Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
PL
LAXIS
S STA
ANDA
ARD COUR
C RSE
MUUMBAI, INDIA
SEPTEMBER 2012
18-21 S
CONTENTS
Lectures & Exercises on 2D and 3D Modelling
DAY 1 TUESDAY 18.9.12
THEME GEOTECHNICAL FINITE ELEMENT MODELLING
Time Module Description Lecturer
Dr. Juneja/
09:00 10:00 CG1 Geotechnical Finite Element Modelling and Plaxis 2D
Dr William
Dr. Juneja/
10:15 11:15 CG2 Introduction to Mohr‐Coulomb Model
Dr William
DAY 2 WEDNESDAY 19.9.12
THEME
Time Module Description Lecturer
09:00 10:00 CG7 Geometry, Meshing and Element Types in Plaxis* Dr.Cheang
3
DAY 3 THURSDAY 20.9.12
THEME
Time Module Description Lecturer
09:00 10:00 CG13 Initial Geo‐static Stresses in Plaxis Dr.Cheang
Dr. Juneja/
3:15 4:15 CG17 Consolidation Analysis in Plaxis
Dr William
DAY 4
THEME
Time Module Description Lecturer
09:00 10:00 CG19 Introduction to Plaxis 3D Dr Cheang
4
Finite element modelling in
geotechnical engineering
Objectives:
5
Basic concepts of the
Finite element method (deformations)
load
equilibrium stiffness matrix
stress displacement
6
Basic concepts of the
Finite element method (deformations)
A A
3 4 5 2
6 7
0 x 1
7
Basic concepts of the
Finite element method (deformations)
Isoparametric elements with nodes for two-dimensional analysis:
element
node
Triangular elements
Quadrilateral elements
stress
point
Triangular elements
Quadrilateral elements
8
Basic concepts of the
Finite element method (deformations)
Isoparametric elements for three-dimensional analysis:
Tetrahedral elements
Brick elements
Tetrahedral elements
Brick elements
9
Basic concepts of the
Finite element method (deformations)
Interpolation functions for linear 3-node triangular element:
a0 v1x b0 v1 y
u x ( x, y ) a0 a1 x a2 y
a1 v2 x v1x b1 v2 y v1 y
u y ( x, y ) b0 b1 x b2 y
a2 v3 x v1x b2 v3 y v1 y
y
u x ( x, y ) N1v1x N 2 v2 x N3v3 x
3
v1y u y ( x, y ) N1v1 y N 2 v2 y N 3v3 y
v1x N1 1 x y
x
1 2 N2 x
N3 y N : Shape functions
a0 v1
u x ( x, y ) a0 a1 x a2 y a3 x 2 a4 xy a5 y 2
a1 3v1 v2 4v4
u y ( x, y ) b0 b1 x b2 y b3 x 2 b4 xy b5 y 2 a2 3v1 v3 4v6
a3 2v1 2v2 4v4
N1 (1 x y )(1 2 x 2 y )
N 2 x(2 x 1) a4 4v1 4v4 4v5 4v6
y N 3 y (2 y 1) a5 2v1 2v3 4v6
v5y
3 N 4 4 x(1 x y )
6 5 v5x N 5 4 xy
N 6 4 y (1 x y ) u x ( x, y ) N1v1x N 2 v2 x ... N 6 v6 x
x
1 4 2 u y ( x, y ) N1v1 y N 2 v2 y ... N 6 v6 y
CiTG, Geo-engineering, http://geo.citg.tudelft.nl Basic concepts of FEM 12
10
Basic concepts of the
Finite element method (deformations)
Interpolation functions for quadratic 6-node triangular element:
xx ( x, y ) u x x a1 2 a3 x a4 y
yy ( x, y ) u y y b2 b4 x 2 b5 y
xy ( x, y ) u x y u y x (b1 a2 ) (a4 2b3 ) x (2a5 b4 ) y
du x dN1 dN dN
y xx v1x 2 v2 x ... 6 v6 x
v5y dx dx dx dx
3 du y dN1 dN dN
yy v1 y 2 v2 y ... 6 v6 y
6 5 v5x dy dy dy dy
du x du y dN1 dN dN dN
x xy v1x 1 v1 y 2 v2 x ... 6 v6 y
1 4 2 dy dx dy dx dy dx
11
Basic concepts of the
Finite element method (deformations)
Strains for 6-node triangular element:
B ve xx ( x, y ) v1x
v
yy ( x, y )
1y
xy ( x, y ) v2 x
N1 N 2 N 6
v2 y
x v
0 0 ... ... 0 e
x x
N1 N 2
N 6
...
B 0 0 ... ... 0
y y y
N N1 N 2 N 2 N 6 N 6
...
1 ... ... v
y x y x y x
6x
B : Strain interpolation matrix v6 y
CiTG, Geo-engineering, http://geo.citg.tudelft.nl Basic concepts of FEM 15
f1x
f
y 1y
3
f5y f2x
f2 y
f
6 5 f5x e
...
x
1 4 2
...
f
6x
f 6 y
CiTG, Geo-engineering, http://geo.citg.tudelft.nl Basic concepts of FEM 16
12
Basic concepts of the
Finite element method (deformations)
Element stiffness matrix:
K B M B dV B M B wk
e T T
1 0
Hooke’s law: M D
E 1 0
(1 2 )(1 )
0 0
2
1
13
Basic concepts of the
Finite element method (deformations)
Elements and nodes in a FE mesh (global node numbers are indicated):
K v f v i v i 1 v
14
Global stiffness matrix
K
e
K
elements
K 11e1 K 12e1
e1 0
K 21
e1
K 22 K 11e 2 K 12e 2
e2 e2
K 21 K 22
0
CiTG, Geo-engineering, http://geo.citg.tudelft.nl Basic concepts of FEM 21
Strains, stresses
Once v’s are known:
B ve
M
B v
e
i i 1 M
15
Finite element method (deformations)
Read input data K B M B dV B M B wi K
e T T e
K
Form stiffness matrix i 1
i elements
f ex f ex f ex
i
New step
f in B c dV B c wk
i 1 i 1
Form new load vector
T T
Calculate unbalance v 0
Reset displacement increment j j 1
New iteration v K 1 f
Solve displacements v v v
j j 1
Calculate unbalance f in f ex f in
i k
16
Mohr-Coulomb model and soil stiffness
Objectives:
F
1-3
strength
P
stiffness
-1
1
v
1 dilatancy
-1
3 v 3
17
Typical results from soil lab tests
Oedometer test (one-dimensional compression)
Pre-consolidation stress
1 reloading 1
primary loading
1
1
unloading
1 time
1 creep
1
18
Typical results for soil stiffness
Stiffness at different levels of strain
19
Concepts of soil modelling
yy
yx
yz xy
• Relationship between stresses (stress rates)
zy
and strains (strain rates) xx
xz
zx
• Elasticity (reversible deformations) d=f (d) zz
• Example: Hooke’s law
• Plasticity (irreversible deformations) d=f (d,,h)
• Perfect plasticity, strain hardening, strain softening
• Yielding, yield function, plastic potential, hardening/softening rule
• Example: Mohr-Coulomb yielding
• Time dependent behaviour (time dependent deformations)
• Biot’s (coupled) consolidation d=f (d,,t)
• Creep, stress relaxation
• Visco elasticity, visco plasticity
Lin. elast. perfectly-plast. EP strain-hardening EP strain-softening
CiTG, Geo-engineering, http://geo.citg.tudelft.nl MC model and soil stiffness 8
20
Hooke’s law
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
xy 2
yz 0 0 0 0 1
2
0 yz
1
zx 0 0 0 0 0 2 zx
Inverse:
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
Hooke’s law
21
Model parameters in Hooke’s law:
d1
Two parameters:
- d1
- Young’s modulus E
- Poisson’s ratio
d3
- 1
Meaning (axial compr.):
d1 E
E
d1 1
d 3 - 1
d1 1
3
CiTG, Geo-engineering, http://geo.citg.tudelft.nl MC model and soil stiffness 11
Bulk modulus:
dp
dp E
K dv
d v 31 2
Oedometer modulus:
- d1
d E 1 - d1
Eoed 1
d1 1 1 2
22
Stress definitions
• In general, soil cannot sustain tension, only compression
• PLAXIS adopts the general mechanics definition of stress and strain:
Tension/extension is positive; Pressure/compression is negative
yy yy
yy yy
• In general, soil deformation is based on stress changes in the
grain skeleton (effective stresses)
• According to Terzaghi’s principle: σ’ = σ - pw
Symbolic: d ' D d
e
d D e 1
d '
23
Plasticity
d ij d de
ij
p
ij (strain rates)
Plasticity
24
When do plastic strains occur?
f<0
Within the yield contour: f<0 f>0
On the yield contour: f=0
Outside the yield contour: f>0 (impossible stress state)
25
The Mohr-Coulomb failure criterion
Origin: F ’n
T
T
’
A c’
F ’n
CiTG, Geo-engineering, http://geo.citg.tudelft.nl MC model and soil stiffness 19
’3
’1
26
The Mohr-Coulomb failure criterion
c cos MC criterion:
c
-3 -1 -n
-s*
t* = ½(’3 - ’1)
s* = ½(’3+’1)
1
2 '3 '1 c' cos ' 12 '3 '1 sin '
27
Visualisation of the M-C failure criterion
’
c’
’n
-’1
2c' cos '
a
b 1 sin '
1
1 sin '
a b
1 sin '
-’3
CiTG, Geo-engineering, http://geo.citg.tudelft.nl MC model and soil stiffness 23
1 1
2 '3 '2 c' cos ' 12 '3 '2 sin '
2 '2 '3 c ' cos ' 2 '2 '3 sin '
1 1
3
2 '1 '2 c ' cos ' 2 '1 '2 sin '
2 1 1
28
Reformulation into yield functions
1
2 '3 '1 c' cos ' 12 '3 '1 sin '
29
Plastic potentials of the M-C model
g1a 12 '3 '2 12 '3 '2 sin c' cos
g1b 12 '2 '3 12 '2 '3 sin c' cos
g 2 a 12 '1 '3 12 '1 '3 sin c' cos
g 2b 12 '3 '1 12 '3 '1 sin c' cos
g 3a 12 '2 '1 12 '2 '1 sin c' cos
g 3b 12 '1 '2 12 '1 '2 sin c' cos
Failure in a simple
shear test: xx
t*
yy
g
d ije D e ijkl d 'kl 0
1
d ijp d d xx 0
'ij
g ' ' yy 1
d xxp d d xx 2 sin 0
' xx 4 t*
g ' ' xx 1
d yyp d d yy 2 sin d sin
' yy 4 t*
g '
d xyp d d xy d cos
' xy t*
d yy d yyp
p tan
CiTG, Geo-engineering, http://geo.citg.tudelft.nl
d MC dand
xy model xy soil stiffness 28
30
Failure in a simple shear test:
d yy d yyp
tan xy
d xy d xyp
xy
yy
dilatancy xy
31
The LEPP Mohr-Coulomb model
Model parameters:
Parameter determination
32
MC approximation of a CD triax. test
1-3 E ’50 ’3 = confining pressure
-1
v
2 sin
1 sin
-1
1-2’
CiTG, Geo-engineering, http://geo.citg.tudelft.nl MC model and soil stiffness 33
Eoed
-1
(1 )(1 2 )
Eoed E
(1 )
33
Stiffness parameter – suggestions
15000 cu 5000 cu
- Clay:
u
E50 or G50
I p [%] I p [%]
Ip = plasticity index
Eoed 1'
150..500
p ref p ref
34
Stiffness parameter – suggestions
-1 Order of magnitude Eoed (clay):
500
Eoed '1 (correlation)
Ip
pref Eoed
Eoed 3..5 qc (correlation)
-1
Eoed
d 1
1 E E
(1 )(1 2 )
Eoed
d1 1 1 2 (1 )
This E–value applies to primary compression
CiTG, Geo-engineering, http://geo.citg.tudelft.nl MC model and soil stiffness 37
1
Eoed
1
1(1)
Eoed
1(0)
-1
35
Stiffness parameter – suggestions
G
cu
36
Possibilities and limitations of the
LEPP Mohr-Coulomb model
37
Elastoplastic analysis of a footing
ELASTOPLASTIC ANALYSIS OF A
FOOTING
Computational Geotechnics 1
38
Elastoplastic analysis of a footing
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 20°.
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 γ 0 B ∗ Nγ
0
Nq = eπ tan ϕ tan2 (45 + 12 ϕ0 )
0
q − 1) cot ϕ
Nc = (N
0
2(Nq + 1) tan ϕ
(V esic)
Nγ = 1.5(Nq − 1) tan ϕ0 (Brinch Hansen)
(Nq − 1) tan(1.4 ϕ0 ) (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
39
Elastoplastic analysis of a footing
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
• Geometry input
• Calculation of vertical and horizontal load representing building weight and wind force
A) Geometry input
• General settings
• Mesh generation
B) Calculations
• Construct footing
C) Inspect output
Computational Geotechnics 3
40
Elastoplastic analysis of a footing
GEOMETRY INPUT
Start PLAXIS by double-clicking the icon of the PLAXIS 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).
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.
In order to enter the proper settings for the footing project, follow these steps:
4 Computational Geotechnics
41
Elastoplastic analysis of a footing
• In the Project tabsheet, enter “Exercise 1” in the Title box and type “Elasto-plastic
analysis of drained footing” or any other text in the Comments box.
• In the General options 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.
• The Acceleration box indicates a fixed gravity angle of -90°, which is in the vertical
direction (downward). Independent acceleration components may be entered for pseudo-
dynamic analyses. Leave these values zero and click on the Next button below the
tabsheets or click on the Model tabsheet.
• In the Model tabsheet, keep the default units in the Units box (Length = m; Force = kN;
Time = day).
• In the Geometry dimensions box the size of the considered geometry must be entered.
The values entered here determine the size of the draw area in the 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.
• The Grid box contains values to set the grid spacing. The grid provides a matrix of dots
on the screen that can be used as reference points. It may also be used for snapping to
regularly spaced points during the creation of the geometry. The distance of the dots is
determined by the Spacing value. The spacing of snapping points can further be divided
into smaller intervals by the Number of snap intervals value. Enter 1.0 for the spacing
and 4 for the intervals.
• Click on the Ok button to confirm the settings. Now the draw area appears in which the
geometry model can be drawn.
Computational Geotechnics 5
42
Elastoplastic analysis of a footing
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.
Create sub-soil
• Position the cursor (now appearing as a pen) at the origin (point 0) of the axes (0.0; 0.0).
Click the left mouse button once to start the geometry contour.
• Move along the x-axis to (14.0; 0.0). Click the left mouse button to generate the second
point (number 1). At the same time the first geometry line is created from point 0 to point
1.
• Move upward to point 2 (14.0; 4.0) and click again.
• Move to the left to point 3 (0.0; 4.0) and click again.
• Finally, move back to the origin (0.0; 0.0) and click the left mouse button again. Since
the latter point already exists, no new point is created, but only an additional geometry
line is created from point 3 to point 0. PLAXIS will also automatically detect a cluster
(area that is fully enclosed by geometry lines) and will give it a light colour.
6 Computational Geotechnics
43
Elastoplastic analysis of a footing
Create footing
• Position the cursor at point 4, (6.0, 4.0) and click the left mouse button once.
• Move vertical to point 5, (6.0; 4.25). Click the left mouse button to generate a vertical
line.
• Move horizontal to point 6, (8.0; 4.25). Click the left mouse button to generate a horizontal
line.
• Generate a second cluster by clicking the left mouse button on coordinate (8.0; 4.0).
• Click the right mouse button to stop drawing.
This action created the footing.
The proposed geometry does not include plates, hinges, geogrids, interfaces, anchors or
tunnels. Hence, you can skip the corresponding buttons in the second toolbar.
Hints: Mispositioned points and lines can be modified or deleted by first choosing the
Selection button from the toolbar. To move a point of line, select the point or
the line and drag it to the desired position. To delete a point or a line, select the
point or the line and press the Delete key on the keyboard.
> Undesired drawing operations can be restored by pressing the Undo button
from the toolbar or by selecting the Undo option from the Edit menu or by
pressing <Ctrl><Z> on the keyboard.
Hint: The full geometry model has to be completed before a finite element mesh can be
generated. This means that boundary conditions and model parameters must be
entered and applied to the geometry model first.
Hint: During the input of geometry lines by mouse, holding down the Shift key will
assist the user to create perfect horizontal and vertical lines.
Computational Geotechnics 7
44
Elastoplastic analysis of a footing
Prescribed displacements
Click on the Standard fixities button on the toolbar or choose the Standard fixities option
from the Loads menu to set the standard boundary conditions. As a result PLAXIS will
automatically generate a full fixity at the base of the geometry and roller conditions at the
vertical sides (ux =0; uy =free). A fixity in a certain direction is presented as two parallel lines
perpendicular to the fixed direction. Hence, the rollers appear as two vertical parallel lines and
the full fixity appears as cross-hatched lines.
Hint: The Standard fixities option is suitable for most geotechnical applications. It is
a fast and convenient way to input standard boundary conditions.
Vertical load
Click on the Point load - load system A button on the toolbar or choose the Point load
- static load system A option from the Loads menu to enter another point force. Click on the
coordinate (7.0, 4.25) to enter a point force. As a result PLAXIS will automatically generate a
vertical point force on the indicated point with a unity force (f = 1).
Click on the Point load - load system B button on the toolbar or choose the Point load -
static load system B option from the Loads menu to enter a point force. Click on the coordinate
(7.0, 4.25) to enter a point force. As a result PLAXIS will automatically generate a vertical point
force on the indicated point. As a horizontal force is needed, the direction of load B needs to
be changed.
Choose the Selection button from the toolbar. Double click on the geometry point 8 with
coordinate (7.0, 4.25) which will display a box as indicated in Figure 6. Select Point Load -
load system B, click OK and enter 1.0 as x-value and 0.0 as y-value. These values are the
input load of point force B. Click OK to close the window.
8 Computational Geotechnics
45
Elastoplastic analysis of a footing
& Interfaces, Plates, Anchors and Geogrids. The creation of material data sets is generally
done after the input of boundary conditions. Before the mesh is generated, all material data
sets should have been defined and all clusters and structures must have their appropriate data
set.
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 γunsat 16.0 24.0 kN/m3
Weight below phreatic level γsat 18.0 — kN/m3
Young’s modulus Eref 5.0·103 2.0·107 kN/m2
Poisson’s ratio ν 0.35 0.15 —
Cohesion c 5.0 — kN/m2
Friction angle ϕ 20 — °
Dilatancy angle ψ 0 — °
The input of material data sets can be selected by means of the Material Sets button on
the toolbar or from the options available in the Materials menu.
To create a material set for the clay layer, follow these steps:
• 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.
Computational Geotechnics 9
46
Elastoplastic analysis of a footing
• Enter the proper values for the weights in the General properties box according to the
material properties listed in table 1
• See also figure 8 and figure 9. In these figures the Advanced parameters part has been
collapsed.
Figure 7: General tabsheet of the soil and interface data set window for Clay
• Click on 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.
• 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.
• For the concrete of the footing repeat the former 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).
10 Computational Geotechnics
47
Elastoplastic analysis of a footing
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
Computational Geotechnics 11
48
Elastoplastic analysis of a footing
• Drag the data set “Clay” from the Material Sets window (select it and keep the left mouse
button down while moving) to the soil cluster in the draw area and drop it there (release
the left mouse button). Notice that the cursor changes shape to indicate whether or not
it is possible to drop the data set. When a data set is properly assigned to a cluster, the
cluster gets the corresponding colour. Drag the concrete material set to the footing and
drop it there.
• Click on the OK button in the Material Sets window to close the database.
Hint: PLAXIS distinguishes between a project database and a global database of
material sets. Data sets may be exchanged from one project to another using
the global database. In order to copy such an existing data set, click on the
Show global button of the Material Sets window. Drag the appropriate data set
(in this case “Clay”) from the tree view of the global database to the project
database and drop it there. Now the global data set is available for the current
project. Similarly, data sets created in the project database may be dragged
and dropped in the global database.
Hints: Existing data sets may be changed by opening the material sets window,
selecting the data set to be changed from the tree view and clicking on the Edit
button. As an alternative, the material sets window can be opened by double
clicking a cluster and clicking on the Change button behind the Material set box
in the properties window. A data set can now be assigned to the corresponding
cluster by selecting it from the project database tree view and clicking on the
OK button.
> The program performs a consistency check on the material parameters and will
give a warning message in the case of a detected inconsistency in the data
Mesh generation
When the geometry model is complete, the finite element model (mesh) can be generated.
