure mechanism with
the lowest factor of safety

10
DISPLACEMENT FINITE ELEMENT METHOD > PHI-C-REDUCTION
Factor of safety: Many possible definitions
available soil resistance

1.8
mobilized soil resistance
failure load
5.9
working load
PLAXIS: Safety factor on soil resistance
5
11
Strength reduction method: Phi-c-reduction

Same numerical tool as for serviceability design
Automatically detects most critical failure mechanism

12
Reduction of strength parameters c and tan() until failure is reached.

The factor of safety is then defined as:
tan
c tan tanreduced
Lowered incrementally
Msf
creduced tanreduced
6
13
Calculation procedure:
Create a Safety Phase
Accept the default increment for Msf = 0.1 from the multiplier tab-sheet
Calculate
Carefully examine Msf vs. displacement curve in Plaxis Curves to assure
that failure is indeed reached
If so, the value of Msf is assumed to be the factor of safety on soil resistance
Notes:
In order to check failure, select a control point within the (expected) failing
body
Use sufficient number of load steps
Choose elasto-plastic behaviour for wall, anchors and geotextiles with realistic
full plastic values in order to prevent excessively high structural forces
Displacements etc. AFTER safety analysis are meaningless

14
INFLUENCE OF STRUCTURAL ELEMENTS
wall elastic
Msf = 1.95
7
15
INFLUENCE OF STRUCTURAL ELEMENTS
wall elastic - perfectly plastic

Msf = 1.73

16
MAIN ADVANTAGES OF PHI-C-REDUCTION METHOD
Requires no a-priori assumptions on the failure mechanism

Critical surface is found automatically as slope failure occurs naturally
through the zones due to insufficient shear strength to resist shear stresses.
No requirement of assumptions on e.g. inter-slice shear force distribution
Applicable to complex conditions
Numerical tool same as for deformation analysis
Powerful alternative approach
8
17
INFLUENCE OF MESH DISCRETIZATION
20,00 20,00 20,00
2
1
15,00
5,00
60,00
Parameters:
' = 20 [] , c = 10 [kN/m]
E = 105 [kN/m] , = 0.3 [-] , = 20.0 [kN/m]

18

642 elements (6-noded) FoS: 1.37
642 Elemente (15-noded) FoS: 1.33
3211 Elemente (6-noded) FoS: 1.34
Incremental shear strain contours after /c-reduction

9
19
INFLUENCE OF MESH DISCRETIZATION DIFFERENT FE-CODE

3-knotiges
3-noded Dreieck mit 1 Remeshing
elements FoS: 1.59
1.59 3-noded elements FoS: 1.391.39
6-noded elements FoS: 1.43 6-noded elements FoS: 1.36
6-knotig 1.43 6-knotig 1.36
4-noded elements FoS: no failure 4-noded elements FoS: 1.90

8-noded elements FoS: 1.55
4-knotig kein Bruch 8-noded elements 4-knotig 1.90
FoS: 1.38
8-knotig 1.38
adaptive mesh refinement

FoS: 1.29

20
COMPARISON OF FOS FOR DIFFERENT METHODS AND FE-CODES
Method of Janbu FE-code FE-code

Bishop Spencer
Analysis (corrected) A B*
FoS 1.38 1.37 1.37 1.33-1.37 1.29-1.38
*) fine meshes, with quadratic elements
Note: for high friction angles (> 35 to 40) differences between LEM
and FEM may become larger and flow rule has more influence
10
21
15-noded elements Factor of Safety
5 1.90
11 1.62
(very coarse) 38 1.52
(coarse) 82 1.51
(medium) 170 1.50
(fine) 414 1.45
(very fine) 871 1.43
3733 1.43
15749 1.43

22
NUMBER OF LOAD STEPS

1.16 1.16
1.12 1.12
Sum-Msf
Sum-Msf
1.08 1.08
1.04 1.04
1.0 1.0
0.0 0.3 0.6 0.9 1.2 1.5 0.0 0.3 0.6 0.9 1.2
displacement displacement
1. Arrows of incremental 2. Shadings of incremental 3. Shadings of incremental

displacements displacements shear strains
11
23
UNDRAINED STABILITY OF A SLOPE
H = 12m 2
D cu = 50 kPa
1
Plaxis:
F = 1,35
cu = 100 kPa
cu 50 D
Stability charts: F N0 6.6 1.38 , N0 f ( , ) (Taylor,1948)
Pd 12 20 H

24

Automatic detection of most critical shear surface
Plaxis:
F = 1.34
cu = 50 kPa
cu = 60 kPa
12
25

Automatic detection of most critical shear surface
Plaxis:
F = 1.19
cu = 50 kPa
cu = 50 kPa

26
SOME FINAL ADVICE FOR PHI-C-REDUCTION
Always inspect the incremental displacements or strains as computed in the

last load step to make sure that failure is reached.
The mesh used in the calculation needs to be sufficiently fine.
Mesh: Refine and redo the phi-c analysis until the factor of safety remains
constant upon further refinement of the mesh.
Always use the arc-length time stepping procedure within the phi-c-reduction
(default)
Use a small tolerated error (maximum should be the default error of 1%)
Beware of threedimensional effects
Check for local, not relevant, failure mechanisms
13
27
FINITE ELEMENT LIMIT ANALYSIS (FELA)
Based on the lower/upper bound theorems of plasticity

Assumes a rigid-perfectly plastic material, an associated flow rule and small
deformations
Limit analysis basics:

Lower bound method Upper bound method
Requires a stress field that satisfies the Is based on a velocity field that satisfies the
boundary conditions, the yield criterion boundary conditions and implies an associated flow
and fulfils equilibrium. rule (strain and velocity compatibility conditions).
Any loads supported by such a stress Internal work (due to plastic shearing) is equated to
field are a lower bound on the true the external work (minimum internal power
failure load. dissipation).
Any loads derived from such a velocity field are an
upper bound on the true failure load.

28
FINITE ELEMENT LIMIT ANALYSIS (FELA)

Associated vs a non-associated flow rule:
Flow rule defines direction of plastic strain increment.
In displacement based FEM in general, a non-associated flow rule (with a dilatancy angle
smaller than the friction angle ) is usually employed.
Displacement finite element method: Finite element limit analysis:

assumes normality rule (plastic strain rates
are normal to yield surface)
Davis (1968) suggested the use of reduced
strength parameters, c* and * (in
combination with an associated flow rule)
c* c
tan * tan
PLAXIS: cos cos

stays constant as long as reduced >
with:
1 sin sin
(non-associated flow)
Once reduced = => reduced = reduced Point of interest:
(assosciated flow) Non-associated SRFEA vs FELA using c* and *
14
29
Homogenous slope - drained conditions:
Hs = 10m
Ds = 1 and 5
s = 15, 30, 45 and 60
Material Material set 1 Material set 2

Unit Frictional material Cohesive-frictional material
unsat [kN/m3] 17.0 19.0
c [kPa] 0 20.0
[] 35.0 25.0

30
FELA: mesh refinement is

performed adaptively as part
of the analysis DS = 5 and S = 45
SRFEA: mesh is fixed and a

sensitivity analysis is required
to ensure that it is sufficiently
fine
SRFEA: strength reduction finite element analysis (phi-c-reduction)

15
31
SRFEA: deformed mesh and

deviatoric strains
FELA - UB: deformed mesh

and power dissipation
FELA - LB: deformed mesh

and plastic multiplier
Material set 2, Ds = 5

32
3.5
DisplacementFEAassociatedflowrule
FactorofsafetyFoS[]
3.0 FELimitanalysisassociatedflowrule
2.5
Cohesive-frictional material
2.0
1.5
Purely frictional
material
1.0
10.0 20.0 30.0 40.0 50.0 60.0
SlopeinclinationangleS []
Limit finite element analyses: average values of FoSavg (FoSavg = (FoSLB+FoSUB)/2).

Maximum difference between SRFEA and FE limit analysis is less than 0.5%.
16
33
3.5 DisplacementFEAassociatedflowrule
FELimitanalysisassociatedflowrule
FactorofsafetyFoS[]
3.0 DisplacementFEAnonassociatedflowrule
FELimitanalysisDAVISapproach
2.5
Cohesive-frictional material
2.0
1.5
Purely frictional
material
1.0
10.0 20.0 30.0 40.0 50.0 60.0
SlopeinclinationangleS []
With non-associated flow rule in the displacement based FEA FoS reduces slightly,
but calculations may show numerical instabilities!
Flow rule has a minor influence in the considered example.
Davis approach yields lower FoS values for both numerical methods (conservative).

34
Slope inclination S SRFEA FELA - lower bound FELA - upper bound

(associated)
15 2.64 2.61 2.65 Material 1
30 1.23 1.21 1.25
45 0.70 0.70 0.72
60 0.41 0.40 0.42
Slope inclination S SRFEA FELA - lower bound FELA - upper bound

(associated)
15 3.32 3.31 3.36
30 1.95 1.97 2.01 Material 2
45 1.44 1.46 1.50
60 1.13 1.16 1.21
HOMOGENEOUS SLOPE: COMPARISON OF FACTORS OF SAFETY
> Phi-c-reduction compares extremely will with rigorous limit analysis
17
35
INFLUENCE OF DILATANCY FOR HIGH FRICTION ANGLES
= = 45
= 15
= 0

36
associated, = 45
non associated, = 0
18
37
associated, = 45 non associated, = 0

38
EXAMPLE OF STEEP SLOPE
Reference: Conte, E.; Silvestri, F.; Troncone, A. (2010)

Stability analysis of slopes in soils with strain-softening behaviour.
Computers and Geotechnics, Vol. 37 (5), 710-722.
19
39
RESULTS
high cohesion
Plaxis2Dvs Slide6.0
Slide6.0 FoS Plaxis2D
frictionangle cohesionc MorgensternPrice FoS for= FoS for=0
[] [kPa] [] [] []
29.0 1.194 1.035 1.013
15
33.0
21.5
1.300
1.097
1.131
1.018
1.112
KOLLAPS

22,5
28.0 1.321 1.187 1.134
17.0 1.097 1.056 KOLLAPS
30
23.0 1.308 1.230 1.139
14.0 1.103 1.074 KOLLAPS
35
19.5 1.299 1.247 .,126
11.0 1.105 1.082 KOLLAPS
40
16.0 1.294 1.260 1.108
8.0 1.098 1.080 KOLLAPS
45
12.5 1.292 1.266 1.083
5.5 1.104 1.085 KOLLAPS
high friction
50
9.5 1.305 1.286 1.060

40
INFLUENCE OF TENSION CUT OFF / TENSION CRACK

Extreme case:
= 0 Plaxis:
c = 50 kPa FoS = 1.255
= 65 tco = 100 kPa
Slide - no tension
crack: FoS = 1.259
Plaxis:
FoS = 1.092
Slide - with tension
tco = 0 kPa
crack: FoS = 1.094
20
41
FOS = 1.08

42
FOS = 1.08
21
43
CROSS SECTION

44
PLAN VIEW
Cross section
investigated
22
45
FINITE ELEMENT MODEL
Ground conditions: fill (sandy gravel) and weathered rock,

not always clearly identified
Strength parameters for stability analysis given in

geotechnical report: = 40, c = 0 kPa

46
FINITE ELEMENT MODEL

No failure in geotextile
reinforced slope
Area with reduced strength
for worst case analysis
23
47
SLOPE BEFORE CONSTRUCTION

Strength parameters based on back analysis of
failure: = 40, c = 11 kPa
FOS of natural slope: ~1.25

48
SLOPE AT FAILURE
= 40, c = 11 kPa
~0.98
Note: Limit equilibrium analysis (method of slices) did not predict failure
24
49
SLOPE AFTER FAILURE
= 40, c = 11 kPa
~1.11

50
SLOPE AFTER FAILURE
Assuming reduced
strength:
= 35, c = 2 kPa
> failure, i.e. not
realistic
25
51
DESIGN OF RECONSTRUCTION

52
SLOPE AFTER PARTIAL REMOVAL OF FAILED SLOPE
= 40, c = 11 kPa
~1.31
26
53
RECONSTRUCTED SLOPE AT FULL HEIGHT
= 40, c = 11 kPa
3 x 720 kN (a = 4 m)
~1.38

54
RECONSTRUCTED SLOPE AT FULL HEIGHT WITH TOP ANCHORS
= 40, c = 11 kPa
3 x 720 kN (a = 4 m)
2 x 720 kN (a = 4 m)
~1.53
27
55
RECONSTRUCTED SLOPE AT FULL LOAD
= 40, c = 11 kPa
3 x 720 kN (a = 4 m)
2 x 720 kN (a = 4 m)
~1.46

56
Assuming reduced
strength:
= 35, c = 2 kPa
28
57
Assuming reduced
strength:
= 35, c = 2 kPa
> ~1.06

58
FOS OF SLOPE CONSIDERING RAINFALL INFILTRATION
Rainfall is one of the main factors

causing slope failures.
Rainfall induced slope failures are

common problems in steep residual
soil slopes in the tropics.
Residual soils frequently exist in an

unsaturated state.
The infiltration of rainfall will increase

the groundwater level and water
pressure and decrease matric suction
of unsaturated soils.
Slope failures caused by heavy

rainfall (Indonesia, 2010)
29
59
FOS OF SLOPE DURING RAINFALL INFILTRATION

Hydraulic Characteristics
Van Genuchten (1980) presented a set of closed-form equations of hydraulic
characteristics for unsaturated soils which is based on the capillary model of Mualem
(1976). The Van Genuchten model introduced the relation between saturation and suction
pore pressure head (p):
gc
S
Sr
Sr
1
ga
gn

p
e
s
i
d
u
s
a
t
e
s
i
d
u
p
u
w
g

w
Sresidu: residual degree of saturation of the soil that describes the part of water that remains
in the soil even at high suction heads.
Ssat: degree of saturation of the soil when the pores are filled with water.
ga, gn and gc are empirical parameters, and it is assumed that:

1
gn n
gc

60
FOS OF SLOPE DURING RAINFALL INFILTRATION

Hydraulic Characteristics
The effective degree of saturation (Se) is obtained as:

S Ss
S rS
S

e
s
i
d
u

a t
r
e
s
i
d
u
The relative permeability based on Mualem - Van Genuchten is:

2
gng
1

gngne

kr
S
1
1
S

n
gl

e
l
where gl is an empirical parameter.
30
61

Shear strength of unsaturated soils
Bishop (1959) has modified Terzaghis classical effective stress theory and
proposed the following expression for the effective stress of unsaturated soils:
= ( - ua) + (ua - uw)
and : effective and total stress respectively

ua : pore air pressure
uw : pore water pressure
(ua uw) : matric suction
: matric suction coefficient, varies from 0 to 1 covering the range from dry to
fully saturated conditions.
Assuming (ua 0) and S leads to the simplified expression:
= ( - S uw)

62

Simple benchmark example
Homogeneous slope.
The international soil classification system USDA series is used for
determining the hydraulic parameters for the Van Genuchten model.
Mohr Coulomb failure criterion.
The height of the slope is 10 m and the gradient (horizontal to vertical)
is 2:1.
20m 20m 20m
10m
15m
31
63

Soil parameters for Mohr Coulomb Model
Description Symbol Unit Value
Unit weight kN/m3 20
Elastic modulus E kPa 7 500
Poisson's ratio - 0.35
Effective cohesion c' kPa 20
Effective friction angle ' o 20

64

Parameters for Van Genuchten Model
Soil Ksat (m/s) ga (1/m) gn (-) gl (-)

Sand 8.25E-05 14.50 2.68 0.50
Loamy Sand 4.05E-05 12.40 2.28 0.50 Hydraulic characteristics for
Sandy Loam 1.23E-05 7.50 1.89 0.50 different soils based on
USDA series with Van
Loam 2.89E-06 3.60 1.56 0.50 Genuchten parameters
Silt 6.94E-06 1.60 1.37 0.50
Silty Loam 1.25E-06 2.00 1.41 0.50
Sandy Clay Loam 3.63E-06 5.90 1.48 0.50
Clayey Loam 7.22E-06 1.90 1.31 0.50
Silty Clay Loam 1.94E-06 1.00 1.23 0.50
Sandy Clay 3.33E-06 2.70 1.23 0.50
Silty Clay 5.50E-07 0.50 1.09 0.50
Clay 5.50E-08 0.80 1.09 0.50 Source: Plaxis 2D Reference Manual 2010
32
65
FOS OF SLOPE CONSIDERING RAINFALL INFILTRATION (kN/m2)

100 kN/m2
Initial conditions 100.00

90.00
80.00
70.00
Initial suction in the model is 0 kN/m2
60.00
assumed to increase linearly 50.00

40.00
above ground water level until 30.00

20.00
ground surface. 10.00

0.00
0 kN/m2 -10.00
Soil Water Characteristic Curve

(SWCC)

66

Modelling assumptions
Initial ground water level horizontal at level of the toe of the slope.
Rainfall intensity: 10 mm/hour lasting 3 days (72 hours).
Minimum and the maximum pore pressure head: -0.1 m (min) and 0.1m (max).
Left, right and lower boundary are impervious boundaries.
Rainfall 10 mm/hour
General
33
67