PLAXIS includes a fully automatic mesh generation procedure, in which the geometry is
automatically divided into elements of the basic element type and compatible structural elements,
if applicable. The mesh generation takes full account of the position of points and lines in the
geometry model, so that the exact position of layers, loads and structures is reflected by
the finite element mesh. The generation process is based on a robust triangulation principle
that searches for optimised triangles, which results in an unstructured mesh. This may look
disorderly, but the numerical performance of such a mesh is usually better than for regular
(structured) meshes. In addition to the mesh generation itself, a transformation of input data
(properties, boundary conditions, material sets, etc.) from the geometry model (points, lines
and clusters) to the finite element mesh (elements, nodes and stress points) is made.
In order to generate the mesh, follow these steps:
• Click on the Generate mesh button in the toolbar or select the Generate option from
the Mesh menu. After the generation of the mesh a new window is opened (PLAXIS
Output window) in which the generated mesh is presented (see Figure 11).
12 Computational Geotechnics
49
Elastoplastic analysis of a footing
Figure 11: Generated finite element mesh of the geometry around the footing
If necessary, the mesh can be optimised by performing global or local refinements. Mesh
refinements are considered in some of the other exercises. Here it is suggested to accept the
current finite element mesh.
Hints: By default, the Global coarseness of the mesh is set to M edium, which is
adequate as a first approach in most cases. The Global coarseness setting
can be changed in the M esh menu. In addition, there are options available to
refine the mesh globally or locally.
> At this stage of input it is still possible to modify parts of the geometry or to add
geometry objects. In that case, obviously, the finite element mesh has to be
regenerated.
Press the close button to close the output program and return to PLAXIS input.
Creating the input for this project now finished. Press the green Calculation button on the
toolbar to continue with the definition of the calculation phases.
Computational Geotechnics 13
50
Elastoplastic analysis of a footing
CALCULATION
After the finite element model has been created, the calculation phases need to be defined.
This analysis consists of four phases. In the initial phase the initial pore pressures and
stresses are generated, in the first phase the footing is constructed, during the second phase
the vertical load is applied and in the third phase the horizontal load is applied.
When starting the PLAXIS Calculation program the Calculation mode window appears. In
this window the user can choose how he wants PLAXIS to handle pore pressures during the
calculation. This is important when calculating with undrained behaviour and/or groundwater
flow. In this first exercise this is not important and so the default setting of Classical mode is
chosen. Press <OK> to close the Calculation mode window. PLAXIS now shows the General
tabsheet of the initial phase (see Figure 12).
14 Computational Geotechnics
51
Elastoplastic analysis of a footing
• Make sure the Calculation type is set to K0 -procedure on the General tabsheet. This is
the default setting.
• On the Parameters tabsheet press the Define button located in the Loading input box.
This will start a window presenting the problem in Staged construction mode. In Staged
construction mode it is possible to switch on and off various parts of the geometry,
change loads, apply strains etc.
• In the initial condition of this exercise, that is the situation before we start constructing
our project, the footing is not present. Therefore the footing has to be deactivated. In
order to do so, click on the area that represents the footing so that it will change color
from the material set color to white. The footing is now disabled.
• Click on Water conditions in the button bar in order to move to the Water conditions
mode of the program.
• Position the cursor (appearing as a pen) at coordinate (0.0, 4.0) and click the left mouse
button to start the phreatic level.
• Move along the x-axis to position (14.0, 4.0). Click the left mouse button to enter the
second point of the phreatic level.
The pore pressures are generated from the specified phreatic level and the water weight.
Directly after the generation, a PLAXIS Output window is opened, showing the pore pressure
as presented in Figure 13. The colors indicate the magnitude of pore pressure. The pore
pressures vary hydrostatically, ranging from 0 kN/m2 at the top to -40 kN/m2 at the bottom.
• Click on Update in order to save the changes made and return to the PLAXIS Calculations
program. This completes the definition of the initial conditions.
Hints: For the generation of initial stresses based on the K0 procedure it is necessary
to specify the coefficient of lateral earth pressure, K0 . This K0 value is defined
per material set and therefore has to be set when entering material set data. If
the K0 value is not explicitly set PLAXIS uses a value according to Jaky’s
formula (K0 = 1-sin(ϕ)).
> The K0 procedure may only be used for horizontally layered geometries with a
horizontal ground surface and, if applicable, a horizontal phreatic level.
Computational Geotechnics 15
52
Elastoplastic analysis of a footing
• Click on the Next button . This will introduce a new calculation phase and
present the corresponding tabsheets for the first calculation stage. Enter a suitable name
in the Number/ ID box (e.g. ‘Construction of footing’).
• Select the second tabsheet called Parameters. On this sheet Staged construction is
selected by default in the Loading input combo box. Click the Define button. This will
open the window presenting the problem in Staged construction mode.
• Click on the cluster that represents the strip footing, in order to switch on the footing
(original colour should reappear).
• Click on Update to conclude the definition of the first calculation phase. Updating will
automatically present the calculation window.
• Select the Parameters tabsheet. On this tabsheet accept the selection Staged construction
in the Loading input combo box. Click on the Define button. This will open the window
presenting the problem in Staged construction mode.
• Click on the point forces in the middle of the footing, a Select items window comes up.
Select the Point load - Load System A to activate point load A and press the Change
button to change the load value. Change the y-value to -50 kN/m and press the Ok
button.
16 Computational Geotechnics
53
Elastoplastic analysis of a footing
• The point load A is now active (blue) and has a load value of 50 kN/m.
• Press Update.
Computational Geotechnics 17
54
Elastoplastic analysis of a footing
• Directly below the Number/ID box select from the Start from phase dropdown list the
second calculation phase. By selecting this the 4th phase will be a continuation of the
2nd phase, hence we will continue to apply the vertical load without having the additional
horizontal load that was applied in phase 3.
• Select the second tabsheet called Parameters. On this sheet choose the selection Total
multipliers in the Loading input group box. Select the third tabsheet called Multipliers by
either clicking on the Define button or directly selecting the tabsheet.
• Enter a ΣMloadA of 10. In this way the working force is increased to a maximum load of
10 x 50 = 500 kN/m.
In PLAXIS two methods exist to increase an active load. The magnitude of the
activated load is the input load multiplied by the total load multiplier. Hence, in
this excersise ΣMloadA x (input load of point load A) = Active load A
The value of the input load A can be changed using Staged construction as
Loading input while using Total multipliers as Loading input may be used to
change the load multiplier.
• Click on the Select points for curves button in the toolbar. This will result in a plot of
the mesh, showing all generated nodes. Click on the node, located in the centre directly
underneath the footing. For a correct selection of this node it may be necessary to use
the zoom option . After selection of the node it will be indicated as point A. Press
the Update button to proceed to calculations.
18 Computational Geotechnics
55
Elastoplastic analysis of a footing
INSPECT OUTPUT
After each successful execution of a calculation phase PLAXIS will indicate the phase with
a green check mark ( ). This indicates a successful calculation phase. If during execution
either failure or an error occurs, PLAXIS marks the stage with a red cross ( ).
• While phase 3 is highlighted, press the View calculation results button that will start
the output program, showing the deformed mesh for the situation with both horizontal
and vertical load applied, as presented in figure 17.
Computational Geotechnics 19
56
Elastoplastic analysis of a footing
• Check the various types of output, such as the deformed mesh, displacement contours,
effective (principal) stresses etc. These can be found from the Deformations and
Stresses menus.
• Still in the Output program, select from the dropdown list at the right of the toolbar the
output step belonging to phase 4.
• From the Displacements menu in the Output program now select Incremental
displacements and then the option |∆u|. Display the incremental displacements as
contours or shadings. The plot clearly shows a failure mechanism (see Figure 18).
• In the Curves manager select the button New to define a new curve. Now the Curve
generation window opens.
• On the x-axis we want to plot the settlement of our chosen point in the middle of the
footing. In the x-axis box choose point A from the dropdown list and then below in
Deformations and then Total displacements choose |u|.
• On the y-axis we want to plot the force applied on the footing, which is a global value
not connected to a specific node or stress point. In y-axis box choose Project from the
dropdown list to indicate we want to plot a global value, and then in Multipliers choose
ΣMLoadA.
20 Computational Geotechnics
57
Elastoplastic analysis of a footing
• Figure 19 shows the Curve generation window after applying the steps mentioned.
The input value of point load A is 50 kN/m and the load multiplier ΣMloadA reaches approximately
4.6. Therefore the failure load is equal to 50 kN/m x 4.6 = 230 kN/m. You can inspect the load
multiplier by moving the mouse cursor over the plotted line. A tooltip box will show up with the
data of the current location.
Computational Geotechnics 21
58
Elastoplastic analysis of a footing
Table 2: Results for the maximum load reached on a strip footing on the drained sub-soil for
different 2D and 3D meshes
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.
22 Computational Geotechnics
59
Elastoplastic analysis of a footing
ADDITIONAL EXERCISE:
UNDRAINED FOOTING
Computational Geotechnics 23
60
Elastoplastic analysis of a footing
24 Computational Geotechnics
61
Elastoplastic analysis of a 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’ 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 νu .
Aims
• The understanding and application of undrained soil behaviour
A) Geometry input
B) Calculations
• Construct footing
C) Inspect output
Computational Geotechnics 25
62
Elastoplastic analysis of a footing
GEOMETRY INPUT
Mesh generation
The mesh generator in PLAXIS allows for several degrees of refinement. In this example
we use the Refine global option from the Mesh menu, which will re-generate the mesh,
resulting in an increased number of finite elements to be distributed along the geometry lines.
Notice the message that appears about staged being reconstructed: the program will take into
account the newly generated mesh for the previously generated initial conditions and staged
construction phases. From the output window, in which the mesh is shown, press the continue
button to return to the Input program.
Hint: After generation of a finer mesh, the geometry may be refined until a
satisfactory result appears. Besides the option Refine global several other
methods of refinement can be used.
26 Computational Geotechnics
63
Elastoplastic analysis of a footing
Hint: After re-generation of the finite element mesh new nodes and stress points
exists. Therefore PLAXIS has to regenerate pore water pressures and initial
stresses. This is done automatically in the background when regenerating the
mesh. Also, the new mesh is taken into account for any change to calculation
phases with the exception of ground water flow analysis.
After generating the mesh one can now continue to the calculation program. Click on the
Caculations button to proceed to the calculations program. Click ‘yes’ to save the data.
CALCULATIONS
• 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.
INSPECT 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 results normally, after
loading, in excess pore pressures. The excess pore pressures may be viewed in the output
window by selecting:
• Select in the calculation program the phase for which you would like to see output results.
• Start the output program from the calculation program by clicking the View output button .
• Select from the Stresses menu the option Pore pressures and then pexcess , this results in
Figure 22 .
The excess pore pressures may be viewed as stress crosses ( ), contour lines ( ),
shadings ( ) 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.
Computational Geotechnics 27
64
Elastoplastic analysis of a footing
Figure 22: 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 23.
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 approx. 70 kPa.
28 Computational Geotechnics
65
Elastoplastic analysis of a footing
Computational Geotechnics 29
66
Elastoplastic analysis of a footing
30 Computational Geotechnics
67
Elastoplastic analysis of a footing
Qf
B
= c · Nc + 12 γ 0 B · Nγ
0
Nq = eπ tan ϕ tan2 (45 + 12 ϕ0 )
0
q − 1) cot ϕ
Nc = (N
0
2(Nq + 1) tan ϕ
(V esic)
Nγ = 1.5(Nq − 1) tan ϕ0 (Brinch Hansen)
(Nq − 1) tan(1.4 ϕ0 ) (M eyerhof )
γ 0 = γw − 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 γ 0 B · Nγ = 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 )
Qf
L=
B
III
I
II
Computational Geotechnics 31
68
Non-linear calculations
in PLAXIS
Non-linear calculations 1 / 27
Content
– Learning objectives
– Introduction
– Multipliers
– Iteration process
– Plastic points
– Recommendations
Non-linear calculations 2 / 27
69
Learning objectives
Non-linear calculations 3 / 27
Introduction
• Load multipliers
Non-linear calculations 4 / 27
70
Introduction
• Load multipliers
• Miscellaneous parm’s
Non-linear calculations 5 / 27
Introduction
• Load multipliers
• Miscellaneous parm’s
• Load-displ. curve
Non-linear calculations 6 / 27
71
Introduction
• Load multipliers
• Miscellaneous parm’s
• Load-displ. curve
• Iteration process
Non-linear calculations 7 / 27
Introduction
• Load multipliers
• Miscellaneous parm’s
• Load-displ. curve
• Iteration process
• Plastic points
Non-linear calculations 8 / 27
72
Load multipliers
Applied load = Load multiplier x Input load
Defaults:
• Load multiplier = 1
• Input load = 1 unit
Loading input:
• Staged construction: Change Input load
• Total multipliers: Change Load multiplier (M…)
• Incremental multipliers: Change Load multiplier (M…)
Non-linear calculations 9 / 27
Load multipliers
MdispX : Tot. mult. prescribed x-displacements
MdispY : Tot. mult. prescribed y-displacements
MloadA : Tot. mult. loads system A
MloadB : Tot. mult. loads system B
Mweight : Tot. mult. soil & structural weights
Maccel : Tot. mult. pseudo-static acceleration
Msf : Tot. mult. Phi-c reduction process
Mstage : Tot. mult. staged-construction process
Non-linear calculations 10 / 27
73
Load multipliers – Incremental multipliers input
Non-linear calculations 11 / 27
Load multipliers – Total multipliers input
Non-linear calculations 12 / 27
74
Miscellaneous parameters
PMax : Maximum (excess) pore pressure in the model
ΣMarea : Relative part of the mesh area currently active
Force-X : Reaction force due to horizontal prescr. displ.
Force-Y : Reaction force due to vertical prescribed displ.
Stiffness : Current (relative) Stiffness Parameter
Time : Elapsed model time (usually in days)
Dynamic time : Elapsed model time for dynamics (s)
Non-linear calculations 13 / 27
Load‐displacement curve
– Evaluation of calculation progress:
• Multipliers
• Stiffness (CSP)
• Pmax
• Load-displacement curve
• Iterations
• Global error
• Plastic points
Non-linear calculations 14 / 27
75
Iteration process
Calculation phase
Load steps (q)
Equilibrium iterations
constitutive model
q displacement strain stress reaction
Equilibrium?
Non-linear calculations 15 / 27
Iteration process
Load q
Elastic stiffness (K)
qex Non-linear
iterations behaviour
Unbalance
Load step q
qin
Settlement of Node A
Non-linear calculations 16 / 27
76
Iteration process
Current step ≤ Max. step Additional steps
Iteration ≤ Max. iterations Maximum iterat.
Unbalance Global error ≤ Tolerance Tolerated error
Control parameters
Non-linear calculations 17 / 27
Iteration process – Control parameters
Non-linear calculations 18 / 27
77
Iteration process – Over‐relaxation
A B
Overrelaxation
A
uo u
Non-linear calculations 19 / 27
Iteration process – Arc‐length control
Non-linear calculations 20 / 27
78
Iteration process – Arc‐length control
qP qP
0 0+ +
1 P = I Pe – Pc I = const.
K
Pq0
P0 0 Arc length control
uo u
Non-linear calculations 21 / 27
Iteration process – Desired minimum / maximum
Non-linear calculations 22 / 27
79
Iteration process – Desired minimum / maximum
q
• Converged within desired minimum Scaling up
number of iterations:
– Scaling up load step by a
factor 2 Scaling down
• Not converged within desired
maximum number of iterations:
– Scaling down load step by a
Scaling up
factor 2
Non-linear calculations 23 / 27
Plastic points
- 1 Cap (HS, SS and SS-Creep model)
Cap point
Mohr-Coulomb
failure surface
Mohr-Coulomb point
f<0
Shear hardening Cap & Hardening point
yield surface (HS model)
Hardening point
-3
Apex point -1
Tension point
-3
Tension cut-off: Principal tensile stress is set to zero
Non-linear calculations 24 / 27
80
Plastic points
Local error criterion: Constitutive stress c:
Stress that follows from the constitute
model (Mohr- Coulomb)
eq
Equilibrium stress eq:
Stress that is in equilibrium with the
c external load
|| c eq ||
Local Error
|| c ||
Inaccurate point:
Local error > Tolerated error
Convergence requirement:
Inaccurate stress points ≤ 3 + (plastic soil points) /10
Inaccurate interface points ≤ 3 + (plastic interface points) /10
Non-linear calculations 25 / 27
Recommendations
– Use mostly defaults
– Monitor and evaluate calculation progress
– In case of bad convergence or numerical failure, check input
– Use output facilities to trace input errors
– In case input is right, consider control parameters
– Don’t change control parameters without understanding
consequences!
– Don’t increase tolerated error to speed up convergence!
Non-linear calculations 26 / 27
81
Non-linear calculations 27 / 27
82
Hardening Soil Model
William Cheang
Notes by:
Professor Helmut Schweiger ( TU Graz)
Professor Pieter Vermeer
A/Professor Tan Siew Ann (NUS)
83
INTRODUCTION
INTRODUCTION
84
85
86
87
88
89
LINES OF EQUAL SHEAR STRAINS (Tatsuoka & Ishihara, 1974)
90
Do you need plasticity when unloading (back into the yield locus)?
Yes..if the accumulation plastic volumetric strains are important in cyclically loaded
soils..dynamic liquefaction related boundary value problems
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
Surface Heave in Initial Exc./Cantilever Wall
3 m deep excavation with cantilever wall
20kPa
5m
3m
7m
• 3 analyses with Mohr Coulomb, Hardening Soil & Hardening Soil-Small models using equivalent
soil input parameters
112
Pre-failure Stress-strain Behaviour
1: Mohr Coulomb
1: Linear elastic, perfectly plastic
2: Hyperbolic stress-strain curve
(stiffness degradation for > 1E-4)
3: Non-linear stiffness from very
small strains (1E-6)
0.004 Heave
0.002
Settlement (m)
0.000
-0.002
-0.004
Settlement
-0.006
MC
-0.008 HS
HSsmall
-0.010
113
Predicted Heave at Exc. Level in Cofferdam
Distance in front of wall (m)
-5 -4 -3 -2 -1 0 1 2 3
0.025
MC
HS
Wall
0.020 HSsmall
0.015
Heave (m)
0.010
0.005
0.000
-0.005
MC HS HSsmall
Ux=6mm
Ux=11mm Ux=10mm
114
Predicted Stability of Wall
3 FOS=2.8
2.5
MC Rotation mechanism
2 with FOS 2.8
1.5
3 FOS=2.8
2.5
2
HS
1.5
• “Phi-c' reduction” for predicting FOS
3 FOS=2.8
• FSP III sheetpile properties:
2.5 EI=34440 kNm2/m; EA=3.92×106kN/m
2 HSsmall
Mp=369 kNm/m; Np=3575 kN/m
1.5
Summary of Predictions
Analyses Surface settlement Heave at Wall horizontal FOS for wall
behind wall excavation level displacement stability
MC Heave 4 mm Heave 20 mm 6 mm 2.8
(not OK)
HS Settle 9 mm Heave 11 mm 11 mm 2.8
HSsmall Settle 9 mm Heave 8 mm 10 mm 2.8
115
Variation of Soil Stiffness in Excavation
A. Soil stiffness is not constant and varies with
1. stress-level. Higher stress, higher stiffness
2. strain-level. Higher strain (or displacement), lower stiffness
3. stress-path (recent soil stress history). Rotation of stress path,
higher soil stiffness
4. anisotropy, destructuration
B. During excavation, soil elements at different locations experience
different changes in
1. stress,
2. strain
3. stress-path direction
• A: unloading compression;
•B: unloading extension
• Rotation of stress paths at A & B
116
Soil Stress Paths Near Excavation
20kPa 25 20kPa
Failure line
20
3m A K0
15 A
Exc. A
B 10
B
Exc.