Initial conditions
Initial degree of saturation for the four different hydraulic parameters (suction is the same)
85.43% 60.73%
100.00% 100.00%
(%)
(a) Clay (b) Sandy Clay 100.00
95.00
90.00
85,00
40.39% 14.08%
80.00
75,00
70.00
65,00
60.00
100.00% 100.00% 55,00

50.00
45,00
40.00
35,00
30.00
25,00
20.00
15,00
10.00
(c) Silt (d) Loamy Sand 5.00

0.00
saturation

68

Results clay
Before rain infiltration: suction
FOS = 1.709 (kN/m2)
100.00
90.00
80.00
70.00
60.00
50.00
40.00
30.00
After 3 hours rain infiltration: 20.00

10.00
FOS = 1.709 0.00

-10.00
saturation
(%)
100.00
95.00
90.00
85,00
80.00
75,00
70.00
After 6 hours rain infiltration: 65,00

60.00
FOS = 1.708 55,00

50.00
45,00
40.00
35,00
30.00
25,00
20.00
15,00
10.00
5.00
0.00
Failure mechanism Distribution of suction Degree of saturation
34
69

Results clay
After 48 hours rain infiltration: suction
FOS = 1.694 (kN/m2)
100.00
90.00
80.00
70.00
60.00
50.00
40.00
30.00

10.00
FOS = 1.692 0.00

-10.00
saturation
(%)
100.00
95.00
90.00
85,00
80.00
75,00
70.00
60.00
FOS = 1.686 55,00

50.00
45,00
40.00
35,00
30.00
25,00
20.00
15,00
10.00
5.00
0.00

70

Results loamy sand
Before rain infiltration: suction
FOS = 1.603 (kN/m2)
100.00
90.00
80.00
70.00
60.00
50.00
40.00
30.00

10.00
FOS = 1.596 0.00

-10.00
saturation
(%)
100.00
95.00
90.00
85,00
80.00
75,00
70.00

60.00
FOS = 1.592 55,00

50.00
45,00
40.00
35,00
30.00
25,00
20.00
15,00
10.00
5.00
0.00
35
71

Results loamy sand
After 48 hours rain infiltration: suction
FOS = 1.502 (kN/m2)
100.00
90.00
80.00
70.00
60.00
50.00
40.00
30.00

10.00
FOS = 1.344 0.00

-10.00
saturation
(%)
100.00
95.00
90.00
85,00
80.00
75,00
70.00
60.00
FOS = 1.229 55,00

50.00
45,00
40.00
35,00
30.00
25,00
20.00
15,00
10.00
5.00
0.00

72
FOS OF SLOPE CONSIDERING RAINFALL INFILTRATION - SUMMARY
Clay
Sandy
Clay
Silt
Loamy
Sand
Initial degree of saturation Distribution of suction after Degree of saturation after 72

72 hours infiltration hours infiltration
36
73
FOS OF SLOPE CONSIDERING RAINFALL INFILTRATION - SUMMARY
FOS Analysis Slope Stability

Time during Infiltration (R = 10 mm/hr)
(hours) Clay Sandy Silt Loamy 1.8
Clay Sand
0 1.709 1.680 1.656 1.603 1.7
3 1.709 1.679 1.649 1.596

1.6
6 1.708 1.678 1.644 1.592
9 1.705 1.674 1.637 1.590
FOS
1.5
12 1.704 1.673 1.634 1.586
18 1.702 1.667 1.617 1.582 1.4
24 1.699 1.662 1.602 1.577 Clay
Sandy Clay
36 1.699 1.653 1.573 1.550 1.3 Silt
Loamy Sand
48 1.694 1.644 1.536 1.502
1.2
60 1.692 1.634 1.492 1.344 0 20 40 60 80
72 1.686 1.623 1.423 1.229 time (hours)
37
Exercise 5: Slope Stabilised by Soil Nails
Mr Siva Subramanian & Dr William Cheang
Slope stability for a road construction project
SLOPE STABILITY FOR A ROAD

CONSTRUCTION PROJECT
INTRODUCTION
On the North Island of New Zealand a new road section has to be constructed along the shore
line of a tidal bay, see figure 1.
Figure 1: Situation overview for the newly constructed road
Though the easiest solution would have been to construct the road at a larger distance from
the bay as the slope gradients are easier there, this is not possible as the upper land is
privately owned which for historic reasons cannot be changed. The new road therefore had to
be constructed along the steeper gradient just next to the shore line of the tidal bay.
The hillside is mainly siltstone, weathered at the surface but intact at certain depth. Construction
will take place in summer when the ground water level is low. However, in winter the hillside
side almost fully saturates due to heavy rainfall, which has a significant influence on the
stability. For the construction of the new road part of the slope was excavated. The excavated
material is crushed and mixed with sand and gravel to make fill material to support the road.
During the first winter after the road construction the road started to tilt towards the tidal bay
and after assessing the winter situation the factor of safety was considered too low. The
decision was taken to stabilize the fill and hillside below the road using so-called launched soil
nails: long steel reinforcement bars that are shot with high speed into the ground.
Main goal of the analysis

Determine the factor of safety of the original hillside
Construct the new road under dry (summer) conditions and calculate its factor of safety
Simulate wet (winter) conditions and calculate its factor of safety
Apply stabilising soil nails and calculate the factor of safety in wet conditions
INPUT
Project properties
Start a new project and select appropriate Dimensions according to the size of the geometry
(see figure 2). After closing the Project properties window, open the Snapping options and
make sure to use a snap distance of 0.25m.
(a)
(b)
Figure 2: Soil model (a) and position of the road surface, construction details and soil nails (b)
Soil mode
Due to the complexity of the model the geometry will not be defined using boreholes, but
through soil polygons in Structures mode. Therefore, move directly to Structures mode.
Structures mode
First the intact siltstone is modelled.
Select the Create soil polygon button ( ) and from the submenu that opens,
select the Create soil polygon option.
Now draw a soil polygon starting from (x y) = (0 0) and then to (0 22), (30 16), (37
11), (46.5 7.25), (58 6), (65 6) and finally to (65 0).
Secondly, the weathered siltstone layer will be added. As the bottom of weathered
siltstone layer coincides with the top of the intact siltstone layer its not needed to draw
the complete soil polygon.
From the Create soil polygon submenu now select the option Follow contour.
Click at (x y) = (0 22) and draw a line to (0 25), (25 20), (31 19.25), (35 16), (37.5
14), (43 11), (46 10.25), (58 8.25), (65 8) and finally to (65 6).
Now right click to end the drawing. A soil polygon will be created from the line that
was just drawn and the upper contour of the intact siltstone layer below.
The last part of soil missing is the new fill that will be constructed for the road.
Select again the Create soil polygon option and draw a soil polygon from (x y) = (35
16) to (38 16), (43 11) and (37.5 14)
Now some additional lines must be specified in order to model the construction sequence.
From the Create line menu choose the option Create line.
Draw a line from (x y) = (25 20) to (30 16)
Draw a line from (x y) = (35 16) to (37 11) and finally to (43 11)
The road must be added, including the traffic load:
From the Create line button choose the option Create plate.
Draw a plate from (x y) = (30 16) to (38 16).
Choose the Select button ( ) in order to stop drawing plates.