t (kPa)
7m K0 20kPa B
5
5m
-5
-10
Failure line
A: unloading compression
-15
B: unloading extension 0 10 20 30 40 50 60
s' (kPa)
°
Stress path rotation,
t °
=0°
°
=180°
K0
=90°
s'
Atkinson et al. (1990) °
Triaxial tests on
London Clay
Shear strain (%)
-1 -0.1 -0.01 0.01 0.1 1
=0°, no change in stress path direction
=180°, full reversal of stress path direction
117
Stress Path Dependent CDG Stiffness
Stress-level Test series
Extension
Compress
Compression
Extension
=90°
118
APPENDIX
73
Characteristics:
119
1.Hardening Soil model
Hyperbolic stress-strain relationship in (tri)axial loading:
(Duncan-Chang model)
q
1 qult
q
1 / E0 1 / qult E0
Rf qult
q / E0 Eur
1
1 q / qult
E0 = initial stiffness 1
qult = asymptotic value of q (related to strength)
Rf = ‘failure ratio’ (standard value 0.9)
120
3.Shear hardening in the HS model
Elastoplastic formulation of hyperbolic q-1 relationship:
1 q 2q
f* p
2 1p vp
E50 1 q qa Eur
m m
c cot ' '3 c cot ' '3
E50 E ref
Eur E ref
c cot ' p c cot ' p
50 ref ur ref
1 q 2q
f* p
2 1p vp
E50 1 q qa Eur
m m
c cos ' '3 sin ' c cos ' ' 3 sin '
E50 E ref
Eur E ref
c cos ' p sin ' c cos ' p sin '
50 ref ur ref
121
5.Shear hardening in the HS model
Elastoplastic formulation of hyperbolic q-1 relationship:
Elastic
q MC failure line q MC failure line
plastic
m 3p,fric
2p,fric
1p,fric
p’
p’
122
7.Shear hardening in the HS model
Flow rule: d vp , fric d p , fric
sin m with:
sin m sin cv
sin m
q MC failure line 1 sin m sin cv
m>0 sin ' sin
m
cv sin cv
1 sin ' sin
m<0 (in principle)
Note:
m < 0 is not taken
into account
p’
8.Compaction hardening
q
in the HS model
MC failure line
Yield function (associated):
q2
f cap p 2 pc2
2
Cap
Hardening rule: pc
1 m
pc fc = 0
vp ,cap
1 m p ref pc p’
is determined by K0 nc c 1
is determined by Eoed
m
ref c cot ' '1
Eoed Eoed ref
c cot ' p v
123
9.Compaction and Shear hardening in
the HS model
Cap
Cone
124
11.Small-strain stiffness in the HS
model (HSsmall)
Strain(path)-dependent elastic overlay model:
G0
Gs
1 0.385 / 0.7
G0
1 0.385 / 0.7 2
Gt Gur
G starts again at G0
Gur after full strain reversal
Cyclic loading
Gt
leads to Hysteresis
G0 Gs
Energy dissipation
-c
Damping
+c
G0 G0
125
13.Small-strain stiffness in the HS
model (HSsmall)
G0 0.7
Gt Gs
Gur
126
Parameters of the HS(small) model
Parameters:
E50ref Secant stiffness from triaxial test at reference pressure
Eoedref Tangent stiffness from oedometer test at pref
Eurref Reference stiffness in unloading / reloading
G0ref Reference shear stiffness at small strains (HSsmall only)
0.7 Shear strain at which G has reduced to 70% (HSsmall only)
m Rate of stress dependency in stiffness behaviour
pref Reference pressure (100 kPa)
ur Poisson’s ratio in unloading / reloading
c’ Cohesion
’ Friction angle
Dilatancy angle
Rf Failure ratio qf /qa like in Duncan-Chang model (0.9)
K0nc Stress ratio ’xx/’yy in 1D primary compression
q
3=pref c 1=pref 1
qult
(, c)
E50ref qf=Rf qult
Eurref
0.5 qf 1
Eoedref
1 v
Triaxial test Oedometer test
127
Parameters of the HS model
105
Eoed [MPa] for NC-soils and ´ =
Janbu :
103 m
Eoed Eref
oed
sandy gravel pref
102
sand
10
more general:
m
Norwegian a
Eoed E ref
100 kPa
oed
1 clays pref a
Mexico City Clay with a = c´ cot´
0.1
0 50 100
porosity n [%]
ref
Eoed 1
2
ref
E50 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 * Relationship with Soft Soil model
128
Parameters of the HS model
ref
Eoed E50
ref
Order of magnitude by Schanz
Correlation by Lengkeek
ref
Eoed RD 60 MPa for pref=100 kPa
Schanz (1998)
129
Parameters of the HS model
Eurref (3 to 5) E50
ref
Eurref
G ref
(2.5 to10)G ref
where ref
G
2(1 ur )
0 ur ur
0.7 (1 to 2) 10 4
130
Initial conditions for the HS model
’yy0 c ’yy0 c
Initial stresses:
’yy
’yy0 follows from soil weight
Prestress and pore pressure
’c
Initial
CAP ’xx0 = K0 ’yy0
POP 1
ur
’yy0 1 ur
Initial stress
K 0nc '0yy POP ur
1 ur
POP
1 K0
K0nc
'0yy
ur
’xx0 ’xx K 0 OCR K 0nc OCR 1
1 ur
131
Initial conditions for the HS model
Initial stresses:
q
MC failure line
K0nc line
Output:
pc Cap
pc
' OCR ' OCRiso
p’0, q0 p eq
132
Comparison HS model and MC model
Isotropic compression test:
Custom
1000 MC
HS.vlt
900
800
700
p' [kN/m²]
600
500
400
300
200
100
0
0 0.01 0.02 0.03 0.04 0.05 0.06
v
MC
HS.vlt
200
| 1 - | [kN/m²]
3
100
0
0 -0.01 -0.02 -0.03 -0.04 -0.05 -0.06 -0.07 -0.08
1
133
Comparison HS model and MC model
Drained triaxial test at 3=100 kPa :
Custom
0.009 MC
HS.vlt
0.006
0.003
v
-0.003
400
| 1 - | [kN/m²]
300
3
200
100
0
0 -0.01 -0.02 -0.03 -0.04 -0.05 -0.06 -0.07 -0.08
1
134
Comparison HS model and MC model
Drained / undrained triaxial test at 3=100 kPa :
E1DS
500
MC(u)
HS(u).vlt
MC.vlt
400 HS.vlt
| 1 - | [kN/m²]
300
3
200
100
0
0 -0.01 -0.02 -0.03 -0.04 -0.05 -0.06 -0.07 -0.08
1
300
q [kN/m²]
200
100
0
0 -100 -200 -300
p' [kN/m²]
135
Comparison HS model and MC model
One-dimensional compression test (oedometer):
Custom
1000 MC
HS.vlt
900
800
700
' 1 [kN/m²]
600
500
400
300
200
100
0
0 0.005 0.01 0.015 0.02 0.025 0.03
1
One-dimensional
MC
HS.vlt
-1000
(oedometer):
-800
-700
' 1 [kN/m²]
-600
-500
-400
-300
-200
-100
0
0 -200 -400
' 3 [kN/m²]
136
Comparison HS model and MC model
One-dimensional compression test (oedometer):
Stress state after unloading
HS MC
137
Hardening Soil model
138
Examples of parameter selection
200
150
dense 40
100
50
pref a
E50ref E50
0 'x a
0 1 2 3 4 5
100kPa
Axial strain [%] 20000kPa 32MPa
40kPa
Dense 40
-3 1-sin 1.2
-2
2sin Dense 40 2 sin 4.2
sin 0.29
-1
0 1 2 3 4 5
0
1 17
139
Examples of parameter selection
0 1 ' a
Eoed Eoed
ref
-0,2
Test data pa ' a
Vertical strain [%]
-0,4
-0,6
pa ' a
-0,8 ref
Eoed Eoed
-1 1 ' a
-1,2
850kPa 100kPa
-1,4 53MPa
0 200 400 600 800 1000 1200 0.008 400kPa
Vertical effective stress [kPa]
140
Examples of parameter selection
200
180
Deviatoric stress, q [kPa]
160
140
120
Plaxis 40
100 dense 40
80
60
40
20
0
0,00 1,00 2,00 3,00 4,00 5,00
Axial strain [%]
-3
-2,5
-2 from PLAXIS 40
Dense 40
-1,5
-1
-0,50,00 1,00 2,00 3,00 4,00 5,00
0
0,5
1
141
Examples of parameter selection
0
Test data
-0,2
Vertical strain [%]
-0,4 Plaxis
-0,6
-0,8
-1
-1,2
-1,4
0 200 400 600 800 1000 1200
Vertical effective stress [kPa]
142
Simulation of laboratory tests
Computational Geotechnics 1
143
Simulation of laboratory tests
2 Computational Geotechnics
144
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
• Simulation of laboratory tests
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.
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.
Computational Geotechnics 3
145
Simulation of laboratory tests
4 Computational Geotechnics
146
Simulation of laboratory tests
Computational Geotechnics 5
147
Simulation of laboratory tests
6 Computational Geotechnics
148
Simulation of laboratory tests
Figure 6: Undrained triaxial (CU) tests at cell pressures of 100 kPa and 400 kPa
Computational Geotechnics 7
149
Simulation of laboratory tests
8 Computational Geotechnics
150
Simulation of laboratory tests
For a cell pressure σ30 = 100 kPa a maximum value of approximately |σ10 − σ30 | = 400 kPa is reached at failure.
The Mohr-Coulomb failure criterium is:
1 0
2 |σ1 − σ30 | + 12 (σ10 + σ30 ) · sinϕ − c · cosϕ = 0
Considering it is sand we assume that the cohesion is zero and so the Mohr-Coulomb failure criterium reduces
to:
|σ10 −σ30 |
(σ10 +σ30 ) = sinϕ
Filling in σ30 = 100 kPa and σ10 = 500 kPa as obtained from the test we find for the
friction angleϕ0 = 420
The triaxial test stiffness E50 is the secant stiffness over the first 50% of the failure value for | σ10 − σ30 |. This is
indicated in red in the triaxial test graph of figure 8.
σ 0 =100 kP a 400
E503 = 0.013 = 30800 kP a
Computational Geotechnics 9
151
Simulation of laboratory tests
The triaxial test stiffness ,E 50 , is within the Hardening Soil model defined as:
m m
c cosϕ−σ30 sinϕ σ30
ref ref
E50 = E50 c cosϕ+pref sinϕ , c = 0 ⇒ E50 = E50 − pref
The reference stress pref is chosen equal to the cell pressure of this triaxial test then
ref σ 0 =100 kP a
E50 = E503 ≈ 30000 kPa
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.
σ 0 =100 kP a 400
Eur3 = 0.026−0.021 = 80000 kPa
Under the same assumptions as for the stiffness in triaxial testing counts:
ref σ 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
∆εv
sinψ = −2∆ε1 +∆εv
10 Computational Geotechnics
152
Simulation of laboratory tests
From the oedeometer test results we determine the stiffness Eoed for vertical stresses σy0 = 100 kPa en σy0 =
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.
σ 0 =100 kP a
y 320−0
Eoed = 1.4%−0.33% = 29900 kPa
σy0 =200 kP a 400−0
Eoed = 1.4%−0.47% = 43000 kPa
Within the Hardening Soil model the stress dependent oedometer stiffness is defined as:
m m
c cosϕ−σy0 sinϕ σy0
ref ref
Eoed = Eoed c cosϕ+pref sinϕ , c = 0 ⇒ Eoed = Eoed − pref
ref σ 0 =100 kP a
Eoed = Eoed
3
≈ 30000 kPa
The power m for stress dependent stiffness can now be determined as:
σ 0 =200 kP a m
y
σy0
Eoed 43000 200 m
ref
Eoed
= pref ⇒ 30000 = 100 ⇒ m = 0.5
Computational Geotechnics 11
153
Simulation of laboratory tests
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
∆σx ∆σ30 100
K0N C = ∆σy0 = ∆σ10 = 300 = 0.33
12 Computational Geotechnics
154
Simulation of laboratory tests
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 σ3 while the oedometer test results presented in figure 10 give the strain vs
the vertical stress σy (= σ1 voor oedometer testing).
ref σ 0 =100 kP a
Eur = Eur3 ≈ 110000 kPa
ref
This is a bit high and so a value of Eur = 90000 kPa is chosen.
Computational Geotechnics 13
155
Simulation of laboratory tests
CLAY
We start with the determination of the strength parameters based on the CU triaxial tests.
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:
With p0 = (σ10 + 2σ30 )/3 and q = σ10 − σ30 under triaxial test conditions this can be rewritten as:
2p0 + 13 q
q 6sinϕ 0 6c cosϕ
2 = 2 sinϕ − c cosϕ = 0 ⇒ q = 3−sinϕ p + 3−sinϕ
Hence, the slope M of the Mohr-Coulomb line in p’-q plane is defined as:
6sinϕ 195
M= 3−sinϕ = 200 ⇒ ϕ = 250
From the intersection between Mohr-Coulomb line and the vertical axis where p=0 the cohesion can be determined:
6c0 cosϕ
q= 3−sinϕ = 0 ⇒ c = 0 kPa
14 Computational Geotechnics
156
Simulation of laboratory tests
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.
Considering that both primary loading and unloading/reloading paths are straight lines in the log(p)-εv graph,
hence they have a relation of the form:
εy = εv = A · log(σy0 )
ε2 −ε1 0.370−0.270
A= log(σ2 )−log(σ1 ) = 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(σy0 )
εy = εv = A · ln(10)
dεy dσy0 ln(10)
dσy0 =A· 1
ln(10) · 1
σy0 ⇒E= dεy = A · σy0
σy0
ln(10)
E = Eoed = A · pref − pref
In the Hardening Soil model the oedometer stiffness is defined as (assuming c = 0) :
m
σy0
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:
Computational Geotechnics 15
157
Simulation of laboratory tests
εy = εv = B · log(σy0 )
ε2 −ε1 0.427−0418
B= log(σ2 )−log(σ1 ) = log(120)−log(30) =0.0149
dσy0 ln(10)
Eur = dεy = B · σy0
However, the Eur in the Hardening Soil model is dependent on the smallest principal stress, which is σx0 in an
oedeometer test and not σy0 .
During the unloading process there is no linear relation between horizontal and vertical stress, as in the beginning
of unloading σy0 > σx0 where as after much unloading σy0 < σx0 . Therefore the assumption is made that during
unloading on average σx0 = σy0 .
0
ln(10) ln(10) ln(10) σx
Eur = B · σy0 = B · σx0 = B · pref pref
0
m
ref σx
Eur = Eur − pref
ref
Follows, in a similar way as for the Eoed , that
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.
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
Jaky’s rule:
Poisson’s ratio
The Poisson’s ratio for unloading and reloading is again estimated as νur = 0.2
16 Computational Geotechnics
158
Simulation of laboratory tests
Computational Geotechnics 17
159
Simulation of laboratory tests
18 Computational Geotechnics
160
Simulation of laboratory tests
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.
Computational Geotechnics 19
161
Simulation of laboratory tests
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 .
In the main window select the Oedometer tabsheet and set the parameters as indicated in Figure .
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.
20 Computational Geotechnics
162
Simulation of laboratory tests
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.
Computational Geotechnics 21
163
Simulation of laboratory tests
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 σxx , σyy and σzz 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: τxy = 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 (∆σyy ) until the stress level where the unloading path should start.
Note that the horizontal stresses (∆σxx and ∆σzz ) 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
22 Computational Geotechnics
164
Simulation of laboratory tests
Figure 20: Unloading/reloading cycle in a triaxial test using the General option
Computational Geotechnics 23
165
Derivation of Soil Parameters from
Lab Test Results & Verification in
Plaxis SoilTest
by
RF Shen
23 Nov 2011
Singapore 2011
166
In this exercise, we are going to fully utilize the test
data to derive soil parameters for Hardening Soil
Parameters from most common stress-strain data
provided in a typical SI report, and subsequently
use Plaxis SoilTest to verify the derived parameters
Part 1: Sand
Singapore 2011
167
For sand, one of the most common lab tests is
Triaxial Isotropically Consolidated Drained (CID)Test
Fa/A = q (deviatoric
stress)
Typical sample size 38 mm Ø x 76 mm a = q + r
Singapore 2011
168
450
400
350
Deviator stress (kPa)
300
250
100 Test data
50
0
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07
Axial strain
0.06
0.05
0.04
Volumetric strain
0.03
0.02
0.01
Test data
0
0 0.02 0.04 0.06 0.08 0.1
‐0.01
Axial strain
Singapore 2011
169
Another common lab test is Oedometer Test
Settlement dial gauge
Oedometer Cell
Sample: dia. =75mm
Protruded
lever arm Height = 20mm
Heavy dead weights
Singapore 2011
170
Another common lab test is Oedometer Test
Boundary
conditions
0
0.1
0.2
0.3
0.4
Vertical strain (%)
0.5
0.6
0.7 Test data
0.8
0.9
1
1.1
1.2
1.3
1.4
0 100 200 300 400
Vertical pressure (kPa)
Singapore 2011
171
400
300
Vertical pressure (kPa)
200
100
Test data
0
0 50 100 150 200
Lateral stress (kPa)
Singapore 2011
172
Part 1: Strength parameters
450
400
Since c’ = 0 for sand, it can be
350
simplified to:
Deviator stress (kPa)
300
250
100 Test data
50
0
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07
500 100
sin ' 0.67
Axial strain
500 100
' 42
Singapore 2011
173
Part 1: Strength parameters
0.06
0.05
So,
0.048
0.04
Volumetric strain
0.004
Test data sin 0.27
0
0 0.02 0.04 0.06 0.08 0.1
0.03 0.09
16
‐0.01
Axial strain
pc p’
Singapore 2011
174
Part 2: Stiffness parameters
450
400 400
350
Deviator stress (kPa)
300
250
3’ = 100 kPa
200
150
100 Test data
50
0
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07
0.013
Axial strain
400
E50ref 30800 kPa 30000 kPa
0.013
Singapore 2011
175
Part 2: Stiffness parameters
450
400400
350
Deviator stress (kPa)
300
200
150
100 Test data
50
0
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07
0.021 0.026
Axial strain
400
Eurref 80000 kPa
0.026 0.021
Singapore 2011
176
Part 2: Stiffness parameters
0
0.1
0.2
0.3
0.33
0.4
Vertical strain (%)
0.5
0.6
0.7 Test data
0.8
0.9
1
1.1
1.2
1.3
1.4
0 100 200 300 320 400
Vertical pressure (kPa)
320
ref
Eoed 29900kPa 30000kPa
1.4% 0.33%
m
ref c cot ' '1
Eoed Eoed ref
c cot ' p
Singapore 2011
177
Part 2: Stiffness parameters
0
0.1
320
29900kPa 30000kPa
0.2 ref
Eoed
0.3
0.4
1.4% 0.33%
0.47
Vertical strain (%)
0.5
400
0.6 200 kPa
Eoed 43000kPa
0.7
0.8
Test data
1.4% 0.47%
0.9
1
1.1
1.2
1.3
1.4
0 100 200 300 400
Vertical pressure (kPa) 400
m
c cot ' '1 200
200 kPa m
Eoed 43000
m = 0.5
c cot ' p
ref ref
Eoed 100 30000
Jaki’s formula:
300
Vertical pressure (kPa)
200
100
Test data
0
0 50 100 150 200
Lateral stress (kPa)
x ' 100
K 0NC 0.33
y ' 300
Singapore 2011
178
Summary of Hardening Soil Parameters
Singapore 2011
179
Part 2: Clay
Singapore 2011
180
For Clay, one of the most common lab tests is Triaxial
Isotropically Consolidated UnDrained (CIU)Test
Fa/A = q (deviatoric
stress)
Close the valve = Undrained test =
a = q + r
Excess will accumulate with shearing
350
Test data
300
250
q (kPa)
200
195
150
100
50
0
0 50 100 150 200 250 300 350 400 450
p' (kPa)
Singapore 2011
181
CIU stress path
Gradient:
350
Test data
300
200
195
150
100 ’ = 25
50
0 Intercept:
0 50 100 150 200 250 300 350 400 450
p' (kPa)
c’ = 0
Singapore 2011
182
Another common lab test is Oedometer Test
Test data
0.1
Vertical strain (%)
0.2
0.3
0.4
0.5
1 10 100 1000
Vertical pressure (kPa)
Singapore 2011
183
Oedometer test for clay
0
d y '
Test data
Eoed
0.1
d y
Vertical strain (%)
0.2
d (log y ' )
Gradient _ k
d y
0.3
0.5
1 10 100 1000
Vertical pressure (kPa)
ln y ' 1
d ( y ' )
d (log y ' ) d( )
2.3 1 y' 1 d ( y ' ) 1
Gradient _ k Eoed
d yy d yy 2.3 d yy 2.3 y ' d yy 2.3 y '
Test data
0.1
ref
Eoed 2.3 100 6.02 1350 kPa
Vertical strain (%)
0.2
Eoed y '
0.27 ref
0.3
Eoed pref
0.37 m
0.4
ref c cot ' '1
Eoed Eoed ref
0.5
c cot ' p
1 10 30 100 1000
120 m
Eoed '1
Vertical pressure (kPa)
ref
Eoed 2.3 y ' gradient _ k
ref
Eoed p
gradient _ k
log(120) log(30)
6.02
m=0
0.37 0.27
Singapore 2011
184
Oedometer test for clay
Eur refers to when 3’ =
0
100kPa
Test data
0.1
0.3 x’<100kPa;
When y’ loaded to about
0.418
0.4
300kPa and unload to
0.427
0.5 100kPa, x’ is expected to be
30
closer to 100kPa. As such, we
1 10 100 1000
Vertical pressure (kPa) 120
can approximately accept the
derived Eur.
log(120) log(30)
gradient _ k 66.9 Eur 2.3 100 66.9 15000 kPa
0.427 0.418
80
3’ = 100kPa for
consolidation,
During shearing, 3 = 0
70
60
Excess pore pressure
Deviator stress (kPa)
50 accumulates during
40 shearing 3’ 100kPa
3’ = 100 kPa
30
Typically for NC clay, E50ref
20
may be about
Test data 2~5 times
10 Eoedref or about 2800kPa~7000kPa.
Trial runs to fit the test data gives
0
0 0.01 0.02 0.03 0.04 E50ref =0.06
0.05 3500kPa
0.07
Axial strain
Singapore 2011
185
Summary of Hardening Soil Parameters
Singapore 2011
186
Let’s call it a day!