Right-click on the just created plate and from the popup menu select the option
Create Line load
In the Selection explorer, make sure the line load Distribution is set to Uniform and
qy,start,ref = -10 kN/m/m to create a vertical line load of 10 kN/m downwards, per
meter out-of-plane.
And finally the 3 soil nails are added as well:
From the Create line button menu choose the option Create embedded pile row.
Insert 3 embedded pile rows according to the coordinates given in figure 2.
Material properties
Soil
Enter the material properties for the three soil data sets specified in table 1.
After entering all properties for the three soil types, drag and drop the properties to the
appropriate clusters, as indicated in figure 2.
Table 1: Soil material set parameters

Parameter Symbol Intact Weathered Reinforced Units
siltstone siltstone fill
Material model Model Mohr- Mohr- Mohr-
Coulomb Coulomb Coulomb
Type of behaviour Type Drained Drained Drained
Dry weight unsat 16.0 16.0 19.0 kN/m3
Wet weight sat 17.0 17.0 21.0 kN/m3
0
Youngs modulus E 12000 12000 20000 kN/m2
0
Poissons ratio 0.3 0.3 0.3
Cohesion c0ref 12 10 8 kN/m2
0
Friction angle 35 19 30

Dilatancy angle 0 0 0
Permeabilities kx , ky 1103 0.01 0.1 m/d
Tension cut-off Tension cut-off Disabled Enabled Enabled
Road surface
The road surface is modelled with a plate element. Therefore, create a new plate material set
using the parameters as specified in table 2 and assign it to the plate representing the road
surface.
Table 2: Properties of the road surface (plate)
Parameter Symbol Road surface Unit

Material model Model Elastic
Isotropic Yes
End-bearing No
Axial stiffness EA1 , EA2 2.5105 kN/m
Flexural stiffness EI 500 kN m2 /m
Weight w 3.0 kN/m/m
Poissons ratio 0.0
Soil nails
The 3 soil nails are modelled using embedded pile row elements. Hence, create a new
embedded pile row material set with parameters as specified in table 3 and assign the material
to all 3 soil nails.
Table 3: Properties of the soil nails (embedded pile rows)

Mesh mode
The road surface and the soil nails are automatically refined. However, as possible failure
would be expected in the weathered siltstone layer, this layer has to be refined as well.
The Coarseness factors as specified in figure 3 should be applied to the indicated areas. This
can be done in 2 ways:
1. From the vertical toolbar select the Refine mesh button ( ) and click on the areas to
be refined. For every click on an area or object its coarseness factor will become 70% of
its current value. Hence, to reach a coarseness factor of 0.5 its necessary to click twice
on the area, for a coarseness factor of 0.35 one has to click 3 times on the same area.
2. Select the areas and in the Selection explorer directly enter the appropriate coarseness
factors.
Now select the Generate mesh button ( ) and make sure the Element distribution is set to
Medium. After mesh generation, view the mesh (see figure 4)
Figure 3: Areas of the mesh to be refined
Figure 4: Generated mesh with refinement
Water conditions mode and Staged construction mode

The calculation consists of the initial phase and 12 calculation phases more in order to model
the proper construction sequence and the determination of the factors of safety at key moments
in the construction process.
Initial phase
The initial situation consists of the intact hill side and a phreatic level representing typical
summer conditions as construction starts in summer. In order to define the initial situation,
follow these steps:
Water conditions mode
From the vertical toolbar select the Create water level button ( ) and then the
option Create water level.
Draw a water level from (x y) = (-1 10) to (66 10). This water level will automatically
become the global water level.
The geometry has a non-horizontal soil layering, hence the K0 -procedure cannot
be used.
Open the Phases window and for the initial phase set the Calculation type to Gravity
loading.
Make sure only the clusters representing the original hillside are activated. Hence,
switch off the parts of reinforced soil.
Phase 1 - Stability prior to the construction
Before the construction is started the factor of safety is determined of the initial situation
Open the Phases window and change the Calculation type of this phase to Safety.
Phase 2 - Road excavation
The road excavation should continue from the initial situation and not from the results of the
safety factor determination. To do so:
Select the Initial phase.
Select the Add phase button ( ). A new phase (phase 2) will now be created that starts
from the initial phase.
Now we will define the phase:
In the Phases window, set the Calculation type to Plastic of loading type Staged
construction.
In order to discard the displacements during gravity loading make sure the option
Reset displacements to zero is selected under the Deformation control parameters.
Switch off the upper part of the road excavations, see figure 5.
Figure 5: Phase 2, road excavation
Phase 3 - Construction of the fill

This calculation phase that starts from Phase 2 is again a Plastic calculation, loading
type Staged construction.
Switch on the additional fill

Assign the reinforced fill material set to the 4 clusters of the fill area, see figure 6.
Figure 6: Phase 3, Construction of the fill
Phase 4 - Construction of the road

This calculation phase that starts from Phase 3 is another Plastic calculation, loading
type Staged construction.
Switch on the plate representing the road. Make sure the distributed load representing
the traffic load remains switched off.
Phase 5 - Apply the traffic load

Again a Plastic calculation of loading type Staged construction.
Switch on both parts (left nd right) of the distributed load representing the traffic
load. The plate representing the road surface remains switched on.
We are now finished with the road construction.
Phase 6 - Factor of safety of the road in summer conditions

In order to determine the factor of safety directly after constructing the road use a Safety
phase starting from Phase 5.
Phase 7 - Winter conditions

In winter, the water level inside the hill gradually increases due to rainfall. Only the highest
water level in winter will be modelled, for which a steady-state groundwater flow analysis must
be performed.
The increase of water level should occur after finishing the road construction and not after
determination of the factor of safety of this situation:
Select Phase 5 and press the Add phase button ( ). Now Phase 7 will be created,
starting from Phase 5.
In Water conditions mode
Select the Create water level button and draw a new water level from (x y) = (-1,20)
to (5,20) and further to (20,10) and (66,10).
Choose the Select button ( ) in order to stop drawing water levels.

Right-click on the newly created water level and select the option Make global to
make this new water level the global water level.
Select the bottom boundary of the model, and in the Selection explorer set the
Behaviour of the boundary conditions to Closed.
Now right-click on the bottom boundary and in the menu that pops up select the
option Activate in order to activate the closed boundary.
Open the Phases window and in the General section set the Pore pressure calculation
type to Steady-state groundwater flow.
Phase 8 - Factor of safety of the road in winter conditions

In order to determine the factor of safety directly in winter conditions create a Safety
phase starting from Phase 7.
Phase 9 - Apply top level soil nails

In winter conditions the factor of safety appears to be rather low and therefore it is decided to
improve stability by applying launched soil nails.
The application of the first level of soil nails should occur after calculating winter conditions
and not after determination of the factor of safety of this situation : select phase 7 and
create a new phase
Switch on the topmost soil nail, see figure 7.
Figure 7: Phase 9, Road construction with traffic load and topmost level of soil nails
Phase 10 - Factor of safety in winter conditions with top level soil nails
In order to determine the factor of safety directly in winter conditions with the topmost
level of soil nails installed create a Safety phase. Keep all default settings
Phase 11 - Apply additional soil nails

The application of the first level of soil nails should occur after installing the top level of
soil nails and not after determination of the factor of safety of this situation. Therefore,
create a phase starting from Phase 9
Switch on the 2 other soil nails
Phase 12 - Factor of safety in winter conditions with all soil nails installed
In order to determine the factor of safety directly in winter conditions with the all soil nails
installed create a final Safety phase.
For this Safety phase, set in the Phases window the amount of calculation steps (Max
steps) to 200 in the Numerical control parameters section.
Load-displacement curves
Before starting the calculation choose some points for node-displacement curves. In order
to check failure for the phi/c reduction phases the chosen points should be in the expected
failure zone. As there are several possible slope instabilities, chose at least points at (25,20),
(35,16), (38,16) and (43,11).
Now save the project and start the calculation by pressing the Calculate button.
SUCTION
Beforehand, it was estimated that the factor of safety of the slope before construction should
be in the order of 1.5 as there is no history of significant deformation for either low water table
(summer) and high water table (winter).
However, after the calculation it appears that the factor of safety before construction in summer
conditions is just over 1.2 and it is doubted that the factor of safety in reality is indeed that low.
Therefore the possibility of present suction is taken into account, as suction generally leads to
an increased factor of safety.
Save the project under a different name
Open the Phases window and for all phases uncheck the option Ignore suction in the
Deformation control parameters. Hence, we will allow for suction in all phases,
Mark all phases to be calculated.
Now recalculate the project
OUTPUT
Failure mechanisms
Figure 8 shows the failure mechanisms for all 5 conditions. Note that only for the winter
condition with all soil nails installed, the failure mechanism is different depending on whether
suction was taken into account. For all other conditions the failure mechanism is the same
with or without suction, though the actual factor of safety is different.
(a) Before construction (b) Summer conditions
(c) Winter conditions (no nails) (d) Winter conditions (top nails)
(e) Winter conditions (all nails, no suction) (f) Winter conditions (all nails, with suction)
Figure 8: Incremental displacements showing failure mechanisms
Factors of safety
In order to check the factors of safety, strength reduction curves (M sf vs. displacement of
a control point) must be made in the Curves module. As can be seen from figure 8 it is not
possible to use the same control point for all 6 factors of safety in case we ignore suction, as
the failure mechanisms are in different locations for different situations. Therefore we choose
the control points as:
(x y) = (25 20) for the winter conditions with all nails installed
(x y) = (35 16) for all other conditions.
To create the curve as shown in figure 9 follow these steps:
Open the Curves manager ( ) and choose to start a new chart.
Set the x-axis values to the total displacement of point (x y) = (35 16) and the y-axis
values to the Project multiplier M sf .
Right-click on the chart and choose the option Settings.
In the Settings window, on the tabsheet representing the curve, click the Phases... button
and in the Select phases window that opens, deselect phase 12 (factor of safety of the
winter conditions with all nails installed) so that it will not appear in the graph. To clean up
the graph a bit more, one can decided to deselect all phases that are not Safety phases
as well.
Close the Select phases window but do not close the Settings window.
In the Settings window now select the Add curve button and then from the popup menu
select From current project.
Add a new curve, but now with the total displacements of point (x y) = (25 20) on the
x-axis. The y-axis values remain the Project multiplier M sf .
Back in the Settings window, on the tabsheet representing the newly added curve, click
again the Phases... button. Now deselect all phases but keep phase 12 selected.
Close the Phases window
Addittionally, on the Chart tabsheet of the Settings window one can set the scaling of
the axes. For instance the x-axis from 0 to 2 m. Press the Apply button to confirm this.
We now have a graph with the strength reduction curves for point (x y) = (25 20) for the final
phases and for point (x y) = (35 16) for all other calculation phases.
Please note that in case we do calculate with suction, all graphs can be created from point
(x y) = (35 16) as this point is in the failure zone for all situations (see figure 10).
Figure 9: Factors of safety for key moments in the project without taking into account suction.
Figure 10: Factors of safety for key moments in the project taking into account suction.
From figures 9 and 10 the effect of installing the nails on the factor of safety can be seen. It
can also be seen that taking into account suction gives a factor of safety prior to construction
that is more in accordance of the expected value, while suction only has a minor influence on
the factor of safety in winter conditions as in winter conditions most of the soil is fully saturated.
Hoek-Brown and Rock Jointed Models
CG17
HOEK-BROWN
JOINTED ROCK MODEL
Helmut F. Schweiger
Hoek-Brown / Jointed Rock model

2
Continuum Model for Rock | Plaxis Jointed Rock Model | Examples | Hoek-Brown Model | Example Pressure Tunnel | Concluding Remarks
CONTENTS
Continuum model for jointed rock

Plaxis jointed rock model
Validation with data from literature
Examples
Hoek-Brown model
Comparison with analytical solution
Practical Example: FE-modelling of grouting pressures for
pressure tunnels
Concluding remarks on discontinua
1
3
JOINTED ROCK BEHAVIOUR

Behaviour of rock = Behaviour of intact rock + Joints (joint sets)
approx. even, unfilled peak and residual shear strength depend on joint conditions
T . friction angle joint
n tanT T* .residual friction angle joint
F . friction angle joint material
residual n tanT* F* .residual friction angle joint material
cF . cohesion joint material
approx. even, filled
cF* . residual cohesion joint material
G . friction angle intact rock
n tanF c F G* residual friction angle intact rock
cG cohesion intact rock
residual n tan F* c F*
tensile strength is usually assumed
uneven, (un)filled
to be 0 for all cases
for n n tan (T i) residual n tan ( T* i)
for n n tan G c G residual n tan G*


4
Schematic model of "rough" joint
for n
n tan (T i)
for n
n tan G c G
2
5
Alternative formulation for joint behaviour
JCS
n tan JRC log10 r

n
JRC .. joint roughness coefficient

JCS .. joint compressive strength
R residual friction angle

6
JOINTED ROCK
distance of joints or fault zones large > explicit modelling required
3
7
JOINTED ROCK
distance of joints small (sets of joints) > explicit modelling not

necessarily required, approximation with suitable constitutive
model possible (smeared model, homogeneous model)

8
Assumption:
approximately parallel, continuous, unfilled joint sets
m
d rock d int act rock d joj int set
j 1
d rock d int act rock d jo int set

for each joint set a failure (yield) function is defined:
F j , n 0
4
9
Validity of model
- intact rock
- parallel joint sets
- distance between individual
joints small compared to typical
dimension of structure
- no fault gouge in joints
behaviour of joint sets can be taken into account

Max. 3 joint sets (Plane 1, 2, 3)
anisotropic, elastic perfectly-plastic model
Coulomb perfect plasticity on joints

10
Parameters required
Anisotropic elastic parameters
E1 Young's modulus
1 Poisson's ratio
E2 Young's modulus in Plane 1 direction
G2 Shear modulus in Plane 1 direction
2 Poisson's ratio in Plane 1 direction
Strength parameters in joint directions (Planes i = 1, 2, 3 )

ci Cohesion
i Friction angle
i Dilatancy angle
t, i Tensile strength
Definition of joint orientation (Planes i = 1, 2, 3)

n Number of joint directions
1i Dip angle
2i Dip direction
5
11
Definition of joint direction
(2=90)
1 x

12
Definition of anisotropy and sliding directions:
Dip angle 1
Dip direction 2
6
13
8000
7000
6000
1-3 5000
2
[KN/m ]
4000
3000
45+(/2)
2000
1000
0
20 30 40 50 60 70 80 90
orientation
Winkel of joint 1
der Trennflchen

14
Qualitative comparison of contour lines of stresses with data from

literature ( = 0)

from Wittke 1984
7
15
Qualitative comparison of contour lines of stresses with data from

literature ( = 90)

from Wittke 1984

16
INFLUENCE OF DIRECTION OF JOINT SET
full excavation
MStage (until failure)
unsupported
overburden: 60 m
width of tunnel: approx. 14 m
height of tunnel: approx. 12 m
8
17


18

9
19
INFLUENCE OF K0
influence of K0 (=32; 2652 elements)
1.0

0.9
0.8
0.7
0.6
MStage
0.5
0.4
0.3 Ko = 0,3
0.2 Ko = 0,7
0.1
Ko = 1,0
0.0
0 22.5 45 67.5 90
orientation of joint ( 1)

20
INFLUENCE OF DILATANCY
joints 1 = 45; = 0

MStage = 0.13
10
21
joints 1 = 45; = 5

MStage = 0.18

22
joints 1 = 45; = 10

MStage = 0.23
11
23
Influence of (K0=0,3; =32; 2652 elements)

1,0
0,9
0,8
0,7
0,6
MStage
0,5
0,4
0,3 Dilat. = 0
0,2 Dilat. = 5
0,1 Dilat. = 10
0,0
0 22,5 45 67,5 90
orientation of joints

24
1, 3 .. principal stresses
c .. uniaxial compressive strength
m, s . material parameters
s =1 > intact rock, s = 0 > heavily jointed
m > .. brittle behaviour, m < .. ductile behaviour
12
25
1, d 1
p
HB failure line
{dp}
gf
=0
{d }
p
=+60
{d }
p
max=+90 3, d 3
p
t +90 0
Hoek-Brown criterion: flow rule
1, d 1
p
MC failure line
{d }
p
=0
gf
{dp}
gt = ft
=90
3, d 3
p

26
GENERALISED HOEK-BROWN CRITERION

Hoek E., Carranza-Torres C., Corkum B., 2002. HoekBrown failure criterion - 2002 edition. In: Proceedings
of the North American rock mechanics Symposium, Toronto.
3
1 3 ci (mi s) a
ci
GSI 100 GSI 100
mb mi exp s exp
28 14 D 9 3D
a e GSI / 15 e 20 / 3
1 1
2 6
GSI geological strength index
mi depends on type of rock
D disturbance factor
(due to blasting and/or stress relaxation)
Failure function can be written as:

~ ~
f HB 1 3 f ( 3 ) with f ci (mb 3 s) a
ci
13
27
GEOLOGICAL STRENGTH INDEX

28
INTACT UNIAXIAL COMPRESSIVE STRENGTH CI
Source: RocLab
14
29
DISTURBANCE FACTOR D
Source: RocLab

30
GENERALISED HOEK-BROWN CRITERION - VALIDATION

Carranza-Torres C., 2004. Elasto-plastic solution of tunnel problem using the generalized form of the HoekBrown failure
criterion. Int J Rock Mech Min Sci; 41(3), 4801.
Intact rock parameters:
HB constant, mi [-] 10 Rock mass parameters:
Uniaxial compression strength, sci [MPa] 30 HB constant mb 1.6767
Geological strength index, GSI [-] 50 Parameter s 0.0038
Hydrostatic pressure, p0 [MPa] 15 Parameter a 0.5057
Young's modulus, E [MPa] 5700 Parameter D 0
Poisson's ratio,n [-] 0.3
Plastic radius = 2.58 m

Support pressure = 2.5 MPa
15
31
Relative shear stresses Plastic points
Plastic radius = 3.79 m

Support pressure = 0 MPa

32
Elasto-Plastic Stress Distribution (after Carranza-Torres)
25.0
Support pressure = 0 MPa

20.0
15.0
Stress [MPa]
10.0
Radial Stress (Exact)
Tangential Stress (Exact)