Thank you!
Singapore 2011
187
Mesh & Geometry Selection
1/35
Contents
– Plane strain, Axi-symmetry, 3D
– Model boundaries
• General considerations
• Excavations
• Shallow foundations
• Embankments
• Tunnels
– Conclusions
– References
188
Plane strain
Considerations:
– One dimension is relatively long
– Similar geometry and stress or loading conditions in any cross
section ‘long’ dimension
Consequences:
– No strain ‘long’ dimension (stress can change!)
– No shear stress and arching ‘long’ dimension
– Model represents 1 length unit ‘long’ dimension
y
Plane strain
Examples:
189
Plane strain
NOT a plane-strain situation:
45 m
30 m
45 m
8m
Axi-symmetry
Considerations:
– Geometry is circular
– Similar geometry and stress or loading conditions in any cross
section that includes the central axis
Consequences:
– Stress and strain central axis are radial
– Model represents 1 radian around central axis
190
Axi-symmetry
Examples:
Axi-symmetry
NOT an axi-symmetric situation:
Gravity!
191
3D models
Considerations:
– Do I really need a 3D model?
– If yes, but still I use a 2D model:
• What are the consequences?
• Would this give conservative or optimistic results?
• How large is the error?
3D models
192
3D models
3D models
193
3D models
Model boundaries
General considerations
194
Model boundaries
Stability analysis:
Model boundaries
Deformation analysis:
drained undrained
195
Model boundaries
Dynamic analysis:
~
~ ~ ~ ~
Model boundaries
Stability analysis
Drained
deformation analysis
Undrained
deformation analysis
~
~ ~ ~ ~
Dynamic analysis
196
Model boundaries – Shallow foundations
a w a
limit
w depth
(0.1 to 0.2) accepted
197
Model boundaries – Embankments
a w a
h
198
Model boundaries – Excavations
a w a
l d
½a
Use HSsmall or bottom layer with small-strain stiffness for Eurref (height ½ a)
199
Model boundaries – Excavations
Considering the wall:
TBM or
NATM excavation
w
D D
a a
½a
Suggestions: Face stability: a ½D ; w 2D
Structural analysis: a ½D ; w 2D
Deformation analysis: a D ; w 3D
Use HSsmall or bottom layer with small-strain stiffness for Eurref (height ½ a)
200
Model boundaries – Tunnels
Take account of the following:
Meshing
• Type of element:
Two types of volume elements are available in Plaxis 2D:
node (ux, uy)
201
Meshing
• Type of element (2D):
Which type of element in which situation?
Meshing
• Type of element 3D: 3
12
1
9
8
6
1
4 7
10
2
15 2
3
14
5
4 6 11
13
5
3DT, 3DF: 15-node wedge New Plaxis 3D: 10-node tetrahedral
(quadratic interpolation) (quadratic interpolation)
Do not confuse 15-node wedge in 3D (quadratic) with 15-node triangle in 2D (4th order)!
202
Meshing
General considerations:
Hint:
• Use local element size factors to make meshes fine near loads and
structures and coarse at model boundaries (local element size
factor may be larger than 1.0!).
Meshing
203
Conclusions
Conclusions:
• Model size and boundaries depend, a.o., on type of analysis and type of
behaviour (stability analysis, drained deformation undrained deformation,
dynamic analysis).
References
• Potts D.M., Zdravkovic L. (2001). Finite element analysis in geotechnical engineering –
Application. Thomas Telford, London.
• Schweiger H.F. (2002). Musterlösung und Parameterstudie für dreifach verankerte Baugrube,
Geotechnik 25, 101-109.
• Ruse N.M. (2003). Räumliche Betrachtung der Standsicherheit der Ortsbrust beim
Tunnelvortrieb. PhD thesis. Institut für Geotechnik. Universität Stuttgart.
• Vermeer P.A., Wehnert M. (2005). Beispiele von FE-Anwendungen – Man lernt nie aus. In:
FEM in der Geotechnik (ed. Grabe et.al.). Technische Universität Hamburg-Harburg.
• Brinkgreve R.B.J, Bakker K.J., Bonnier P.G. (2006). The relevance of small-strain stiffness
in numerical simulation of excavation and tunnelling projects. In: NUMGE 2006 (ed.
Schweiger). Taylor & Francis, London. 133-139.
204
STRUCTURAL ELEMENTS & IN PLAXIS 2D
Presentation by
Dr William Cheang
Principal Geotechnical Consultant
Plaxis AsiaPac Pte Ltd
Contents
1. Structural elements available in Plaxis
2. Usage of structural elements in FE modelling
3. Plate elements (Beam and Shell element)
4. Anchor elements (Spring element)
5. Geotextile elements (Membrane element)
6. Interface elements (Zero thickness element)
205
1.Structural elements in Plaxis
4. Interface element
Section 3.44 & 14.1
206
3.1 Plate Element
Overview:
1. 3 or 5 noded line elements (for 6‐noded or 15‐noded element mesh)
2. 3 degrees of freedom per node
3. Plates have:
o Axial forces
o Shear forces
o Bending moments
o Hoop forces (axisymmetry)
4. Elastic or elastoplastic behaviour
5. For modelling walls, floors, tunnels
Plates – elastic parameters
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
6
207
3.3 Plate Element
Plates – elasto‐plastic behaviour
Np
M
Mp
1200 ‐100‐90 ‐80 ‐70 ‐60 ‐50 ‐40 ‐30 ‐20 ‐10 0 10 20 30 40 50 60 70 80 90 100
Envelope 0
Elasto‐plastic plate
1000
Elastic plate
‐5
800
Elasto‐plastic plate
‐10 Elastic plate
600
N
400 ‐15
200 ‐20
0
‐25 8
‐200 0 200 400 600 800 1000 1200 M
M
208
3.5 Plate Element
Effect on Global FOS by c/phi Reduction
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 – weight, in soil
dreal
209
3.7 Plate Element
Plates – weight, excavation
Actual problem In the model
dreal
1
wreal = concrete d real wmodel = soil d real wplate
2 Below GT
soil sat
1
wmodel = wreal wplate = ( concrete soil ) d real Above GT
2
soil 11 unsat
Rotation
spring
5 7
Hinged connection
Rigid connection
(default)
12
Illustration: Connection.P2D
210
3.9 Plate Element
Walls – thin wall vs. thick wall
• Thin wall
– Wall thickness << wall length
– No much end‐bearing, only friction
→ Plate element suffices
• Thick wall
– Wall thickness significant
– End‐bearing capacity needed
→ Use soil elements with material set represen ng wall material
→ In order to obtain structural forces a plate with fictitious properties
may be inserted
13
(Illustration: Beam.P2D): d
14
211
2. Anchor Element
Anchors – fixed‐end
a) To model supports, anchors and struts
a) Elasto‐plastic spring element
b) One end fixed to point in the geometry, other end is fully
fixed for displacement
c) Positioning at any angle
d) Pre‐stressing option
Anchors – node‐to‐node
a) To model anchors, columns, struts and rods
a) Elasto‐plastic spring element
b) Connects two geometry points in the geometry
c) No interaction with the mesh along the anchor rod
d) Pre‐stressing option
15
4.1Anchor Element
Anchors – material properties
Axial stiffness, EA (for one anchor) [kN]
Spacing, Ls (out‐of‐plane distance between anchors) [m]
Maximum anchor force for compression and tension,
|Fmax,comp| and |Fmax,tens| [kN]
Ls
16
212
4.2 Anchor Element
Anchors – pre‐stressing
• Defined in Staged construction phase
• Both tension (grout anchor) or compression (strut)
possible
Tension = positive
17
5.1.Geogrid Element
Geogrids
1. 3 or 5 noded line element
2. Elastic or elasto‐plastic behaviour
3. No flexural rigidity (EI), only axial stiffness (EA)
4. Only allows for tension, not for compression
18
213
5.2 Anchor Element + Geogrid Element
Ground anchors
1. Combination of node‐to‐node anchor and geogrid
2. Node‐to‐node anchor represents anchor rod (free length)
(no interaction with surrounding soil)
3. Geogrid represents grouted part (full interaction with surrounding soil)
4. No interface around grouted part; interface would create unrealistic failure surface
5. Working load conditions only – no pullout
6. If pullout force is known this can be used by limiting anchor rod force
19
5.3 Ground anchors
Axial force distribution along fixed length (modelled using geogrid)
Probable actual distribution of axial
forces in ground anchor
axial forces in geotextile element
214
5.4 Ground Anchors: Influence of node numbers
along structural elements
21
6.1 Interface Element
Interfaces – material properties
1. Soil‐structure interaction
1. Wall friction
2. Slip and gapping between soil and structure
2. Soil material properties
A. Taken from soil using reduction factor Rinter
3. Individual material set for interface possible
22
215
6.2 Interface Element
Interfaces – reduction factor
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 (grouted body) = Rinter≈ 1.0
(interface may not be required)
– Interaction soil/geotextile = Rinter≈ 0.9 – 0.5 (foil, textile)
23
References
24
216
25
217
E3: Excavation Exercise
Briefing of the Project
(for illustration only) 2
218
Proposed secant bored pile wall (SBP)
Dia. = 1 m with 200mm overlapping
Secant wall
d = 1 m
I = 0.049 m^4/m
E= 2.70E+07 kPa
A = 0.79 m^2/m
c/c spacing = 0.80 m
so,
EA/m = 2.65E+07 kN/m Take c/c spacing of 0.8m
EI/m = 8.28E+05 kNm^2/m Take c/c spacing of 1.6m 3
weight = 19 kN/m/m
The ground anchors are made of 32mm dia. Steel bars
at c/c spacing of 1m. The steel bar have a stiffness of
Es=2.1*10^8 kPa and with ultimate strength of 605kN
per anchor.
The anchors will be pre‐stressed to 60% of the
ultimatee strength, namely 363kN/anchor.
The properties of the grout body can be ignored.
219
Proposed ground anchors
Anchor free length (node‐to‐node element with Elastoplasticity ):
d = 0.032 m
E= 2.10E+08 kPa
A = 8.04E‐04 m^2/m
c/c spacing = 1.00 m
so,
EA/m = 1.7E+05 kN/m
Max axial force = 605 kN/m
Anchor grout body (geogrid element with Elastoplasticity):
d = 0.032 m
E= 2.10E+08 kPa
A = 8.04E‐04 m^2/m
c/c spacing = 1.00 m
so,
EA/m = 1.7E+05 kN/m
5
Max axial force = 605 kN/m
The upper 40m of the subsoil consists of a more or
less homogeneous layer of medium dense sand.
Typical soil parameters based on triaxial tests are
presented in the next slide.
Underneath this layer there is a very stiff layer of
gravel which can be acted as the bottom boundary
of the 2D FEM mesh.
The ground water table is very deep and does not
play a role in this analysis.
6
220
Soil parameters
Soil parameters
221
Simulation in Plaxis 2D version
2011
Simulation in Plaxis 2D
10
222
Tied-back excavation using HSsmall
TIED-BACK EXCAVATION
Using the HSsmall model
Computational Geotechnics 1
223
Tied-back excavation using HSsmall
2 Computational Geotechnics
224
Tied-back excavation using HSsmall
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.
AIMS
• Using interface elements
• Pre-stressing of anchors
0 x 4 1
Stage 1
Secant wall
11 12
Stage 2
13 14 Anchor rods
Stage 3
7 8 15
Grout bodies
16 17
9 5 18 10
6
3 2
Computational Geotechnics 3
225
Tied-back excavation using HSsmall
4 Computational Geotechnics
226
Tied-back excavation using HSsmall
MATERIAL PARAMETERS
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 compres-
sion 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.
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 =2.7•107 kN/m2 with a specific weight γ=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.
Computational Geotechnics 5
227
Tied-back excavation using HSsmall
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 =2.1*108 kN/m2 . The anchors have an ultimate strength
of 605 kN per anchor. In combination with a secant wall the anchors may be prestressed
to a maximum level of 60% of the ultimate strength, hence up to 363 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.
6 Computational Geotechnics
228
Tied-back excavation using HSsmall
Computational Geotechnics 7
229
Tied-back excavation using HSsmall
8 Computational Geotechnics
230
Tied-back excavation using HSsmall
GEOMETRY INPUT
• Start a new project
Project properties
• Accept the default values in the Project tab sheet of the Project properties (15-node
elements). For the dimensions see figure 3.
Geometry
(15,0)
(0,0) (70,0)
0 x 4 1
(0,-5) 11 12
(0,-10) 13 14
(30,-15)
(0,-15) 7 8 15
(37.5,-20)
(30,-20) 16 17
(0,-25) (70,-25)
9 5 18 10
6 (37.5,-25)
(15,-27)
(0,-60) (70,-60)
3 2
• Click the Geometry line button and draw the geometry contour and soil layers as
specified in figure 4.
Computational Geotechnics 9
231
Tied-back excavation using HSsmall
• Click the Plate button and draw the secant wall from (15, 0) to (15, -25).
• Click the Interface button and draw the interface from (15,0) to (15, -27) and back to
(15,0). This creates an interface on both sides of the secant wall.
• Click the Node-to-node anchor button and insert both anchor rods. These anchors
connect the beginning of the grout bodies to the wall.
• Finally, click the Geometry line button again to introduce the two levels of excavation.
Fixities
Material properties
• Enter the material properties for the four soil data sets, as determined in table 1of this
exercise.
• After entering all properties for the three soil types, drag and drop the properties to the
appropriate clusters.
• Enter material properties for the plates, anchors and ’geogrids’ as indicated in tables 2,
3and 4.
10 Computational Geotechnics
232
Tied-back excavation using HSsmall
Mesh generation
• From the Mesh menu, set the Global coarseness to Medium and press the Generate
button. This will result in a mesh as shown in figure 5.
• Select the geogrid and plate elements and press Refine line from the Mesh menu. This
will result in a refinement around the selected lines as shown in Figure 6.
Computational Geotechnics 11
233
Tied-back excavation using HSsmall
12 Computational Geotechnics
234
Tied-back excavation using HSsmall
CALCULATION
• When starting the calculation program choose the Classical mode.
The entire construction process consists of five phases. Define the phases, as shown graph-
ically below. For each phase, use the Plastic calculation, Staged construction.
Initial phase
For the initial phase choose the K0 procedure for calculating the initial stresses. As the phreatic
line is located below the geometry the generation of initial pore pressures can be skipped and
since it’s not necessary to switch off any soil for the initial situation it is not needed to define
the initial phase.
Phase 1
• In the first phase, the diaphragm wall is activated and the first excavation takes place.
• Note that though the the interfaces along the wall are activated automatically with the
activation of the wall, the extensions below the diaphragm wall have to be activated
manually.
Phase 2
In the second phase, a new option is used, namely the prestressing of anchors.
• First the grout-body (the geogrid) is switched on by clicking on the ’geogrid’ element.
The element will appear in yellow as soon as it is switched on. The light grey colour
indicates non-active elements.
Computational Geotechnics 13
235
Tied-back excavation using HSsmall
• Now that the grout-body is active, the anchor element needs to be prestressed. By
double clicking on a node-to-node anchor a window will appear as shown in figure 9.
• Select the option Adjust prestress, fill in a prestress force of 300 kN/m (tension) and
press OK.
• 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.
Phase 3, 4 and 5
Now define the remaining phases according to figures 10, 11 and 12.
14 Computational Geotechnics
236
Tied-back excavation using HSsmall
Figure 10: Phase 3: Second excavation Figure 11: Phase 4: Activation and
prestressing of 2nd anchor
Computational Geotechnics 15
237
Tied-back excavation using HSsmall
16 Computational Geotechnics
238
Tied-back excavation using HSsmall
INSPECT OUTPUT
The results of fase 5 is presented in Figure 13. After this final stage the excavation bottom
heave calculated is about 5 cm.
• By double clicking on the 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 geogrids the change of axial forces within the grout body
can be investigated. What is immediately noticeable is that the axial force at the connection
with the anchor rod is significantly lower than the force in the anchor rod itself. This is due the
fact that the end of the anchor rod is not only connected to the grout body, but also to several
soil elements surrounding the end of the anchor rod. Therefore part of the anchor force is
transferred directly to those soil elements while part of the anchor force is transferred to the
geotextile representing the grout body. The amount of force transferred to the soil depends on
the stiffness of the soil; in this exercise it is 25-35% of the anchor force. However, this effect
has very little influence on other calculation results. That is, it is not so important for other
calculation results how the anchor rod transfers its force; directly to the soil or by means 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 350 kNm/m (figure 14)
• 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 Ctrl + Shift pressed on the
keyboard while double-clicking on the interface. In figure 15the 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.
Computational Geotechnics 17
239
Tied-back excavation using HSsmall
Figure 14: Bending moments in the sec- Figure 15: Effective normal stresses in the
ant 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 boundar-
ies (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 γp that specifies the maximum
shear strain level reach in the stress history.
18 Computational Geotechnics
240
Tied-back excavation using HSsmall
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
16.
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 (70, -0.1) for the second point and press OK. This will create a
cross section from the secant wall until the right boundary of the model 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 12-13 mm, see figure 17.
Computational Geotechnics 19
241
Tied-back excavation using HSsmall
20 Computational Geotechnics
242
Tied-back excavation using HSsmall
Strength parameters
σ1 −σ3
σ1 +σ3
= sin ϕ
370−100
370+100
= sin ϕ
ϕ = 35o
ψ =ϕ − 30 = 5o
Computational Geotechnics 21
243
Tied-back excavation using HSsmall
Stiffness parameters
Since excavation is considered in this exercise, the input of Young’s modulus E should be
based on unloading, rather than on primary loading. For the same reason, Poisson’s ratio
should also be based on unloading, which results in a somewhat lower value.
The triaxial test has a cell pressure σ 3 = 100 kPa. This corresponds with reference pressure,
so E50 = Eref
50 .
22 Computational Geotechnics
244
Tied-back excavation using HSsmall
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 π·(0.5)4
Iwall = D
= πr
4D
= 4·0.8
= 61.3 · 10−3 m4 /m
Computational Geotechnics 23
245
Undrained Soil Behaviour
Contents
• Drained / undrained (conditions and analysis)
• Drained / undrained soil behaviour
– Typical results from drained and undrained triaxial tests
– Strength parameters
– What is the critical case: drained or undrained?
• Modeling undrained behaviour with Plaxis
– Three methods
– Effective stress analysis: how does it actually work
• Undrained shear strength
– Undrained behaviour with Mohr-Coulomb Model
– Undrained behaviour with Hardening Soil Model
– Influence of dilatancy
• Summary
246
Drained / undrained (conditions and analysis)
• In drained analysis water is assumed to immediately flow out
upon loading and therefore no pore pressures are built up.
This is appropriate when
– Permeability is high
– Rate of loading is low
– Short term behaviour is not of interest for problem considered
k = Permeability
k E oed Eoed = Oedometer modulus
T t w = Unit weight of water
γ w D2
D = Drainage length
t = Construction time
T = Dimensionless time factor
U = Degree of consolidation
247
Contents
• Drained / undrained (conditions and analysis)
• Drained / undrained soil behaviour
– Typical results from drained and undrained triaxial tests
– Strength parameters
– What is the critical case: drained or undrained?
• Modeling undrained behaviour with Plaxis
– Three methods
– Effective stress analysis: how does it actually work
• Undrained shear strength
– Undrained behaviour with Mohr-Coulomb Model
– Undrained behaviour with Hardening Soil Model
– Influence of dilatancy
• Summary
248
Triaxial test (OC) – drained / undrained
Typical results from drained (left) and undrained (right) triaxial tests
on overconsolidated soils
249
Strength parameters
Mohr-Coulomb parameters in terms of effective stress (real soil behaviour)
c tan
c
3 ' s’ 1 '
1 3 3 c c
1 sin ; t s sin
2 2 tan tan 9
Strength parameters
MC parameters in terms of total stresses (only undrained conditions!)
Cu
1 3 1 3
2 F
2 F
cu ,
-Cu
Effective stresses
250
What is the critical case: drained or undrained?
t
1.
short
short long
2.
ESP
long
s, s’
1.
ESP
s, s’
• For very soft NC soil, factor of safety against failure may be lower for
short term (undrained) conditions for unloading problems (e.g.
excavations)
• For very stiff OC soil, factor of safety against failure may be lower for
short term (undrained) conditions for loading problems (e.g.
embankment) 12
251
Contents
• Drained / undrained (conditions and analysis)
• Drained / undrained soil behaviour
– Typical results from drained and undrained triaxial tests
– Strength parameters
– What is the critical case: drained or undrained?
• Modeling undrained behaviour with Plaxis
– Three methods
– Effective stress analysis: how does it actually work
• Undrained shear strength
– Undrained behaviour with Mohr-Coulomb Model
– Undrained behaviour with Hardening Soil Model
– Influence of dilatancy
• Summary
13
252
Undrained behaviour with PLAXIS
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 3 ' B (1 2 ')
K total u
3 1 2 u 1 ' 3 B (1 2 ')
Notes:
• Skempton B-value can be entered explicitely for undrained materials in order to
simulate effect of partially saturated soil on the effective and excess pore
pressures.
• This procedure gives reasonable relation between νu and B only for ´ < 0.35 !
• Real value of Kw/n ~ 1.106 kPa (for n = 0.5) 15
uf
u
TSP
ESP
cu
s, s’
253
FE modeling of undrained behaviour (method C)
TSP=ESP
cu
s, s’
ESP TSP
cu
s, s’
254
Undrained behaviour with PLAXIS
Notes on different methods:
– Method A:
• Recommended, but be careful with MC model
• Soil behaviour is always governed by effective stresses
• Increase of shear strength during consolidation included
• Essential for exploiting features of advanced models such as the
Hardening Soil model, the Soft Soil model and the Soft Soil Creep model
– Method B:
• Only when no information on effective strength parameters is available
• May be a safer choice than Method A when using MC-model
• Cannot be used with the Soft Soil model and the Soft Soil Creep model
– Method C:
• NOT recommended
• No information on excess pore pressure distribution (total stress
analysis)
19
Contents
• Drained / undrained (conditions and analysis)
• Drained / undrained soil behaviour
– Typical results from drained and undrained triaxial tests
– Skempton's parameters A and B
– Strength parameters
– What is the critical case: drained or undrained?
• Modeling undrained behaviour with Plaxis
– Three methods
– Effective stress analysis: how does it actually work
• Undrained shear strength
– Undrained behaviour with Mohr-Coulomb Model
– Undrained behaviour with Hardening Soil Model
– Influence of dilatancy
• Summary
20
255
Undrained behaviour of Mohr-Coulomb
t
cu,MC tan(φ)
cu,real
Mohr-Coulomb
Real soil
s’
1
cu c ' cos ' s 'sin ' c ' cos ' vo ho sin '
2
1
cu c ' cos ' 'v 0 1 K 0 sin '
2 21
s’
256
Influence of constitutive model
Parameter sets for Hardening Soil model
ref ref ref ref nc
Model Number E50 Eur Eoed c ur p m K0 Rf
2 2 2 2 2
kN/m kN/m kN/m ° ° kN/m - kN/m - - -
HS_1 30 000 90 000 30 000 35 0 / 10 0.0 0.2 100 0.75 0.426 0.9
HS_2 50 000 150 000 50 000 35 0 0.0 0.2 100 0.75 0.426 0.9
HS_3 15 000 45 000 15 000 35 0 0.0 0.2 100 0.75 0.426 0.9
HS_4 30 000 90 000 40 000 35 0 0.0 0.2 100 0.75 0.426 0.9
HS_5 30 000 90 000 15 000 35 0 0.0 0.2 100 0.75 0.426 0.9
HS_6 50 000 150 000 30 000 35 0 0.0 0.2 100 0.75 0.426 0.9
HS_1
HS_2
125 HS_3
HS_4
HS_5
100 HS_6
total stress path
q [kN/m ]
2
75
50
25
0
0.00 25.00 50.00 75.00 100.00 125.00 150.00
24
p' [kN/m2]
257
Parameter variation – Hardening Soil
Simulation of undrained triaxial compression test – HS model - q vs 1
150
125
100
q [kN/m ]
2
75
50 HS_1
HS_2
HS_3
25 HS_4
HS_5
HS_6
0
0.00 1.00 2.00 3.00 4.00 5.00 6.00
25
1 [%]
70
excess pore pressure [kN/m ]
2
60
50
40
30
HS_1
HS_2
20
HS_3
HS_4
10 HS_5
HS_6
0
0.00 1.00 2.00 3.00 4.00 5.00 6.00
26
1 [%]
258
Influence of dilatancy
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
Influence of 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
28
1 [%]
259
Influence of dilatancy
Simulation of undrained triaxial compression test – MC / HS model - q vs p´
300
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
29
p' [kN/m2]
Influence of dilatancy
Simulation of undrained triaxial compression test – MC / HS model - pw vs 1
100
90 MC non dil
MC dil
80
HS_1 non dil
excess pore pressure [kN/m ]
2
70 HS_1 dil
60
50
40
30
20
10
-10
-20
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
30
1 [%]
260
Summary
• Undrained analysis should be performed in effective stresses and
with effective stiffness and strength parameters (Method A)
• Undrained shear strength is result of the constitutive model
• The Mohr-Coulomb model generally overestimates the undrained
shear strength in a Method A calculation. This can be „solved“ by
doing a Method B analysis, but this is a „trick“ that generally
generates incorrect excess pore pressures
• One should not use dilatancy in an undrained analysis
31
References
Atkinson, J.H., Bransby, P.L. (1978)
The Mechanics of Soils, An Introduction to Critical State Soil Mechanics. McGraw Hill
Ortigao, J.A.R. (1995)
Soil Mechanics in the Light of Critical State Theories – An Introduction. Balkema
Schweiger, H.F. (2002)
Some remarks on pore pressure parameters A and B in undrained analyses with the Hardening
Soil Model. Plaxis Bulletin No.12
Skempton, A.W. (1954)
The Pore-Pressure Coefficients A and B. Geotechnique, 4, 143-147
Vermeer, P.A., Meier, C.-P. (1998)
Proceedings Int. Conf. on Soil-Structure Interaction in Urban Civil Engineering, Darmstadt, 177-
191
32
261
Modelling of Groundwater in PLAXIS
CONTENTS
A. Introduction
1. Groundwater in Geotechnical Engineering
2. Plaxis
B. Definitions Porewater Pressures in Plaxis
1. Active
2. Steady-state
3. Excess
C. Generation of Porewater Pressures in Plaxis
1. Porewater Pressure due to Hydrostatic Condition
2. Pore Pressures due to Groundwater Flow (Steady or Transient States)
D. Hydraulic models
1. Fully Saturated Soils
2. Partially Saturated Soils
E. Case Histories
1. Excavations
2. Embankments and Dams
3. Slopes
F. References
262
Modelling of Groundwater in Plaxis
A. INTRODUCTION
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 boundary
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
263
Modelling of Groundwater in Plaxis
B. DEFINITIONS OF POREWATER
PRESSURES IN PLAXIS
2. Advanced
3. Flow
264
A1.ACTIVE POREWATER PRESSURE
2. Steady-state pre pressures are generated due to water conditions (hydraulic boundaries)
assigned to soil clusters (layers)
265
A3.EXCESS POREWATER PRESSURE
B. CALCULATION MODES
266
B1.CLASSICAL MODE
A. Steady-state pore pressures
1. Phreatic lines
2. Steady-state groundwater flow analysis
3. Transient-state groundwater flow analysis =
Steady-state background pore pressure
B. Excess pore pressures
1. Undrained material type in combination with
“Plastic” calculation
2. Consolidation analysis
INPUT KERNEL
B2.ADVANCED MODE
A. Consolidation analysis
B. Transient groundwater flow analysis
KERNEL
267
B3.FLOW MODE
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
268
GENERATION OF STEADY-STATE PORE PRESSURES
2. Cluster Dry
269
C1. PHREATIC LINE
270
GENERAL & CLUSTER PHREATIC LINES
Cluster: Dry
271
Cluster: User-defined Pore Pressure Distribution
272
COMBINATION: PHREATIC & CLUSTER OPTIONS
Cluster Dry
Interpolated
Case Histories
273
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
274
GROUNDWATER FLOW: STEADY STATE
1. Boundary conditions:
2. Soil permeabilities
275
Pore pressures – steady-state
Steady-state flow
3 28 29 6 9 30 31 2
4 8 11 5
General General
General
16 17
21 26
19 18
20 23 24 27
22 25
13 14 12
15 7 10
0 1
276
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
277
Groundwater flow – flow in unsaturated soil
krel
1
hp = -ε
hp = 0
m
0 hp
β
278
Groundwater flow – flow in unsaturated soil
Soil Water Characteristic Curve (SWRC or RC)
1 g n
S ( h p ) S res ( Ssat S res ) 1 g a h p
gn ( g )
n
2
g n 1
g n g n
krel S S e l 1 1 S e n
g g 1
with S Sres
Se
Ssat Sres
279
Groundwater flow – flow in unsaturated soil
1 if hp 0
Linear in Saturation
hp
S hp 1 if hps h p 0
hps
0 if hp 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
280
Groundwater flow - material data sets
Upper soils:
< 1m below soil surface
Lower soils:
all deeper soils
Hydraulic Properties of
European Soils
Particle distribution:
• < 2μm
• 2μm - 50μm
• 50μm – 2mm
281
Groundwater flow - material data sets
Particle distribution:
• < 2μm
• 2μm - 50μm
• 50μm – 2mm
12 soils data sets
No difference between
upper and lower soils
Relative permeability
Degree of saturation
282
Modelling of Groundwater in Plaxis
E. EXAMPLES
283
284
285
286
REFERENCES
A. Galavi, V. (2010), Groundwater flow, fully coupled flow deformation and undrained analyses in
Plaxis 2D and 3D. Technical Report, Plaxis B.V.
287
Excavation and dewatering
Computational Geotechnics 1
288
Excavation and dewatering
2 Computational Geotechnics
289
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 ground anchors. The Hardening Soil 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 ow calculation
to generate the new water pressure distribution.
INPUT
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.
The relevant part of the soil consists of three distinct layers. From the ground surface to a depth of 3 m there is
a ll of relatively loose ne sandy soil. Underneath the ll, 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 ll layer) Below the sand layer there is a loam layer which extends to large depth.
Geometry model
The symmetric problem can be modelled with a geometry model of 32 m width and 20 m depth. The proposed
geometry model is given in gure 2. A ground anchor can be modelled by a combination of a node-to-node anchor
and a geogrid (yellow line). The geogrid simulates the grout body whereas the node-to-node anchor simulates
the anchor rod. The diaphragm wall is modelled as a plate. The interfaces around the plate are used to model
soil-structure interaction eects. They are extended under the wall for 1.0 m to allow for sucient exibility and
accurate reaction forces. Interfaces should not be used around the geogrids that represent the grout body. In
general, it is a good habit to extend interfaces around corners of structures in order to allow for sucient freedom
of deformation and to obtain a more accurate stress distribution. When doing so, make sure that the extended
part of the interface is always turned o in the water conditions mode.
Computational Geotechnics 3
290
Excavation and dewatering
(10,0)
(0,0) (45,0)
0 1 2
(0,-3) 10 12 3
(45,-3)
(0,-7)
9 13 (19,-9)
(0,-10) 16
8 14 (22,-11)
17 18
19
(17,-14)
(0,-17) 7 11 4
(45,-17)
15
(10,-18)
(14,-11)
(0,-35) (45,-35)
6 5
Material properties
The soil consists of three distinct layers. The parameters of the dierent layers are shown in table 1. The interfaces
around the wall will be left impermeable in order to block the ow through it. Since the interfaces in the loam
layer below the wall (the extended part of the interfaces) do not inuence the soil behaviour, therefore their
strength is not reduced and the permeability must be changed to permeable. This will be achieved during the
denition of the staged construction phases.
Threshold shear strain γ0.7 1.0 · 10−4 1.0 · 10−4 1.5 · 10−4
Reference small-strain shear modulus Gref
0 180.0 · 103 350.0 · 103 180.0 · 103 kN/m2
Advanced parameters Default Default Default
Horizontal permeability kx 1.0 0.5 0.1 m/day
Vertical permeability ky 1.0 0.5 0.1 m/day
Interface strength reduction Rinter 0.65 0.7 Rigid
Coecient for initial horizontal stress K0 Automatic Automatic Automatic
4 Computational Geotechnics
291
Excavation and dewatering
The properties of the concrete diaphragm wall are entered in a material set of the plate type. The concrete has a
Young's modulus of 35 GPa and the wall is 0.35 m thick. The properties are listed in table 2.
For the properties of the ground anchors, two material data sets are needed: One of the Anchor type (anchor rod)
and one of the Geogrid type (grout body). The Anchor data set contains the properties of the anchor rod and
the Geogrid data set contains the properties of the grout body. The data are listed in tables 3 and 4.
Mesh generation
For the generation of the mesh it is advisable to set the Global coarseness parameter to Medium. In addition, it
is expected that stress concentrations will occur around the two grout bodies and in the lower part of the wall,
hence local renements are proposed there.
After generating the mesh, continue to the calculation.
Computational Geotechnics 5
292
Excavation and dewatering
CALCULATION
The calculation consists of the initial phase and six phases.
• In the rst phase the wall is constructed.
• In the second phase the rst 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 rst 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 ow analysis to calculate the new pore water pressure distribution, which is a
part of the denition of the third calculation phase.
• In the fth phase the second anchor is installed and prestressed.
• The sixth phase is a further excavation (and de-watering) to the nal depth of 10 m.
The calculation will be done using 2 alternative methods. In the rst method the water will be lowered using
steady-state groundwater ow analysis. This method assumes that excavation is suciently slow that the ow
eld will reach a steady-state situation for every excavations step. For rather slow excavations in high permeable
soils this is a reasonable assumption. In the second method the water will be lowered using a transient ow
analysis. This method is the preferred method if the excavation is suciently fast that no steady-state situation
will be reach during excavation.
Initial phase
• Set the Calculation type to K 0 procedure for calculating the initial stresses.
• Press the Dene button on the Parameters tabsheet to dene the initial situation
• In Staged construction mode make sure that all soil is activated and all structural elements are deactivated,
then continue to Water conditions mode.
• Draw a horizontal phreatic level from (x,y) = (-2,-3) to (20,-3), (30,-3) and (47,-3).
• Pore pressures will be generated based on this phreatic line. To do so, make sure the Generate by phreatic
level button is selected.
6 Computational Geotechnics
293
Excavation and dewatering
Phase 1: Construction of the diaphragm wall
• Construction of the diaphragm wall takes 5 days. Therefore, ll in a Time interval of 5 days on the
Parameters tabsheet.
• Go to the phase denition by pressing the Dene button on the Parameters tabsheet.
• In Staged construction mode activate the wall and the interface extensions below the wall. The interfaces
along the wall are activated automatically.
• Also on the Parameters tabsheet, enter a construction time of 7 days in the Time interval eld.
• Go to the phase denition by pressing the Dene button on the Parameters tabsheet.
• In Stage construction mode of the phase denition activate the upper geotextile representing the grout body
of the rst anchor.
• Double-click on the top node-to-node anchor, the properties window for the node-to-node anchor appears.
• Select the option Adjust prestress and enter a 120 kN/m prestress force.
• No water ow can occur through a axis of symmetry. Therefore the axis of symmetry must be a closed ow
boundary. To do so, select the Closed boundary button and draw a closed boundary from (x,y) = (0,0)
to (0,-35). Check that the bottom of the geometry is also a closed boundary.
• During excavation the water level will be lowered. Due to high permeabilities water will be drawn from
outside the excavation, hence a groundwater ow analysis has to be performed. Therefore, make sure the
Groundwater button is set to Groundwater ow steady state by clicking the down arrow and choosing
the correct option.
• The groundwater head boundary conditions needed for the groundwater ow analysis can be applied in a
simple manner by using the general phreatic level. In order to do so, make sure no cluster is selected (for
instance by clicking completely outside the geometry so that the general phreatic line is red) and then draw
a new general phreatic level from (x,y) = (-2,-7) to (20,-7), (30,-3) and (47,-3).
Computational Geotechnics 7
294
Excavation and dewatering
Phase 5: Prestress second anchor
• Prestressing the second row of anchors will take 1 day, hence enter a 1 day construction time.
• In Stage construction mode of the phase denition activate the lower geotextile representing the grout body
of the second anchor.
• Double-click on the lower node-to-node anchor, the properties window for the node-to-node anchor appears.
• Select the option Adjust prestress and enter a 200 kN/m prestress force.
• Close the properties window and continue to Water conditions mode.
• The phreatic line should be still the same as in the previous calculation phase and also the option Ground-
water ow steady-state should still be selected.
• Return to the Calculations program.
• 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) = (-2,-10) to (20,-10), (30,-3) and (47,-3).
8 Computational Geotechnics
295
Excavation and dewatering
Method 2: Transient groundwater ow
In this method a fully coupled analysis will be performed. This analysis couples transient groundwater ow,
consolidation and deformations implying that the groundwater ow eld, development and dissipation of excess
pore pressures and deformation are calculated simultaneously all inuencing each other. This type of analysis
should be performed if developement of excess pore pressures is expected inuencing the ow eld or when
signicant 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 ow eld will not reach a steady-state during
excavations. The addidional eects of coupling the ow eld 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 Consolidation.
It is possible to re-use the project made for the calculation using the method of steady-state analysis:
• From the menu Tools select the option Calculation mode. In the window that now appears select Advanced
mode.
• Change the calculation phases according to the description below. Note that only the changes relative to
the steady-state calculation method are mentioned.
Initial phase
• No changes have to be made
• Dene the staged construction phase and switch to Water conditions mode.
• During excavation the water level will be lowered. However, due to the short construction time it's unlikely
that the ow eld will be steady state and therefore a transient groundwater ow analysis will be done.
Therefore, make sure the Groundwater button is set to Groundwater ow transient by clicking the
down arrow and choosing the correct option. The phreatic level remains unchanged.
Computational Geotechnics 9
296
Excavation and dewatering
Phase 5: Prestress second anchor
• Set the calculation type to Consolidation on the Parameters tabsheet.
• Dene the staged construction phase and switch to Water conditions mode.
Though the phreatic level in the excavation doesn't change, the ow eld is not steady-state yet outside the
excavation. Therefore this phase needs transient ow analysis without making further changes.
• Also on the Parameters tabsheet, set the number of Additional steps to 500.
• Dene the staged construction phase and switch to Water conditions mode.
10 Computational Geotechnics
297
Excavation and dewatering
OUTPUT
Figure 3 gives the total displacements for the nal phase for both the calculation with steady-state groundwater
ow and the transient groundwater ow.
The excavation using steady-state ow gives a maximum displacements of about 24 mm while excavation using
transient ow gives a maximum displacement of about 23 mm.
Figure 3: Total displacements for the steady state ow analysis (left) and the transient ow analyis (right)
Figure 4 shows the vertical displacements for the nal phase for both calculations. For the displacements behind
the wall the excavation using steady-state analysis clearly gives more vertical displacements over a larger distance
from the excavation than the excavation with transient ow.
Figure 4: Vertical displacements for the steady state ow analysis (left) and the transient ow analyis (right)
The extreme bending moments are about -165 kNm/m and 75 kNm/m for the excavation using steady-state
groundwater ow analysis while the extremen bending moments for the excavation using transient groundwater
ow are about -170 kNm/m and 95 kNm/m.
Figure 6 shows the horizontal displacements of the top of the wall as a function of construction time for both the
excavation using steady-state ow and transient ow.
Computational Geotechnics 11
298
Excavation and dewatering
Figure 5: Bending moments in the wall for the steady state ow analysis (left) and the transient ow analyis
(right)
12 Computational Geotechnics
299
GENERATION OF INITIAL STRESSES & PHI-C REDUCTION
ANALYSIS
William Cheang Plaxis AsiaPac
Research and Numerical Experiments
Cheang, Lee & Tan
CONTENTS
a. Ko-Procedure
b. Gravity Switch On
b. Some Examples
c. Pointers
d. Appendix
e. References
300
Ko and Gravity Switch-On Procedure
PART 1: INITIAL STRESSES
INITIAL STRESSES
1. Initial stresses represent the equilibrium state of the undisturbed soil and consist of:
a) Soil weight
b) Loading history
a) K0 procedure
b) Gravity loading
301
K0-PROCEDURE
1. Generation of initial stresses during Initial Phase Material Set
' h 'v K 0
GRAVITY LOADING
1. Calculation of initial stresses by weight
loading.