5.0
Plaxis HB-Model Radial Stress
Plaxis HB-Model Tangential Stress

0.0
0.0 5.0 10.0 15.0 20.0 25.0 30.0
Distance from Tunnel Center [m]
16
33
Elasto-Plastic Stress Distribution (after Carranza-Torres)
30.0
Support pressure = 2.5 MPa

25.0
20.0
Stress [MPa]
15.0
10.0
Radial Stress (Exact)
Tangential Stress (Exact)
5.0
Plaxis HB-Model Radial Stress
Plaxis HB-Model Tangential Stress
0.0
0.0 5.0 10.0 15.0 20.0 25.0 30.0
Distance from Tunnel Center [m]

34
SHORT PROJECT DESCRIPTION
Longitudinal section
1. Project description
17
35
SHORT PROJECT DESCRIPTION
Typical cross section Detail A:

Detail A

36
PLAXIS MODEL
132m
120m
18
37
Detail of numerical model

38
Modelling of grout pressure

Whirlpool
activated pore
tunnel interior
pressure
Queenston
Q10
19
39
Calculation phases:
0. Initial phase (K0 = 1.5)
1. Excavation (Mstage = 0.2)
2. Activation of shotcrete (Mstage = 1.0)
3. Activation of final lining
4. Pressure phase 1(8 bar, Egap <<)
5. Pressure phase 2 (15 bar)

40
RESULTS
Constant pressure application: p = 15bar
Analytical PLAXIS
Solution
Normal force 10334 kN/m 10220 kN/m 1%
Displacements r = 4.06 mm Pux = 4.06 mm 0%
20
41
Diametrical displacements for unsymmetric pressure application as

measured on site
2 1
1
4
3

D =
6
5
8
8 7

42
Unsymmetric pressure
application in Plaxis
21
43
RESULTS
Unsymmetric pressure application: resulting diametrical strains
D [mm] D D
1-8 3-6 4-5 2-7 [mm] [-]
Displacements on site -9.66 -3.73 -3.44 -6.43 -5.82 -4.6151E-04
Results in PLAXIS -9.60 -3.92 -3.45 -6.79 -5.94 -4.7139E-04

K0 = 1.5
Comparison with Hoek- Brown model
D [mm] D D
1-8 3-6 4-5 2-7 [mm] [-]
MC -9.60 -3.92 -3.45 -6.79 -5.94 -4.7139E-04
HB -9.60 -3.92 -3.45 -6.79 -5.94 -4.7141E-04

44
DISCONTINUUM METHODS (E.G. DISCRETE ELEMENT METHOD)
Allow modelling of finite displacements and rotations of deformable or rigid bodies

Individual bodies can separate and new contacts are detected
Advantages
Modelling of blocky structures (discontinua)
For explicit solution algorithms no equation system required
Suitable also for studying micromechanical behaviour of granular materials
Disadvantages
In 3D long calculation times
For static problems artificial damping required
Influence of various input parameters difficult to judge, i.e. joint stiffnesses
(> may cause numerical problems, lot of experience required)
22
45
DISCONTINUUM METHODS ( E.G. DISCRETE ELEMENT METHOD)
Such displacement patterns cannot be obtained

from continuum models as discussed

46
WHICH MODEL
Selection often done according to availability / familiarity rather than problem oriented
With simplification relevant mechanisms often lost, and thus neglected in the support
design
discontinuum model jointed rock model (continuum) isotropic continuum model
Acknowledgement: W. Schubert
23
Modelling of Tunnels in 2D
CG18
MODELLING OF TUNNELS IN 2D
Helmut F. Schweiger
2D Modelling of Tunnels
2
Introduction | NATM Excavation Sequence | 3D Effects in 2D | Calibration of pre-relaxation | Modelling Support Measures
CONTENTS
Introduction
Typical excavation sequence for NATM-tunnels
Modelling 3D-effects in plane strain analysis
Calibration of pre-relaxation factors
Modelling support measures
1
3
INTRODUCTION
3-D Models
easy modelling of excavation sequence

computational effort high
essential for analysis of face stability
Plane Strain
assumption of pre-relaxation factors

excavation sequence in cross section
face stability cannot be considered
"state of the art" in practice
4
TYPICAL NEW AUSTRIAN TUNNELLING METHOD (NATM) EXCAVATION

0 1 2 3 4 5 6 7
0 2 4 6
primary support primary support

initial stress state tunnel invert
of top heading of bench
1 3 5 7
unsupported zone transition top head transition bench final

at tunnel face - bench excavation invert excavation lining
2
5
DEFORMATION AHEAD OF FACE

tunnel face monitoring section
excavation
sequence
Chainage
settlements ahead
of tunnel face
settlements of
Settlements
unsupported zone
settlements after
installation of
monitoring
section
6
2D MODELLING - LOAD REDUCTION METHOD
E0 0( )
approximate values for :
= 0.2 0.5
for top heading
excavation
s = . 0 = 0.4 0.8
for side drift
excavation
(Laabmayr & Swoboda 1986)
tunnel lining
PLAXIS: Mstage 1-
PRE-RELAXATION
EXCAVATION
Note: Mstage definition changed
3
7
2D MODELLING - STIFFNESS REDUCTION METHOD
E0
approximate values for :
= 0.3 0.5
Ec = . E0 (Schikora & Fink 1982)
tunnel lining
PRE-RELAXATION
EXCAVATION
8
PRE-RELAXATION
4
9
FULL EXCAVATION WITH LINING
10
FULL EXCAVATION WITHOUT LINING
5
11
CHOICE OF AND
values depend on:
ground conditions
length of unsupported section
advance rate
time of construction of invert
experience of personnell (workmanship)
...........
Alternatively use 3D analysis to

estimate pre-relaxation factors
Note: these are different for
displacements and lining stresses
12
WHICH METHOD?
Working Group 1.6 of DGGT

> Load Reduction Method
Stiffness Reduction Method

Influence of
Poisson ratio
Constitutive model
Correlation of and very difficult
6
13
EXAMPLE NATM TUNNEL
Weathered Silt-, Claystone

Overburden 25.0 m
Top Heading Excavation
y = 60 m
Z = 60 m
y = 142 m
Finite Element Meshes

x = 70 m
x = 70 m
14
PARAMETERS FOR DIFFERENT CONSTITUTIVE MODELS
Parameters for different constitutive models

Eoed,ref E50,ref Eur,ref c m K0 POP K0,nc ur G0,ref 0,7
MODEL [MN/m] [MN/m] [MN/m] [kN/m] [] [] [] [kN/m] [] [] [MN/m] []
1)MC, E135 E=135MN/m 35 27 0,54

2)HS, EMC=Eoed 45 45 135 35 27 0,8 0,54 0,2
A 0.7 500
B auto 500
C 0,7 0
3)HS, EMC=Eur 20 20 60 35 27 0,8 0,7 500 0,54 0,2
7)HSS, EMC=Eoed 45 45 135 35 27 0,8 0,7 500 0,54 0,2 225 2*104
9)HSS, EMC=Eur 20 20 60 35 27 0,8 0,7 500 0,54 0,2 100 2*104
Advanced Course on Computational Geotechnics, New Delhi, India, 29 - 31 October 2014 14
7
15
3D STAGED CONSTRUCTION
Tunnel excavation in step i

Activation of lining with parameters shotcrete young in step i+1
Increase of cohesion in anchored region in step i+1
Increase of lining stiffness to shotcrete old in step i+2
16
2D STAGED CONSTRUCTION
Pre-relaxation (MStage < 1,0)

Activation of lining with parameters shotcrete young
Increase of cohesion in anchored region
(MStage < 1,0)
Increase of lining stiffness to shotcrete old (MStage =1,0)

MStage<1.0 MStage<1.0
Prerelaxation Shotcreteyoung+anchors Shotcreteold

8
17
Calibration of pre-relaxation factors by means of crown settlements

Chainage [m]
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140
-5
Crown settlements [mm]
-15
Good agreement between 2D
and 3D for final settlement
-25
Significant differences in
relaxation factor depending
-35 on constitutive model applied
-45
-55
3D 1) MC drained 2D 1) MC drained
3D 2) HS E45 (POP500,K0=0.7) drained 2D 2) HS E45 (POP500, K0=0.7) drained
3D 3) HS E20 (POP500,K0=0.7) drained 2D 3) HS E20 (POP500, K0=0.7) drained
3D 7) HSS E45 (POP500, K0=0.7) drained 2D 7) HSS E45 (POP500, K0=0.7) drained
3D 9) HSS E20 (POP500, K0=0.7) drained 2D 9) HSS E20 (POP500, K0=0.7) drained
Zelger 2012
18
Calibration of pre-relaxation factors by means of surface settlements