2. Disadvantage: Non-physical displacements
are created.
3. Advantage: Equilibrium satisfied in all cases.
For 1D compression:
'n 'v Non-physical displacements
1 reset in subsequent phase)
so
K0
1
302
GRAVITY LOADING
1. Procedure
a. Initial phase
a. Skip K0 procedure, ΣMweight remains zero
b. Phase 1
a. Choose Plastic calculation, Total multipliers
b. Set weight multiplier ΣMweight = 1
c. Phase 2
a. Select Reset displacements to zero to discard all displacements from raising the gravity
GRAVITY LOADING
Points
1. Undrained material
303
Since Plaxis 2D version 2010
GRAVITY LOADING
Cases where gravity loading should be used instead of K0-procedure:
304
SPECIAL CASES
Initial stresses
SPECIAL CASES
1. 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 actual project.
existing buildings
our
project
our project
305
Phi-C Reduction in Plaxis
Main advantages:
2. Critical surface is found automatically as slope failure occurs naturally through the zones due to
insufficient shear strength to resist shear stresses.
5. Information such as stresses, movements and pore-pressures and numerical tool as for deformation
analysis
306
Safety factor
• Many possible definitions
a v a ila b le s o il r e s is ta n c e
1 .8
m o b iliz e d s o il r e s is ta n c e
fa ilu r e lo a d
5 .9
w o r k in g lo a d
Lowered incrementally
tan reduced
c tan
Msf
creduced tan reduced
307
SAFETY FACTOR ANALYSIS
Calculation procedure:
1. Create a phi/c reduction phase (in version 2010 select ‘Safety’)
2. Accept the default increment for MSF=0.1 from the multiplier
tab-sheet.
3. Calculate
4. Carefully examine ΣMsf vs. displacement
curve in Plaxis Curves
Notes:
a. Select control point within (expected) failing body
b. Use sufficient number of load steps (250-500?)
c. Use a sufficiently fine mesh (Check for mesh sensitivity)
d. Limit the maximum structural forces by choosing elastoplastic behaviour
for walls, anchors and geotextiles.
1.12
Sum-Msf
1.12
Sum-Msf
1.08
1.08
1.04
1.04
1.0
0.0 0.3 0.6 0.9 1.2 1.5 1.0
0.0 0.3 0.6 0.9 1.2
displacement
displacement
308
SAFETY FACTOR ANALYSIS
2
1
H = 12m
D cu = 50 kPa
cu = 100 kPa Plaxis:
F = 1,35
cu 50 D (Taylor,1948)
Stability charts: F N0 6.6 1.38 , N 0 f ( , )
Pd 12 20 H
309
EXAMPLE 2: HOMOGENOUS SLOPE WITH AND WITHOUT FOUNDATION LAYER
Note:
1 Griffiths & Lane (1999)
Model Slope G2 (757 elements-15n) : 2 Bishop & Morgenstern (1960)
FOS= 1.323 (1.4001, 1.7522,1.2794,1.3795, 1,3756) 3.Taylor (1937)
4.Janbu
5.Bishop
6.Morgenstern-Price
Cu2
Cu2/Cu1 = 2
cu = 50 the
kPasame strength with upper soil
Cu1
Plaxis FOS = 1.19
Cu2/Cu1 = 1.0
Cu2
cu = 60 kPa
310
EXAMPLE 4: UNDRAINED CLAY SLOPE WITH A THIN LAYER
2.0
1.8
1.6
1.4
Factor of safety)
1.2
Model Slope G3A cu2/cu1=1 (3436 elements- Model Slope G3D cu2/cu1=0.8 (3436 elements-
1.0
15n) 15n)
0.8
0.4
0.2
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1
Cu2 /Cu1)
Note:
Model Slope G3B cu2/cu1=0.6 (3436 elements- Model Slope G3F cu2/cu1=0.5 (3436 elements- 1 Griffiths & Lane (1999)
15n) 15n) 2 Bishop & Morgenstern (1960)
FOS= 1.319(1.401, 1.404) FOS= 1.112(1.251) 3.Taylor (1937) (green line)
4.Janbu
Model Slope G3E cu2/cu1=0.4 (3436 elements- Model Slope G3c cu2/cu1=0.2 (3436 elements-
15n) 15n) :
FOS= 0.903(1.051) FOS= 0.470 (0.591, 1.304)
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
24
311
SOME POINTS
1. Always inspect the incremental displacements or strains as computed in the last
load step to make sure that failure is reached.
3. Mesh: Refine and redo the phi-c analysis until the factor of safety remains constant
upon further refinement of the mesh.
4. Always use the arc-length time stepping procedure within the Phi-C reduction
(default)
312
The Case of Griffith and Marquez 2007 (Case 1)
27
Case 1(Griffith and Marquez, 2007)- A simple 2:1 slope stability verification
Incremental disp
313
Slip Surface : Spoon under iso-surfaces
L/H=8 L/H=12
314
Case 2: Inclined Side Face (to be continued)
31
REFERENCES
1. Matsui, T. & San, K.C. (1992) Finite element slope stability analysis by shear strength reduction technique. Soils
and Foundation, Vol.32 (1),pp.59-70
2. Zienkiewicz, O.C., Humpheson, C. & Lewis, R.(1975) Associated and non-associated visco-plasticity and plasticity
in soil mechanics. Geotechnique 25(4).pp. 671-689.
3. Ugai, K.(1989). A method of calculation of total safety factor of slope by elastoplastic FEM. Soils and Foundation
29(2). pp.190-195.
4. Farias, M.M., Naylor, D.J.(1998). Safety analysis using finite elements. Computer and Geotechnics, Vol 22(2) p.p.
165-181.
5. Griffiths, D.V., Lane, P.A. (1999). Slope stability analysis by finite elements. Geotechnique 49 (3), pp.387-403.
6. Griffith, D.V. & Marquez, R,M (2007). Three-dimensional slope stability analysis by elastoplastic finite elements,
Geotechnique 57, No. 6, 537–546.
315
Slope stability for a road construction project
Computational Geotechnics 1
316
Slope stability for a road construction project
2 Computational Geotechnics
317
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 gure 1.
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 signicant inuence 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 ll material to
support the road.
During the rst 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 ll and
hillside below the road using so-called launched soil nails: long steel reinforcement bars that are shot with high
speed into the ground.
• Construct the new road under dry (summer) conditions and calculate its factor of safety
• Apply stabilising soil nails and calculate the factor of safety in wet conditions
INPUT
Start a new project and select appropriate General settings according to the size of the geometry (see gure 2)
and make sure to use a snap distance of 0.25m. Use 15-node elements as basic element type since in this exercise
we will deal with failure behaviour.
Computational Geotechnics 3
318
Slope stability for a road construction project
Figure 2: Geometry model (a) and position of the road surface and soil nails (b)
Geometry
• Enter the geometry as indicated in gure 2a. The order in which geometry points are created is arbitrary.
• Introduce the 3 soil nails by using geogrids according to the coordinates given in gure 2b.
• Introduce the road surface by using a plate element from (22,16) to (28,16)
• Introduce the trac load by applying a vertical distributed load of -10 kN/m2 on the road surface.
Material properties
Soil and interfaces
• Enter the material properties for the three soil data sets specied in table 1.
• After entering all properties for the three soil types, drag and drop the properties to the appropriate clusters,
as indicated in gure 3.
4 Computational Geotechnics
319
Slope stability for a road construction project
Weath
ered s
iltston
e
A A A
Intact siltstone
Road surface
The road surface is modelled with a plate element. Therefore, create a new plate material set using the parameters
as specied in table 2 and assign it to the plate representing the road surface.
Soil nails
The 3 soil nails are modelled using geogrid elements. Hence, create a new geogrid material set with parameters
as specied in table 3 and assign the material to all 3 soil nails.
Computational Geotechnics 5
320
Slope stability for a road construction project
Mesh generation
• Set the Global coarseness to Medium.
• Select all clusters that fall within the boxed area (12 clusters in total) while keeping the <Shift> button
pressed and then select the option Rene cluster from the Mesh menu in order to rene the mesh in the
selected area. This will give a mesh as given in gure 4.
6 Computational Geotechnics
321
Slope stability for a road construction project
CALCULATION
The calculation consists of the initial phase and 12 calculation phases more in order to model the proper con-
struction sequence and the determination of the factors of safety at key moments in the construction process.
When starting the Calculations program select Classical mode for calculating undrained behaviour and consoli-
dation.
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 dene the initial situation, follow these steps:
• The geometry has a non-horizontal soil layering, hence the K0 -procedure cannot be used. Therefore, set the
Calculation type to Gravity loading.
• Dene the Staged construction settings and make sure only the clusters representing the original hillside are
activated. Also make sure all structural elements (road surface and soil nails) are switched o.
• In Water conditions mode, eEnter a phreatic level by two coordinates (-1, 10) and (56, 10).
• One the General tabsheet make sure this calculation phase is Safety.
• In order to discard the displacements during gravity loading make sure the option Reset displacements to
zero is selected on the Parameters tabsheet.
• On the Parameters tabsheet press the Dene button to dene the phase
Computational Geotechnics 7
322
Slope stability for a road construction project
• Open the material set database and assign the reinforced ll material set to the 4 clusters of the ll area,
see gure 6.
• Close the material sets database and press the Update to return to the Calculations program.
• Switch on the plate representing the road by clicking on it. Make sure the distributed load representing the
trac load remains switched o.
8 Computational Geotechnics
323
Slope stability for a road construction project
• Switch on both parts (left ánd right) of the distributed load representing the trac load. The plate repre-
senting the road surface remains switched on.
• Select the Phreatic level button and draw a new phreatic line from (-1,20) to (5,20) and further to (20,10)
and (56,10).
• If there is no closed ow boundary yet on the bottom of the geometry (indicated with a thick black line)
then select the Closed boundary button and draw a closed ow boundary at the full bottom of the geometry.
• Select Steady state as groundwater analysis type and press the Update button to return to the Calculations
program.
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.
Computational Geotechnics 9
324
Slope stability for a road construction project
Figure 7: Phase 9, Road construction with trac 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 12 - Factor of safety in winter conditions with all soil nails in-
stalled
• In order to determine the factor of safety directly in winter conditions with the all soil nails installed create
a Safety phase. Keep all default settings
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 (15,20), (25,16), (28,16) and (33,11).
10 Computational Geotechnics
325
Slope stability for a road construction project
INSPECT OUTPUT
Factors of safety
Check the factors of safety in the Curves program. Create a new curve of displacements vs. Sum-Msf for the
point at coordinates (25,16). See gure 8.
1.70
1.60
FoS ≈ 1.6 (all nails installed)
1.50
Σ
1.30
1.10
FoS ≈ 1.15 (winter conditions, no nails)
1.00
0.00 0.50 1.00 1.50 2.00
|U| [m]
For all situations check the failure mechanism. This can be done by for instance with the graph of incremental
shear strains (∆γs ) of every Safety phase. This will show the change in shear strains in the last calculation step,
hence when failure occurred, and will show any shear bands that may have occurred. See gures 9-13. It becomes
clear that installing the top nails disturbs the failure mechanism. However, it is only after having installed the
lower nails as well that the sliding of the road ll no longer is the most critical mechanism. Both failure of the
slope above the road and a very large hillside sliding mechanism with considerably higher factor of safety (almost
1.6) are now the critical mechanisms.
Computational Geotechnics 11
326
Slope stability for a road construction project
Figure 10: Most critical failure mechanisme after construction, summer conditions.
Figure 11: Most critical failure mechanisme after construction, winter conditions.
12 Computational Geotechnics
327
Slope stability for a road construction project
Figure 12: Most critical failure mechanisme after installing top soil nails
Figure 13: Most critical failure mechanisme after installing all soil nails
Computational Geotechnics 13
328
Overview of Soil Models, Parameters,
Possibilities & Limitations
Content
• Overview of soil models, parameters, possibilities & limitations
in PLAXIS 3D
– Mohr-Coulomb model
– Hardening Soil model (HS + HSsmall)
– Soft Soil model
– Soft Soil Creep model
– Hoek-Brown model
• Standard soil tests with different models
• Which model in which situation?
329
Overview of soil models in PLAXIS 3D
Mohr-Coulomb model
Linear-elastic perfectly-plastic behaviour
• Hooke’s law:
d xx 1 0 0 0 d xx
d 1 0 0 0 d yy
yy
d zz E 1 0 0 0 d zz
d xy (1 )(1 2 ) 0 0 0 1
2 0 0 d xy
d yz 0 0 0 0 1
0 d yz
2
d zx 0 0 0 0 0 1
2
d zx
330
Mohr-Coulomb model
Linear-elastic perfectly-plastic behaviour
• Yield function:
f 12 '3 '1 12 '3 '1 sin ' c 'cos '
Mohr-Coulomb model
Parameters:
E Young’s modulus
ν Poisson’s ratio
c Cohesion
Friction angle
ψ Dilatancy angle
331
Mohr-Coulomb model
Possibilities:
• First order approach of soil behaviour in general
• (Drained) failure behaviour quite well described
Limitations:
• Linear elastic behaviour until failure (no strain- or stress- or stress path-
dependent stiffness behaviour)
• Be careful with efffective strength in undrained behaviour
332
Hardening Soil model
Shear hardening:
2p,fric
1p,fric
p’
q
Hardening Soil model MC failure line
Compaction hardening:
Elastoplastic formulation of
p - v relationship in Cap
primary compression pc
fc = 0
pc p’
c 1
v
333
Hardening Soil model
Small-strain stiffness:
• Strain-dependent stiffness
• Hysteresis
• Energy dissipation
• Damping
334
Hardening Soil model
Possibilities:
• Better non-linear formulation of soil behaviour in general
• Distinction between primary loading and unload./reloading
• Memory of pre-consolidation stress
• Different stiffness in different stress paths
• Well suited for unloading situations with simultaneous deviatoric loading
(excavations)
• Strain-dependent stiffness (hysteresis, damping) (HSsmall only)
Limitations:
• No peak strength and softening
• No secondary compression (Creep); No anisotropy
• E50 / Eoed > 2 difficult to input
335
Soft Soil model
1D compression (determines M)
336
Soft Soil model
Possibilities:
• Reasonable model for primary loading of normally
consolidated clay and soft soils
• Failure behaviour better than (modified) Cam-Clay
Limitations:
• Less suitable for over-consolidated clay and in certain unloading stress
paths; not suitable for sand
• No time-dependent behaviour (secondary compression)
337
Soft Soil Creep model
338
Soft Soil Creep model
Parameters:
* Modified compression index
* Modified swelling index
* Modified creep index
ur Poisson’s ratio for unloading / reloading
c’ Effective cohesion
’ Effective friction angle
Dilatancy angle
K0 Horiz./vertical stress ratio in normally consolidated
nc
1D compression (determines M)
Limitations:
• ‘Side role’ of over-consolidation ratio OCR
• Influence of K0nc-parameter (M)
• No softening
339
Hoek-Brown model -
Hoek-Brown failure criterion (2002):
a
'3
'1 '3 ci mb s -
ci -
GSI 100
mb mi exp
28 14 D ci = Intact uniaxial compressive
1 1 GSI 20 strength
a exp exp
2 6 15 3 GSI = Geological Strength Index
GSI 100 mi = Intact rock parameter
s exp D = Disturbance Factor
9 3D
Hoek-Brown model -
-’1
Uni-axial compressive strength:
c ci s a
- -
Tensile strength:
s ci
t
mb
-c
-’3
t
340
Hoek-Brown model -
Parameters:
Hoek-Brown model -
Possibilities:
• Continuum approach of rock strength
- -
Limitations:
• Only applicable to rock
• No anisotropy
341
Standard soil tests with different models
Model 1 (Mohr-Coulomb) Model 2 (Hardening-Soil) Model 3 (Soft Soil)
E = 20000 kN/m2 E50 = 20000 kN/m2 * = 0.012
= 0.3 Eoed = 20000 kN/m2 * = 0.0024
c = 0 kN/m2 Eur = 60000 kN/m2 = 0.2
= 30° m = 1.0 c = 0 kN/m2
= 0° pref = 100 kN/m2 = 30°
= 0.2 = 0°
c=0 kN/m2 K0nc = 0.5
= 30°
= 0°
K0nc = 0.5
342
Standard soil tests with different models
Undrained triaxial test at 3 = 100 kPa
343
Standard soil tests with different models
Oedometer test with unloading
344
Standard soil tests with different models
Direct Simple Shear test at xx = yy = 100 kPa
345
346
Consolidation analysis using finite elements
Ronald Brinkgreve (with input from Vahid Galavi)
Plaxis BV, The Netherlands
Content
• Theory of consolidation
• FEM for consolidation analysis
• Validation: One-dimensional consolidation
• New features in PLAXIS 2D 2010
• Conclusions
347
Theory of consolidation
Considering:
Theory of consolidation
Pore water flow:
k k
Darcy’s law: q pw v 2 pw t
w w
Total change of volumetric strain in time, k = permeability
considering homogeneous permeability: w = unit weight of water
v ( v1 v 2 ) n pw k 2
pw
t t K w t w
348
Theory of consolidation
General 3D case:
v 1 p ' 1 ( p pw ) 1 p 1 pw
t K ' t K ' t K ' t K ' t
1 p 1 n pw k
2 pw
K ' t K ' K w t w
E'
where K' = bulk stiffness of soil skeleton and p = mean total stress
3 1 2 '
kK' pw p
Considering incompressible water: 2 pw
w t t
Theory of consolidation
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 = constrained modulus of soil skeleton
(1 ')(1 2 ')
k Eoed pw
Considering incompressible water: 2 pw
w t t
349
Theory of consolidation
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
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)
350
FEM for consolidation analysis
Consolidation analysis based on Biot’s theory of consolidation:
Coupling between deformation and flow of pore water (excess pore pressures)
351
FEM for consolidation analysis
Mandell-Cryer effect:
A C B D
C
axisymmetry
A
B E
D F E
352
FEM for consolidation analysis
Soil properties:
• Unit weight (sat,unsat)
• Stiffness (E,)
• Strength (,c)
• Material type (drained, undrained, non-porous), Permeability (k)
353
FEM for consolidation analysis
Soil properties:
• Unit weight (sat,unsat)
• Stiffness (E,)
• Strength (,c)
• Material type (drained, undrained, non-porous), Permeability (k)
354
FEM for consolidation analysis
Soil properties:
• Unit weight (sat,unsat)
• Stiffness (E,)
• Strength (,c)
• Material type (drained, undrained, non-porous), Permeability (k)
• Change of permeability (Ck)
• Creep
no creep
creep
time (log-scale)
355
FEM for consolidation analysis
Boundary conditions:
• Fixities
• Loads
• Closed consolidation boundaries (no outflow; otherwise ‘draining’ boundaries: pw=0)
Mesh: Stress
• Elements: Interpolation of primary variables point
• Nodes: Primary variables (displacements, pore pressures)
• Stress points: Derived variables (strains, stresses, Darcy velocities)
356
FEM for consolidation analysis
Calculations:
• Consolidation – Staged construction > Time interval t
• Consolidation – Minimum pore pressure > |p-stop|
• Consolidation – Incremental multipliers > Time increment
l2
tcritical
Cv
Note: smaller steps may give
stress oscillations
l = element length
= 80 for 15-node triangles
= 40 for 6-node triangles
357
FEM for consolidation analysis
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
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
* *
358
FEM for consolidation analysis
Output:
• Deformations
• Stresses
• Excess pore pressure
• History curves
(e.g. pore pressure as function of time)
359
Validation: One-dimensional consolidation
360
New features of PLAXIS 2D 2010
Unsaturated soil modelling:
Bishop stress
Suction (a new variable)
Retention curves (Mualem-Van Genuchten + user defined models)
Existing Plaxis soil models (Bishop stress)
User defined soil models (Bishop stress and suction)
• Flow mode
Steady state groundwater flow
361
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 pore pressure
• Adding creep gives more realistic time-dependent behavour and leads to
‘delayed’ consolidation
• Recent development: Fully coupled flow-deformation analysis and unsaturated
soil behaviour
362
Geotextile reinforced embankment
GEOTEXTILE REINFORCED
EMBANKMENT WITH CONSOLIDATION
Computational Geotechnics 1
363
Geotextile reinforced embankment
2 Computational Geotecnics
364
Geotextile reinforced embankment
INTRODUCTION
In 1979 a test embankment was constructed in the Netherlands near the town of Almere. The
objective of this test was to measure the influence of geotextile reinforcement on the short
term stability of an embankment on soft soil. Two test embankments were constructed on top
of a layer, one with and one without geotextile. The construction procedure was such that a
ditch was excavated in the clay layer while at the same time a retaining bank was made with
the excavated clay. A cross-section of the reinforced test embankment is given in figure 1.
line of symmetry
geotextile
retaining bank
1
sand fill 2
2
soft clay
1,5
strong sand layer
1 3.5 3.5 1 3 14 7
Cone penetration tests gave an average cone resistance of qc = 150 kPa for the clay. The clay
is considered to be normally consolidated. The behaviour is assumed to be undrained (the
retaining bank should be drained, however). The saturated weight of the clay is 13.5 kN/m3.