Chainage [m]
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140
0
-2
Surface settlements [mm]
-4
-6 Poor agreement between 2D
-8 and 3D for final settlement
-10
-12 Significant differences in
-14 relaxation factor depending
-16 on constitutive model applied
-18
-20
-22
-24
3D 1) MC drained
2D 1) MC drained
3D 2) HS E45 (POP500,K0=0.7) drained
2D 2) HS E45 (POP500, K0=0.7) drained
3D 3) HS20 (POP500,K0=0.7) drained
2D 3) HS E20 (POP500,K0=0.7) drained
3D 7) HSS E45 (POP500, K0=0.7) drained
2D 9) HSS E20 (POP500, K0=0.7) drained Zelger 2012
9
19
Final note on pre-relaxation:

Pre-relaxation is a way of considering 3D
effects in a 2D but the amount of pre-
relaxation to be taken into account in order
to match 3D results depends on a number of
Mstage
factors, such as
reference displacement
constitutive model
drained/undrained conditions

Zelger 2012
20
Introduction | NATM Excavation Sequence | 3D Effects in 2D | Calibration of pre-relaxation | Modelling Support Measures|
MODELLING OF SUPPORT MEASURES

Shotcrete
strength and stiffness highly time dependent
increase of Young's modulus for subsequent
excavation steps is a practical approach
alternatively complex constitutive model can be used
new model in PLAXIS
Rock bolts
in practice often by means of increase of cohesion
special elements of various types (bars, beams, ..)
embedded piles in PLAXIS
Additional support measures
elefant foot
micropiles at sidewall
Final concrete lining
additional calculation without modelling of excavation sequence in
detail (subgrade reaction method)
from FE-analysis assuming that shotcrete lining
carries no load in the long term
10
21
SUPPORT TOP HEADING
sometimes not sufficient support for beam

of top heading > numerical problems
temporary invert for top heading
micropiles, elefant foot
22
EXAMPLE FOR ANALYSIS WITH ADDITIONAL SUPPORT MEASURES
Calculation phases
Phase 0 Phase 1 Phase 2 Phase 3
Initial stresses Mstage=0.3 Mstage=1.0 Mstage=0.3

K0 procedure excavation top shotcrete young shotcrete top heading
heading young > old
berwimmer 2011 excavation bench
11
23
EXAMPLE FOR ADDITIONAL SUPPORT MEASURES

Calculation phases
Mstage=1.0 Mstage=0.3 Mstage=1.0 Mstage=1.0

shotcrete bench shotcrete bench shotcrete invert shotcrete invert
young young > old young young old
berwimmer 2011 excavation invert
24
INFLUENCE OF GEOMETRIC REPRESENTATION OF LINING
Continuum element Plate in centre line Plate at outer Plate at inner

of lining boundary of lining boundary of lining
berwimmer 2011
12
25
Vertical displacement of tunnel crown
Phase & No.

Phase uy
0.0
0 1 2 3 4 5 6 7 8
-5.0
-10.0
continuum element
uy [cm]
cross section 12
uy [10 *m]
plate inner
cross boundary
section 13
-15.0
plate centre
cross section 14
plate outer
cross boundary
section 15
-20.0
-25.0
-30.0
phase [-]
berwimmer 2011
26

Displacements utot
Phase 3 Phase 7
continuum element
plate inner boundary
plate centre
plate outer boundary
13
27

Axial forces N [kN/m]
Phase 3 Phase 7
plate inner boundary
plate centre
plate outer boundary
continuum element
28
Improved plate model
Horizontal beam element Additional cluster at base

14
29
Phase No.
continuum element
horizontal beam
cluster at base
plate only
berwimmer 2011
30
INFLUENCE OF ELEFANT FOOT
Phase & uy
Phase No.
0.0
0 1 2 3 4 5 6 7 8
-5.0
0.6 mdreieckklein
wide
y [cm]
*m]
-10.0
0.4 m0.2wide
m
-2
0.2 m0.4wide
m
uy [10
-15.0 0.6 m
d of lining
u
cross section 13
plate only
-20.0
-25.0
phase [-]
15
31
INFLUENCE OF MICROPILES
Phase & uy
Phase No.
0.00
0 1 2 3 4 5 6 7 8
-2.00
-4.00 ohne
u [cm]
27
-6.00
uy [10y-2 *m]
35
-8.00 43
51
-10.00
27+35
-12.00 alle
-14.00
-16.00
phase [-]
Influence of inclination of micropile (27 to 51)
32
2D EMBEDDED PILES FOR ANCHORS IN TUNNELLING

Comparison with phase2
circular tunnel D = 9.4 m, two step excavation (pre-relax. 70%/35%)
6-noded finite elements
rock: E = 850 MPa, = 0.2, K0 = 0.4
100 m c = 400 kPa, phi = 28
lining: E = 5 GPa / 15 GPa (young/old)
anchors: d = 3.2 cm, E = 210 GPa
Nmax = 230 kN, Tmax = (fully bonded)
100 m
16
33
RESULTS: LINEAR ELASTIC ROCK

Normal force in tunnel lining (end of excavation)
identical
34
RESULTS: LINEAR ELASTIC ROCK, NO SHAFT FAILURE

Anchor force 1 for various stages
good match
EP reduce to
~0 at anchor
end
17
35
RESULTS: ROCK MOHR-COULOMB, SHAFT RESISTANCE=5.0 KN/M

anchor forces at end of excavation (Swellex anchor in phase2)
36
NEW SHOTCRETE MODEL

governed by vp = 1p + 2p + 3p from tension yield surface
p
fty ftp 1 ftu 1 vp

with tup 28

2 Gt 28 ftp 28 f tu 28
tu Leq f tp228

Leq 2
Ael ftp28
nint 3

Leq = 1m
Gt28
tension
ftu28

ftp ftu
18
37
NEW SHOTCRETE MODEL - APPLICATION

NATM tunnel excavation with temp. side drift walls
Q1
Q5
N1
N5
N1
N5
38

NATM tunnel excavation with temp. side drift walls
th centre
th left 5 th right
3 1
bench bench
left bench right
centre
4 2
6
19
39

elastic lining (E = 5/15 GPa) vs. shotcrete model (no softening)
+3.5 MPa
+15 MPa
Tensile Stresses
40

tension softening vs. no softening
crack in softening +3.5 MPa
Tensile Stresses
20
Exercise 6: Tunnelling in Rock
Dr William Cheang
Excavation of a tunnel in rock
EXCAVATION OF A TUNNEL IN ROCK
This excercise has been created by
Prof. Helmut F. Schweiger

Technical University Graz, Austria
INTRODUCTION
In this example, a circular deep tunnel within a 1-layer homogeneous rock mass under a
hydrostatic pressure is modelled. As the hydrostatic pressure is the equivalent of 1500m
water it is not so practical to model this with the use of a phreatic level. In order to take into
account the hydrostatic pressure there are two possible method:
1. Apply the hydrostatic pressure as a distributed load
2. Apply the hydrostatic pressure by means of a thin heavy soil layer
Method 1) works fine if the hydrostatic pressures is applied during the calculation. However,
in our case the pressure is there from the initial situation. As distributed loads are not taken
into account during the K0 -procedure using a distributed load would require gravity loading for
applying the distributed load. This would cause additional problems to solve as gravity loading
may give plasticity in the rock joints that should not occur.
Method 2) therefore is applied in this exercise. It is chosen to use a thin layer with the same
properties as the rock (that is, a Jointed Rock material set) but with high weight. Alternatively
one can use a Linear Elastic material set with high weight, but in this case its necessary that
the Linear Elastic material has low stiffness so that it cannot prevent sliding along the joints to
reach the soil surface.
INPUT
The geometry and mesh of the model are illustrated in figure 1.
The refinement area around the tunnel can be modelled by inserting an additional circular
tunnel without lining or interface at the same location as the tunnel.
Note: Though the geometry is symmetric it is not possible to only model half the
geometry. The rock itself has an inclined stratification under an angle of
450 , what makes the model asymmetric!
Material properties
The material data sets of the rock and the thin heavy layer representing the hydrostatic
pressure (referred to as Equivalent pressure) are shown in table 1.
After excavation of the tunnel it is finished by applying a 0.2m thick concrete lining with a uni-
axial cylindrical compression strength of 55 N/mm2 and 3% reinforcement with FeB500 steel
bars (yield stress 435 N/mm2). Table 2 shows the material properties for the tunnel lining.
Mesh generation
For the generation of the mesh set the Global coarseness parameter to Medium. In addition,
select the 3 clusters that form the area around and in the tunnel, and perform a cluster
Figure 1: Geometry of the tunnel
refinement by choosing the menu option Mesh -> Refine cluster. After the mesh has been
generated, repeat the cluster refinement for just the inner two (circular) areas.
CALCULATION
The tunnels is excavated and at a certain distance behind the tunnel face the lining is applied.
The tunnel is supposed to be self supporting and so the main purpose of the lining is to protect
the tunnel from rock fall. However, it has to be taken into account that the lining will deliver
limited support for the tunnel due to continuous stress redistribution and deformations after the
tunnel lining is installed. This is done by calculating 80% of the effect of the excavation with
an unsupported tunnel and 20% of the effect of the excavation with lining installed. Hence, the
-method is used with = 0.2.
In PLAXIS, at the beginning of the staged construction calculation phase, the changes made
cause a certain unbalance between the externally applied forces on the mesh and the internal
stresses in both soil and structural elements. This unbalance physically cannot exists and has
to be solved in the calculation phase. This is done with the use of a multiplier, -Mstage, that
indicates the amount of the unbalance that has already been solved. This means that at the
beginning of the calculation phase -Mstage=0 as all unbalance still has to be solved and at
the end of a succesfully calculated phase -Mstage=1 indicating the full unbalance could be
solved. However, in our case we only want to solve 80% of the unbalance with an unsupported
excavation before applying the tunnel lining which implies that we have to tell the calculation
kernel that not the full unbalance should be solved but that the calculation must stop when
-Mstage=0.8 is reached.
Table 1: Rock properties
Parameter Symbol Rock Eq. pressure Units