A plasticity index of Ip = 50% is assumed. Due to the limited soil data, parameters should
be selected using engineering judgement and by using the correlations given in the lecture
"Evaluation of soil stiffness parameters". To obtain an undrained shear strength for the clay
layer it is suggested to use the correlation su ≈ qc /15. Having no data for the effective cohesion
and the effective friction angle, they have to be estimated from the undrained shear strength in
order to do a consolidation analysis. For the determination of a stiffness parameter for the clay
layer it is suggested to use the correlation Eu ≈ 15000 · su /Ip (%). The shear modulus G is one
third of the undrained Young’s modulus Eu . The effective Poisson’s ratio should be chosen
such that a realistic K0nc is obtained in one-dimensional compression (K0nc = ν 0 /(1 − ν 0 ) ≈ 0.5).
The effective Young’s modulus is calculated from the shear modulus E 0 = 2G(1 + ν 0 ). The
fill was reported to be fully saturated loose sand with a saturated weight of 18 kN/m3 . The
behaviour is considered to be drained. The effective strength properties are estimated at ϕ0 =
30° and c’ = 3 kPa. K0nc is assumed at 0.5. For the stiffness one should take E’ = 4000 kPa
and ν 0 =0.33.
Computational Geotechnics 3
365
Geotextile reinforced embankment
AIMS
• Calculation of two alternatives within one project.
SCHEME OF OPERATIONS
1. Determination of stiffness & strength properties (clay)
2. Geometry input
3. Calculation
(a) Initial conditions (Pore pressure generation, Initial geometry configuration, Genera-
tion of initial stresses)
(b) Switch on geotextile, excavate ditch + raise retaining embankment
(c) Apply first hydraulic fill
(d) Apply second hydraulic fill
(e) Determine factor of safety
(f) Repeat this using consolidation phases instead of plastic phases.
4. Inspect output
4 Computational Geotecnics
366
Geotextile reinforced embankment
Note: The main purpose of the exercise is to assess the failure mechanism and
the factor of safety, which has the following consequences for the model:
• The geometry size is chosen such that the failure mechanism fits
within the model boundaries. This means the geometry can be fairly
small.
Computational Geotechnics 5
367
Geotextile reinforced embankment
GEOMETRY INPUT
General settings
Start a new project and select appropriate General settings. Use 15-node elements as basic
element type since in this exercise we will deal with failure behaviour.
(9.5,7.5) (12,8.5)
10 11
(33,8.5)
(8,7.5)
7 9 8 (33,7.5)
(4.5,5.5) (12,5.5) (26,5.5)
(0,5.5) 1 6 12 13 2
(33,5.5)
(0,3.5) 4 5 (1,3.5)
(0,2) 14 15
(33,2)
y
(0,0) 0 x 3
(33,0)
• Enter the geometry as indicated in the previous graph. The order in which geometry
points are created is arbitrary.
• Click the Geogrid button to introduce the geotextile (from (4.5, 5.5) to (26.0, 5.5)).
• Click the Standard fixities button for the standard boundary conditions.
• After entering all properties for the three soil types, drag and drop the properties to the
appropriate clusters, as indicated in figure 3.
6 Computational Geotecnics
368
Geotextile reinforced embankment
3
2 3
1
1
Geotextile
• In the project database select the data type Geogrids and create a new material set. In
this material set, enter 2500 kN/m as stiffness. Note that this is the stiffness in extension.
In compression no stiffness is used.
• Drag the geogrid data set to the geotextile in the geometry and drop it there. The
geotextile should flash red once, indicating the properties have been set.
Mesh generation
• From the Mesh menu select the option Global coarseness. In the window that appears,
set the mesh coarseness to Medium and click on the Generate button, which will present
the following FE mesh composed of 15-node elements.
Computational Geotechnics 7
369
Geotextile reinforced embankment
• Select the clay layer (this consists of two clusters, see also hint) and press Refine cluster
from the Mesh menu. This will result in a refinement in the clay layer that will be needed
for the consolidation analysis. See figure 5.
Close the window showing the generated mesh and continue to the Calculations program.
8 Computational Geotecnics
370
Geotextile reinforced embankment
CALCULATION
The calculation consists of two alternatives for the construction of the embankment: without
and with consolidation taken into account. After both alternatives the factor of safety is
determined. In the calculations list 8 phases are needed, 4 phases for each alternative. First
start with the fully undrained construction, that is without taking consolidation into account.
When starting Plaxis Calculations, choose Classical mode.
Initial conditions
• Select the initial phase in the phase list and then press the Define button on the Para-
meters tabsheet in order to define the initial phase. The input window now opens in
Staged Construction mode.
• Deselect all material clusters and geotextile elements that are not present at the start of
the analysis. As we want to model the entire construction sequence from the beginning,
switch off:
– Geotextile elements
– Material clusters for the fill
– Material cluster for retaining bank
• Now continue to the Water conditions mode by clicking the equally named button.
• Enter a phreatic level at ground level by two coordinates (0, 5.5) and (33, 5.5). Click on
the Water pressures button to generate the pore pressures.
Computational Geotechnics 9
371
Geotextile reinforced embankment
After this, we will construct the embankment taking into account consolidation:
10 Computational Geotecnics
372
Geotextile reinforced embankment
Computational Geotechnics 11
373
Geotextile reinforced embankment
INSPECT OUTPUT
In order to get a good idea of the displacement mechanism, one can view the contours
of incremental displacements. Figure 6 shows this plot of the final calculation step for the
undrained construction. It clearly shows the effect of the geotextile reinforcement. Figure 7
shows the incremental displacement for the consolidated construction. Here the embankment
has a more gradual settlement without showing an upcoming failure mechanism.
The axial forces of the geotextile can be visualised by double clicking on the geotextile. This
will first present the displacement of the geotextile. On using the menu item Forces, one can
select Axial forces N.
At the ends of the geotextile the axial force must be zero, but due to the discretisation and
some numerical inaccuracy this is not completely achieved. The maximum axial forces is
approx. 8 kN/m. figure 9 shows the axial forces for the consolidated construction. The
maximimum axial force here is only 5-6 kN/m.
Finally, the factors of safety are checked. In order to do so follow these steps:
• Start the curves manager by selecting the Curves manager option from the Tools menu.
12 Computational Geotecnics
374
Geotextile reinforced embankment
• In the curves manager (see figure 10) select New in the Charts tabsheet. This presents
the Curve Generation window as shown in figure 11.
• On the x-axis we want to show the displacements of the point at the toe of the embank-
ment, hence choose Point A and Deformations → Total displacements → |u|.
• On the y-axis we want to show the strength reduction factor, hence select Project and
Multiplier → ΣM sf on the y-axis.
The created curve indicates a safety factor around 1.4 for the undrained construction and a a
safety factor of 2.1 for the consolidated construction of the embankment, as can be seen in
figure 12.
From the graph above, the factor of safety can be determined. Always look for a steady state
solution (slight variations in the load multipliers, increasing displacements). In most case, the
phi/c reduction calculation shows some variation at the beginning of the calculation. Note
that the displacements resulting from a Safety analysis are non-physical. Hence the total
displacements are not relevant. An incremental displacement plot of the last step, however,
shows the failure mechanism that corresponds the calculated value for ΣM sf .
Addicionally, figures 13 and 14 show the failure mechanisms with the lowest factor or safety
for both the undrained and consolidated case.
Computational Geotechnics 13
375
Geotextile reinforced embankment
Consolidated: ΣMsf=2.1
Undrained: ΣMsf=1.4
14 Computational Geotecnics
376
Geotextile reinforced embankment
Computational Geotechnics 15
377
Geotextile reinforced embankment
FACTORS OF SAFETY
The factors of safety are checked with the Curves program, see figure 19.
16 Computational Geotecnics
378
Geotextile reinforced embankment
Consolidated: ΣMsf=1.4
Undrained: ΣMsf=1.1
Computational Geotechnics 17
379
Geotextile reinforced embankment
18 Computational Geotecnics
380
Geotextile reinforced embankment
Computational Geotechnics 19
381
Basic concepts of PLAXIS 3D
Plaxis 3D Input
General toolbar
Mode switches
Selection explorer
Drawing area
Model explorer
Mode toolbar
Command line
382
Plaxis 3D Input : Modes
SOIL STRUCTURES
383
Command line
• All the actions carried out using either the mouse or the explorers are
translated into commands.
• Alternatively, PLAXIS 3D allows to carry out actions using keyboard
input by directly typing the corresponding commands in the command
line.
– The Session tab displays the commands executed in the active session
– The Model history tab displays all the commands executed in the project
Other functionality
• Selection by either clicking individual objects or at once by defining a selection
box in the draw area. Criteria can be applied to the type of items to be
selected.
• Group creation for fast model creation when the same operations have to be
undertaken over a large number of objects
384
Soil Mode
Soil mode
Borehole 4
• Definition of soil volumes and initial water levels Borehole 3
• Based on the concept of boreholes Borehole 1
385
Borehole definition
• Defining soil layer heights in the Soil
layers tabsheet
• Defining water conditions in the
Water tabsheet:
– Specific Head
– Hydrostatic distribution,
– Interpolate from adjacent layers
– Dry
– User-defined pore pressures
• Defining Initial Soil conditions in the
Initial conditions tabsheet
– Specify OCR, POP, K0x and K0y for
the K0 procedure
Material Sets
386
Structures Mode
387
Defining the geometry: Array
• Generate multiple copies of a selection, arranged in a rectangular pattern
388
Defining the geometry: Other functions
• Decompose into surfaces: creates outer surfaces of selected volumes
• Decompose into outlines: creates contour of selected surfaces
• Intersect: splits selected geometric objects along their intersection
• Combine: merges selected geometric objects of the same kind
• These functionalities are only accessible from the RMB context menu in
the draw area
Loads
• Generated by clicking the Create load button or by right-clicking on
any geometric objects in the draw area
Point load
Line load
Surface load
389
Default Boundary Conditions
Structures
390
Importing Geometry
• Possibility to import from external sources in different formats like
– 3D Studio files (*.3DS)
– AutoCAD native (*.DWG)
– Interchange (*.DXF) file format
Mesh Mode
391
Mesh density in Plaxis 3D
• Global coarseness:
– Defines an average element size based on model dimensions and relative
element size factor Re (Very coarse / Coarse / Medium / Fine / Very Fine)
– Reference element size = 0.05 * Re * (Model diagonal length)
• Local refinement (Fineness factor):
– Element size can be locally refined or coarsened
– Element size = (Fineness factor) * (Global coarseness) * (Ref. elem. size)
• Color code
– Depending on their degree of local refinement, geometric objects are
displayed in different color in the Draw area (gray for Fineness factor of 1
and green otherwise with darker colour when getting more refined)
392
Generate Mesh
393
Water Levels
394
The Staged Construction Mode
Phase explorer
395
The Phase explorer
• For creating and editing the calculation phases
396
397
Working in the Geometry Modes
of Plaxis 3D
William WL Cheang
Plaxis AsiaPac
Introduction
• The Geometry modes of Plaxis 3D comprises the Soil mode and the
Structures mode
• They are meant to fully define the model geometry in terms of:
– Soil stratigraphy
– Structural elements
– Soil structure interfaces
– Loads
– Boundary conditions
• The Geometry modes are indicated using blue tabsheets and precede the
Calculation modes (green tabsheets) when building up a model from
scratch
398
The Soil Mode
399
Soil Mode Toolbar
Create Borehole
Import Soil …
Show Materials
400
Boreholes Borehole 4
Borehole 3
• Boreholes are locations in the draw
area at which the information on the Borehole 1
height of the constitutive soil layers
and location of the water table is Borehole 2
given
• If multiple boreholes are defined,
PLAXIS 3D will automatically
interpolate between boreholes and
derive the corresponding position
and height of the soil layers from the
available borehole information.
• Each defined soil layer is used
throughout the whole model contour
Creating Boreholes
401
Defining Water Conditions
• Water conditions can be specified
from the Modify soil layers dialog box
in the Water tabsheet
• Available options are
– Head
– Hydrostatic
– Interpolate
– Dry
– User-defined
402
Importing Top and Bottom Model Surfaces
• Definition of the top and
bottom soil layer surfaces can
be achieved as a result of
surface import operation
• Import formats include 3DS,
ITS, DWG, DWF and SLT
• Feature available for VIP
members only
Importing Soils
• The geometry of the soil
can be imported from
predefined files instead of
using the Borehole tool
• The same import formats
as for importing soil
surfaces are available
(3DS, ITS, DWG, DWF and
SLT)
• Feature also only
available for VIP members
403
Material Sets
• Definition of material model parameters
for constitutive soil layers and
structural elements
• Dialog box consistent among all Plaxis
products
• Available from many places in Plaxis
3D:
– Present in each model
– Many shortcuts from different dialog
boxes where material sets assignment
is required
• Entry point to the SoilTest facility
404
Introduction to Structure Mode
• Meant to define structural
elements and loading
405
Changing Movement Limitation Settings
• Selecting one the six sides view
of the default views gives access
to the Movement limitation dialog
boxes
406
Points
• Can be generated by cilcking the Create point button
• The following items can be assigned to a point
– Point load
– Point prescribed displacement
– Fixed-end anchor
Point load
Fixed-end anchor
Lines
• Can be generated by clicking the Create line button
• The following items can be assigned to a line
– Beam
– Line load
– Line prescribed displacement
– Node-to-node anchor
– Embedded pile
407
Surfaces
• Can be generated by clicking the Create surface buttom
• The first three created points define the surface plane by default
• Existing surfaces can be edited from the Surface points dialog box of
from the pop-up submenu of Create surface button
Mode points/lines
Insert points
Delete points
Rotate
• Rotate any selection around a rotation point compared to global axis
• Possibility to directly rotate the selected objects from the draw area
using Euler angles
408
Extrude
• Lines and surfaces can be extruded to create surfaces and volumes
correspondingly:
– From the Extrude dialog box
– By dragging and dropping the bottom surface to the top surface location
Array
• Generate multiple copies of a selection, arranged in a rectangular pattern
409
Further Geometrical Operations
• On top of the aforementioned geometrical operations (rotate, extrude and
array), Plaxis 3D also offers
– Decompose into surfaces: create outer surfaces of selected volumes
– Decompose into outlines: create outer lines (including points) of selected
surfaces
– Intersect: Split select geometric objects along their intersection
– Combine: Merge selected geometric objects of the same kind
• These functionalities are only accessible from the RMB context menu oin
the draw area
Loads
• Can be generated by clicking the Create load button or by invoking the
RMB context menu on any geometric objects in the draw area
Point load
Line load
Surface load
410
Prescribed Displacements
• Can be generated by clicking the Create prescribed displacements
button or by invoking the RMB context menu on any geometric objects
in the draw area is a very similar way as for Load definition
411
Structures
• Can be generated by clicking the Fixed-end anchor
Structure button
Beam
• Can also be created from the RMD
context menu after selecting: Node-to-node anchor
– points (fixed-end anchor) Embedded pile
– lines (beam, node-to-node anchor
or embedded pile) Plate
Positive interface
Negative interface
Importing Geometry
• Possible to import from external sources in different formats like 3D
Studio files (*.3DS), AutoCAD native (*.DWG) and interchange (*.DXF) file
formats:
– Click to import surface
412
Working in the Calculation Modes of Plaxis 3D
William Cheang
Plaxis AsiaPac
Introduction
• The Calculation modes of Plaxis 3D comprises the Mesh mode, the Water
Levels mode and the Staged Construction mode
• They are meant to fully define the model geometry in terms of:
– The finite element mesh
– Changes in water pressure distribution
– Construction phases
– Calculation settings
• The Calculation modes are indicated using green tabsheets and follow the
Geometry modes (blue tabsheets)
413
The Mesh Mode
414
Mesh Density Definition in Plaxis 3D
• Global coarseness:
– Define an average element size based on model dimensions and relative
element size factor Re (Very coarse / Coarse / Medium / Fine / Very Fine)
– Ref. Elem. Size = 0.05 * Re * Model Diagonal Length
• Fineness Factor:
– Element size could be locally refine or coarsen
– Element size = Fineness Factor * Global Coarseness * Ref. Elem. Size
• Color code
– Depending on their Fineness Factor, geometric objects are displayed in
different color in the Draw Area (gray for fineness factor of 1 and green
otherwise with darker color as getting more refined)
• Refine mesh
Coarsen mesh
Reset local coarseness
Generate mesh
View mesh
415
Changing Local Mesh Density
• By clicking the toolbar
buttons Refine mesh or
Coarsen mesh and
selecting the desired
items in the Draw Area
Generate Mesh
• Can be done by clicking the
Generate mesh button from the
Mesh Toolbar or the RMB
context menu
416
View Mesh and Select Points
• The View mesh and the
Select points for curves will
open the Output Program
where
– mesh could be evaluated
– nodes and/or stress points
could be selected at the
location of which Output
needs to be generated
417
Water Levels
• Generated water level is created by specifying a Head in the boreholes (in
the Soil mode) and is the default water level
– A single borehole can be used to create a horizontal water surface that
extends to the model boundaries.
– When multiple boreholes are used, a non-horizontal water surface can be
created by combining the heads in the various boreholes
– Non hydrostatic distribution in the soil may be specified in the Water
tabsheet of theModify soil layers dialog box
• User water levels is available in the Water levels mode and can be defined
as a alternative to Generated water level
• Enable preview of
generation of water
pressure
418
Assigning Water Conditions to Soil Volume
• Water levels can be specified
for each individual volume
• Available options are:
– Global level (default)
– Custom level
– Head
– User-defined
– Interpolate
– Dry
• Can be done from
– The WaterConditions feature
in the Selection explorer.
– The RMB context menu
419
Introduction to the Staged Construction
Mode
• To define relevant
construction stages and
Deactivate
Show materials
Calculate
420
Activate/Deactivate
• Indicates which model parts
should be active or inactive in
the each construction stage
• Can be defined through
– RMB context menu
– The selection or model
explorer
• Selection tool can be
advantageously used for fast
selection of large number of
objects
• Be careful with hidden objects
which might still be active
421
Defining Calculation Stages
• Calculation phase settings can be
edited:
– Phase type
– Phase parameters
• Phase types are:
– Initial stress definition
– Plastic calculation
– Safety factor analysis (phi-c
reduction)
– Consolidation analysis (with closed
or open flow model boundaries)
Phase Parameters
• Advanced phase parameters
can also be controlled from the
Phases dialog box
– Maximum number of
calculation steps
– Reset displacement
– Updated Mesh
– Ignore undrained behaviour
– Parameters for automatic step
size
422
423
Modelling piles in PLAXIS 3D
Learning objectives
• To be able to:
– Model piles in different ways
– Analyse pile forces
• To understand the backgrounds of the embedded pile model
• To recognize the possibilities and limitations of pile modelling
424
Outline
• Volume piles
• Embedded piles
– Concept
– Model
– Properties
– Deformation behaviour
– Elastic region
– Output
• Verification & validation
– Axial loading, pile groups, lateral loading
• Further research
Volume piles
Volume piles:
425
Volume piles
Volume piles:
• Import
cylinder
Volume piles
Volume piles:
426
Volume piles
Volume piles:
Embedded piles – Concept
Sadek & Shahrour (2004):
A three dimensional embedded beam element for reinforced geomaterials
Beam arbitrarily through volume elements
Shear interaction between beam element and surrounding soil.