Material model Material model Jointed Rock Jointed Rock -
Type of behaviour Drainage type Drained Drained -
Weight sat , unsat 24.0 15000 kN/m3
Youngs modulus for intact rock E1 13.0105 13.0105 kN/m2
Poissons ratio for intact rock 1 0.25 0.25 -
Youngs modulus for jointed rock E2 4.0105 4.0105 kN/m2
Poissons ratio for jointed rock 2 0.25 0.25 -
Shear modulus for jointed rock G2 1.33105 1.33105 kN/m2
Number of joint planes Number of planes 1 1 -
Cohesion of plane 1 c 50.0 50.0 kN/m2
0
Friction angle of plane 1 22.0 22.0
0
Dilatancy angle of plane 1 0.0 0.0
0
Dip angle of plane 1 1 45.0 45.0
Coefficient for initial lateral stress K0 1.0 1.0 -
Table 2: Tunnel lining properties
Parameter Symbol Lining Units

Type of behaviour Type Elastoplastic -
Axial stiffness EA 7.7106 kN/m
Flexural stiffness EI 25667 kNm2 /m
Weight w 2.4 kN/m2
Poissons ratio 0.0 -
Full plastic bending moment Mp 350.0 kNm/m
Full plastic normal force Np 10.0103 kN/m
Note: The results of a calculation phase with -Mstage < 1 have no physical
meaning as there is no equilibrium in external forces and internal stresses.
The calculation consists of the initial phase and 2 additional calculation phases
Initial phase
It is assumed there are no pore pressures in the rock, therefore the phreatic level can remain
below the geometry and no pore pressures have to be generated.
Initial stresses are generated using the K0 procedure. Make sure the thin layer representing
the top load is switched on.
Phase 1
In this stage construction phase the tunnel is being excavated while the tunnel lining remains
switched off.
In PLAXIS Calculation go to the Parameters tabsheet.
In the Loading Input box make sure the type of loading input is Staged construction and
then press the button Advanced. A new window will appear.
In this window one can specify that the end value for the -Mstage multiplier as described
above. Fill in a value of 0.8 and press OK to close the window again
Now press the Define button in order to define the staged construction changes, i.e. the
excavation of the tunnel by switching off the cluster representing the tunnel.
Phase 2
In this phase the tunnel lining is applied. Define the staged construction phase and switch
on the lining. No other changes have to be made in this phase. PLAXIS will apply the full
unbalance that exists at the beginning of this phase, which is the remaining 20% unbalance
from excavating the tunnel and the full additional unbalance caused by switching on the tunnel
lining.
Note: The Jointed Rock model is a rather sensitive model for both mesh
refinement and local failure. With this model it may occur more often than
for the other standard consitutive models in Plaxis that the calculation
stops reporting failure, while in fact the failure is only very local and not
important from engineering point of view. In some cases this situation can
be overcome by swichting off the Arc-length control in the Manual settings
of the Iterative procedure.
Arc-length control is a method to obtain reliable load values in case of
failure, but may also lead to detection of local failure mechanisms hat are
tof little interest as stated. As long as a calculation phase ends succesfully
(that is, it was possible to apply all changes requested by the user) the
answers should be the same wether arc-length control is switched on or
not. Only when failure is reached and arc-length control is switched off,
the user will be confronted with the fact that no automatic failure detection
will be done by PLAXIS (hence the user has to decide on whether he
considers the situation failure or not) and the failure load reached is not as
accurate.
In this exercise it may be necessary to switch of arc-length control in the
first calculation phase. However, the user is encouraged to always first try
with arc-length control switched on.
One can select some points for curves, for instance at the crown and the side of the tunnel
and somewhere halfway those two points.
Press the Calculate button to start the calculation.
OUTPUT
Check the influence of the stratification on the stresses and deformations of the rock around
the tunnel. Also, check the bending moments of the tunnel lining:
Figure 2: Plastic points at the end of phase 2
Figure 3: Volume strains at the end of phase 2
Figure 4: Total displacements at the end of phase 2
Figure 5: Bending moments in the lining
FURTHER EXERCISE
Repeat the calculation for dip angles 1 = 0 o and 1 = 90 o and compare results
Figure 6: Plastic points for 1 = 0o (left) and 1 = 90o (right)
Figure 7: Volume strains for 1 = 0o (left) and 1 = 90o (right)
Figure 8: Total displacements for 1 = 0o (left) and 1 = 90o (right)
Figure 9: Bending moments in the lining for 1 = 0o (left) and 1 = 90o (right)

Plaxis Advanced Course New Delhi 2014 (Print)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Plaxis Advanced Course New Delhi 2014 (Print)

Uploaded by

Copyright:

Available Formats

PLAXIS Advanced Course

New Delhi, India.

Technical University of Catalunya, Barcelona

some of the slides were originally created by:

The Finite Element Method

Stability (local and general)

Admissible deformation and displacements

design requirements in geotechnical engineering

Sometimes flow and stability/deformation problems are solved together

Potts & Zdravkovic

geotechnical methods of numerical analysis

methods for numerical analysis

A good analysis, which simulates real behaviour, allows the engineer to

Objectives of the numerical (finite element) analysis

geotechnical finite element analysis

Advantages of numerical (finite element) analysis

(Nearly) unavoidable uncertainties

Some requirements for successful numerical modelling

geotechnical finite element analysis

three final remarks

2. the quality of a tool is important, yet the quality of a result

3. the design process involves considerably more than analysis

Borrowed from C. Viggiani, with thanks

the FEM is a computational procedure that may be used to obtain an

the governing mathematical equations are approximated by a series of

The Finite Element Method: introduction and overview

The FEM is a computational procedure that may be used to obtain an

What kind of problem?

Apply load obtain displacements

Apply head obtain flow

Though we would like to know our solution at any coordinates in our

this lecture presents only a basic outline of the method

The FEM involves the following steps (1/2)

Primary variable approximation

A polynomial form is assumed, where the order of the polynomial

The Finite Element Method: introduction and overview

The FEM involves the following steps (2/2)

where is the element stiffness matrix, is the vector of nodal

Apply load obtain displacements

Displacements Strains Stresses Loads

Combine these steps: K BT DBdV

The first stage in any FE analysis Footing

A mesh consists of elements

examples: multi-anchored diaphragm wall

There is a whole zoo of different finite elements available!

two-dimensional analysis of continua is generally based on the use of

the most used elements are based on an iso-parametric approach

Shape function of node i

Shape functions for 3-node line element

Shape functions for 5-node line element

Illustration for the six-noded triangular element

12 coefficients, depending on the values of the

Illustration for the six-noded triangular element

Constitutive relation (elasticity)

linear isotropic elasticity in plane material stiffness matrix

What happens with inelastic constitutive relations?

It is necessary to apply the external load in separate increments

Element stiffness matrix

Element stiffness matrix

Apply load obtain displacements

Displacements Strains Stresses Loads

Combine these steps: K BT DBdV

Element stiffness matrix

MC criterion: tc cos - s sin