Septanika (2005)
A finite element description of embedded pile model
with limiting
Shaft interaction similar capacities
to Sadek (optional)
& Shahrour (2004)
NEW: - Tip interface
- Shaft interface
427
Embedded piles – kt Model t
kn Skin stiffness:
ks tmax
ks : axial stiffness
pile
kt Kn ,kt : lateral stiffness k
1
kn Skin tractions: urel
ks
ts = qs/length = ks (uspile‐ussoil) ≤ tmax
tskin kt tn = qn/length = kn (unpile‐unsoil)
tt = qt/length = kt (utpile‐utsoil)
Ffoot
kn
soil ks
Base stiffness:
s kb : base/foot stiffness
t Base/Foot force:
Fb = kb (ubpile ‐ ubsoil) ≤ Fmax
kb
n
Embedded piles – Model
Embedded piles:
428
Embedded piles – Properties
Properties (in explorer):
Connection:
• Rigid
(only at beams / plates)
• Hinged
• Free
429
Embedded piles – Properties
Material set with embedded pile properties:
Embedded piles
Bearing Capacity=
½ (Ttop+Tbot)×Lpile + Fmax
Ttop
Lpile
Tbot
Fmax
430
Embedded piles – Deformation behaviour
• Pile bearing capacity is input and not result of FEM calculation
F t
tmax
Specified bearing capacity k
1
urel
Global pile response F
from soil modelling
and pile‐soil interaction Fmax
k
1
u urel
.
Embedded piles – Without elastic region
Load-Displacement Curves - Vertical Pile
EB+CS
1250
1000
750
Load (kN)
500
FINE MESH
MEDIUM MESH
250
COARSE MESH
0
0 50 100 150 200 250 300 350 400
Displacement (mm)
431
Embedded piles –
.
Elastic Region
• Around shaft
• Around foot
Soil stress points inside elastic region are forced to remain elastic
Embedded piles – Output
Displacements, bending moments, axial forces, shaft friction, foot force
u N Ts
C B
A
432
Verification & validation
Verification & validation by Plaxis, METU, TUGraz, TUDelft *
Verification & validation – Axial loading (Plaxis)
433
Verification & validation – Axial loading (Plaxis)
Verification & validation – Axial loading (METU)
Pile load test Alzey Bridge near Frankfurt (Bored Pile)
434
Verification & validation – Axial loading (METU)
Alzey Brigde Single Pile Load Test
3500
PILE CAPACITY
3000
2500
2000
Load (kN)
Skin Friction
PILE CAPACITY
HS-CS
500
HS-CS-Base Res.
Verification & validation – Pile groups (TUDelft)
Pile group example by Poulos:
435
Verification & validation – Pile groups (TUDelft)
Verification & validation – Pile groups (TUDelft)
Average Settlement (mm) Moment (MNm/m)
50,0 1,2
45,0
1,0
40,0
35,0 0,8
30,0
25,0 0,6
FE Ta & Small
FE Ta & Small
Poulos & Davis
20,0
FE + BE Sinha
FE + BE Sinha
Plate (GASP)
Plate (GASP)
Plaxis 3D Fnd
Plaxis 3D Fnd
Strip (GASP)
Strip (GASP)
0,4
15,0
Randolph
10,0 0,2
5,0
0,0 0,0
Differential Settlement (mm) % Load on Piles
10,0
100,0
9,0
90,0
8,0
80,0
7,0 70,0
6,0 60,0
FE + BE Sinha
Plate (GASP)
FE Ta & Small
Plaxis 3D Fnd
Randolph
Strip (GASP)
5,0 50,0
FE Ta & Small
4,0 40,0
FE + BE Sinha
Plate (GASP)
Plaxis 3D Fnd
Strip (GASP)
3,0 30,0
2,0 20,0
10,0
1,0
0,0
0,0
436
Verification & validation – Axial loading (TUGraz)
Verification & validation
437
Verification & validation – Axial loading (TUGraz)
Conclusions from research at TUGraz (based on 3D Foundation):
• Embedded pile gives good results in serviceability states
• Layer-dependent option preferred to obtain realistic shaft friction
• Increased interface stiffness needed at pile tip *
• Pile should end at corner node *
* Implemented in PLAXIS 3D
Verification & validation – Lateral loading (TUDelft)
Validation for lateral loading:
• Comparison with volume pile
• Lateral movement of pile in horizontal soil slice
• Lateral loading of pile top
• Lateral loading by soil movement (embankment construction)
• Comparison with measurements from centrifuge test
• Lateral loading by soil movement (embankment construction)
438
Verification & validation – Lateral loading (TUDelft)
Lateral movement of pile in horizontal soil slice:
> Embedded pile almost behaves as volume pile due to elastic region
Verification & validation – Lateral loading (TUDelft)
Lateral loading by soil movement due to embankment construction
> Bending moments in reasonable agreement with measurements
439
Verification & validation – Lateral loading (TUDelft)
Conclusions from research at TUDelft:
• Embedded piles have capabilities for lateral loading behaviour in case
of rough pile-soil contact (full bonding) and small soil displacements
• When using ‘standard’ mesh around embedded piles (no local
refinement), stiffness and lateral capacity are over-estimated (~30%)
Further research (TUDelft)
Research at TUDelft on pile installation effects:
• Press-replace technique to simulate pile installation with the purpose
to generate data for different situations
• Results are used in generalized model, where (embedded) piles are
‘wished-in-place’ and installation effects are ‘superimposed’
440
Conclusions
• Volume pile
– Pile composed of volume elements or wall elements with pile prop’s
– Massive piles or tubes (wall elements)
– Not feasible for many piles
• Embedded piles
– Efficient way to model different types of piles
– Validated for axial loading, pile groups and lateral loading
Conclusions (cont’d)
• Limitations of embedded piles:
– Primarily for bored piles (no installation effects)
– Primarily for serviceability states
– Mesh-dependency of results
– Full bonding considered in lateral movement
441
Modelling piles in PLAXIS 3D
Ronald B.J. Brinkgreve
442
Embedded Elements in Plaxis 3D
Embedded pile in Plaxis 3D
Embedded pile in Plaxis
3D can be used to
simulate piles to obtain
pile movement profiles,
internal forces readily.
HOWEVER, a good
understanding of the
behavior of embedded
pile in Plaxis 3D is critical
for the proper use of this
very useful element.
Illustration of a pile behind
excavation 2
443
Embedded pile in Plaxis 3D
An embedded pile = Beam element + Interface
element around the beam element to interact
with the surround soil elements.
As such, the definition of an embedded pile
element consists of 2 parts: properties of the
beam & properties of the interface element
(skin resistance and foot resistance).
Embedded pile in Plaxis 3D
For the definition of the
beam element part, it is
much the same as that
defined for the beam
elements.
Be cautious of non‐isotropic
pile (like H‐pile) with the
possibility of different pile
orientation, as will be
illustrated in the next slide!
4
444
Embedded pile in Plaxis 3D
Embedded pile in Plaxis 3D
Illustration: Embeddedpile Orientation.P3D
For a higher resistance
to excavation‐induce
bending moment, the H‐
piles are oriented with
major axis to be bending
toward the excavation
side, has the embedded
pile orientation being
properly configured in
the right‐hand figure?
Illustration of a pile behind
excavation 6
445
Embedded pile interaction
with soil elements
An embedded pile can cross
a 10‐node tetrahedral soil
element at any place with
any arbitrary orientation,
introducing 3 extra nodes
inside the 10‐node
tetrahedral soil element.
An embedded pile crossing an
tetrahedral soil element
7
Embedded pile in Plaxis 3D
Interaction between
embedded pile and
surrounding soil element at
each node is based on:
embedded pile interacting with
surrounding soils
8
446
Embedded pile in Plaxis 3D
While interface element has been
provided along the embedded pile
shaft (good for correct simulation of
axial pile‐soil interaction), take note
that NO interface elements are
provided for the lateral soil sliding
around the embedded pile. As such,
for laterally loaded case, embedded
pile only works well in working
condition when there is not much
pile‐soil lateral sliding occurs, it can
not be used for simulation of embedded pile interacting with
ultimate lateral loading scenario. surrounding soils
9
Performance of embedded pile versus
solid pile in Plaxis 3D
10
447
3D FEM mesh for solid pile
The pile is simulated
by dia. 1m solid
cylindrical object with
surrounding interface
with Rinter=1
The dia. 1m solid
cylindrical object has
concrete elastic properties
with E=3.0E+7kPa 11
3D FEM mesh for embedded pile
Defining the beam
properties: The
embedded pile has the
structural properties
match exactly the dia.
1m bored pile
12
448
Calculated pile capacity
Defining embedded pile interface properties:
Skin resistance: Cu=100kPa, Rinter =1, Thus, Tmax =
3.14*1m*100kPa = 314kN/m.
End‐bearing resistance: qb = 9Cu = 900kPa, Fmax =
0.25*3.14*(1m)^2*900kPa = 706kN
So, Total shaft resistance Fshaft = 314kN/m * 20m = 6280kN
Total base resistance Fmax = 706 kN
13
Total pile resistance Ftotal = Fshaft + Fmax = 6990 kN
Axially loaded embedded pile
Fz =7000kN
=1000kN
=2000kN
=3000kN
=4000kN
=5000kN
=6000kN
14
449
Punching of embedded pile under
theoretical axial load
Fz =7000kN
15
Embedded pile toe resistance under axial loads
Fz =7000kN
16
450
Axial load transfer curves
Axial force (kN)
0 2000 4000 6000 8000
0
‐5
Elevation (m)
‐10
Embeddedpile Fz=1000kN
‐15
Embeddedpile Fz=3000kN
‐20 Embeddedpile Fz=5000kN
706kN Embeddedpile Fz=7000kN
17
‐25
Shaft skin resistance of embedded pile
0 20 40 60 80 100 120
0
‐5
Elevation (m)
‐10
Embeddedpile Fz=1000kN
‐15 Embeddedpile Fz=3000kN
Embeddedpile Fz=5000kN
‐20
Embeddedpile Fz=7000kN
‐25
Shaft friction (kPa) 18
451
Axial load‐settlement behavior
10
20
Pile settlement (mm)
30
40
50
60
70
80 Embeddedpile
EmbeddedpileSolid pile
90
100 19
Axial load (kN)
Axial load transfer curves
Axial force (kN)
0 2000 4000 6000 8000
0
‐5
Elevation (m)
‐10
Embeddedpile Fz=1000kN
Embeddedpile Fz=1000kN
Embeddedpile Fz=3000kN
‐15 Embeddedpile Fz=5000kN
Embeddedpile Fz=3000kN
Embeddedpile Fz=7000kN
Solid pile (1000kN)
‐20 Solid pile (3000kN)
Embeddedpile Fz=5000kN
Solid pile (5000kN)
Solid pile (7000kN)
Embeddedpile Fz=7000kN
20
Solid pile (8000kN)
‐25
10
452
Comparison of skin friction
0 20 40 60 80 100 120
0
‐5
Elevation (m)
‐10 Embeddedpile Fz=1000kN
Embeddedpile Fz=3000kN
Embeddedpile Fz=5000kN
‐15 Embeddedpile Fz=7000kN
Solid pile (1000kN)
Solid pile (3000kN)
‐20
Solid pile (5000kN)
Solid pile (7000kN)
Solid pile (8000kN)
‐25
Shaft friction (kPa) 21
Concluding remarks: Embedded pile performs
satisfactorily under axial loads and conform to
theoretically values, while solid pile exhibit too high end
bearing resistance and much stiffer pile response near
ultimate loading condition, and thus needed to be used
with cautions.
Can tension loading case eliminate the end‐bearing
difference?
Total shaft resistance Fshaft = 314kN/m * 20m = 6280kN
Self‐weight of bored pile = 0.25*3.14*(1m)^2 * 20m *
24kN/m^3 = 377 kN
So, expected total pull out resistance = 6280 + 377 = 6660kN
22
11
453
Can tension loading case eliminate
the end‐bearing difference?
‐50
‐40 Embededpile_use actual load
Embededpile_use actual load
Pile movement (mm)
Solid pile
‐30
‐20
‐10
0
0 1000 2000 3000 4000 5000 6000 7000 8000 23
Tension load (kN)
Load transfer curves under tension loadings
Tension force (kN)
‐8000 ‐7000 ‐6000 ‐5000 ‐4000 ‐3000 ‐2000 ‐1000 0
0
‐5
Elevation (m)
‐10
Embeddedpile Fz=1000kN
Embeddedpile Fz=1000kN
Embeddedpile Fz=3000kN
Embeddedpile Fz=5000kN ‐15
Embeddedpile Fz=3000kN
Embeddedpile Fz=6000kN
Solid pile (1000kN)
Solid pile (3000kN)
Embeddedpile Fz=5000kN
‐20
Solid pile (5000kN)
Solid pile (7000kN)
Embeddedpile Fz=6000kN
Solid pile (8000kN) 24
‐25
12
454
In general, embedded pile performs satisfactorily
under both axial compression loads and tension
loads and generally conform to theoretically
values, while solid pile exhibit too high end
bearing resistance and much stiffer pile response
near ultimate loading condition, and develop very
large suction at the toe of solid pile under tension
load which may not be so reliable, and thus
needed to be used with cautions.
How about the performances under lateral loads?
25
Estimation of lateral pile capacity assuming pile is rigid enough
and has sufficiently high strength, and failure occurs in the clay
Brom's Theory:
Free head, L/d=20, e/d=0, Hu/cud^2=60
26
Hu=6000 kN
13
455
Lateral loading on pile in Plaxis 3D
27
Lateral load – movement curves
400
350
Embeddedpile
Solid pile
300 Solid pile
Lateral pile movement (mm)
250
200
150
100
50
0
0 1000 2000 3000 4000 5000 6000 7000
28
Lateral load (kN)
14
456
Embedded pile in Plaxis 3D
Take note that NO interface element
provided for the lateral soil sliding
around the embedded pile. As such,
for laterally loaded case, embedded
pile only works well in working
condition (FOS=2.0~3.0) when there
is not much pile‐soil lateral sliding
occurring, it can not be used for
simulation of ultimate lateral
loading scenario.
embedded pile interacting with
surrounding soils
29
BM under lateral loadings
14000
12000 Embeddedpile
Solid pile
Pile bending momnet (kNm)
10000
8000
6000
4000
2000
0
0 1000 2000 3000 4000 5000 6000 7000
30
Lateral load (kN)
15
457
Comparison of pile deflection profiles
‐20 0 20 40 60 80 100 120
0
‐2
‐4
‐6
Elevation (m)
‐8
Embeddedpile Fx=1000kN
Embeddedpile Fx=1000kN
‐10
Embeddedpile Fx=2000kN
Embeddedpile Fx=2000kN
‐12 Embeddedpile Fx=3000kN
Embeddedpile Fx=3000kN
Embeddedpile Fx=4000kN
‐14
Solid pile Fx=1000kN
Solid pile Fx=1000kN
‐16 Solid pile Fx=2000kN
Solid pile Fx=2000kN
Solid pile Fx=3000kN
‐18
Solid pile Fx=3000kN
Solid pile Fx=4000kN
‐20
Lateral pile movemetn (mm) 31
Comparison of BM profiles
‐2000 0 2000 4000 6000 8000
0
‐5
Elevation (m)
‐10
Embeddedpile Fx=1000kN
Embeddedpile Fx=1000kN
Embeddedpile Fx=2000kN
Embeddedpile Fx=2000kN
‐15 Embeddedpile Fx=3000kN
Embeddedpile Fx=3000kN
Embeddedpile Fx=4000kN
Solid pile Fx=1000kN
Solid pile Fx=1000kN
‐20 Solid pile Fx=2000kN
Solid pile Fx=2000kN
Solid pile Fx=3000kN
Solid pile Fx=3000kN
Solid pile Fx=4000kN
‐25
32
Lateral pile movemetn (mm)
16
458
Conclusions
• Embedded pile is a good model of single pile response
subjected to both vertical compression loads and
vertical pull out loads.
• Solid pile may give very large end bearing resistance
when subjected to vertical compression load, and
generate large suction force at the base when subjected
to pull out load, and thus must be used with caution.
• Under working load condition with FOS=2~3 when there
is no much pile‐soil lateral slide occurring, embedded
pile give very reasonable pile deflection and pile bending
moment. However, embedded pile can not be used for
simulation of ultimate lateral loading scenario.
33
17
459
E7: Exercise on Piled Raft
Analysis
Based on an actual project: Pile Foundations for Flieden Bridge in Germany
Briefing of the Project
460
Briefing of the Project
y
X = ‐35m
Trench to 35m
X
section
coordinates
(‐4 ‐25 0)
(‐.4 ‐25 ‐.8) z y = ‐25m
(.4 ‐25 ‐.8) X to 25m
(4 ‐25 0)
z = 0 to
‐30m
3
Briefing of the Project
The subsoil consists mainly of tertiary
formations of highly plastic clay with lenses of
lignite coal (clay with brown coal). In this
analysis, a uniform clay layer was idealized with
OCR = 1.3
4
461
Soil parameters
Simulation in Plaxis 3D
Step 1: General setting
Step 2: Add in a borehole
Step 3: Define soil properties
Step 4: Create 6 piles
Step 5: Create 1 pile cap
Step 6: Clone another pile group
Step 7: Create the trench
Step 8: Assign vertical loads
Step 9: Generate mesh with refinement
Step 10: Define stages and view results 6
462
Flieden bridge piled-raft foundation
FLIEDEN BRIDGE
PILED-RAFT FOUNDATION
Computational geotechnics 1
463
Flieden bridge piled-raft foundation
INTRODUCTION
The foundation of the 4-span railway bridge of Flieden in Germany (figure 1) was the first
railway bridge in Germany founded on piled rafts.
The subsoil consists mainly of tertiary formations of highly plastic clay with lenses of lignite
coal (clay with brown coal). To ascertain the adequacy of the piles and determine appropri-
ate design values, pile load tests were first conducted on large diameter bored piles with and
without post shaft grouting (El-Mossallamy et al. 2003). These results conform to the me-
chanical sensitivity of the organic silty clay and lignite coal lenses. It was decided to install all
foundation piles applying post shaft grouting.
INPUT
The bridge piers are consisted of two pillars, each founded on a separate group of 6 piles
underneath a raft. The pile arrangements are shown in Figure 2. The rafts are 1.5 meters
thick and are embedded in the soil with the raft base at a depth of 2.3 meters below the soil
surface. The piles where designed with a diameter of 1.2 m and a length of 18 m. The pillars
transfer two working loads of 20 MN and 22 MN respectively from the superstructure to the
piled raft foundation.
Work flow
In this excercise the model is created in a specific order that has proven to be a rather efficient
way to create the model. Please note that many parts of the model can be created in any
other order as well and the work flow presented here is not the only correct method to create
the project. The work flow to create the project presented here is:
1. Enter dimensions of the project and some general visualisation options
2. Define the underground model using 1 borehole and the appropriate soil material sets
Computational geotechnics 2
464
Flieden bridge piled-raft foundation
4. Copy this 1 pile 5 times to create the 6 piles needed for 1 piled raft
5. Insert the raft, the lower column and the top load
6. Copy the complete piled-raft 1 time to create the second piled raft
8. Generate mesh
Computational geotechnics 3
465
Flieden bridge piled-raft foundation
Geometry
General settings
Start the PLAXIS 3D input program. A Quick select project dialog box will appear in which
you can select an existing project or create a new one; choose Start a new project so that the
Project properties window appears.
1. In the Project properties window on the Model tabsheet the size of the model contour
has to be set. In the Contour box fill in xmin = −35, xmax = 35, y min = −25 and y max = 25.
2. Close the Project properties window, the drawing area will now appear.
3. From the Options menu choose Visualization settings. A new window will open, contain-
ing 2 tabsheets: View and Visibility.
4. On the View tabsheet the grid point distance (Spacing) and number of snap intervals per
grid distance can be set. By default the Spacing is set to 1 m with only 1 snap interval
per grid distance. As can be seen from figure 2 many dimensions of this project have an
accuracy of 0.1 m and therefore just 1 snap interval per 1 m is not sufficient. Therefore,
set the Intervals to 10, this will results in having a snap distance of 0.1m (Spacing /
Intervals).
Subsoil
The first step in creating a model in PLAXIS 3D is the definition of the subsoil, which is done
using boreholes.
1. Select the Create borehole button ( ) and move the mouse to the origin of the system
of axis. Click at (x,y,z) = (0 0 0), this will open the Modify soil layers window.
2. In the Modify soil layers window click the Add button in order to define a new soil layer
in this borehole. Set the top of the borehole to 0.0 m and the bottom to -30.0 m.
3. In order to assign a material set to the newly defined model it is necessary to first define
a material set. To do so, press the Materials button ( ) to open the material
sets database.
4. Though the model only has one soil layer (clay) we will have to define two material sets:
the second material set will be used to represent the concrete needed for both raft and
piles. Therefore, create two material sets according to the material parameters specified
in table 1.
5. After defining the two material sets close the window by clicking OK in order to return to
the Material sets window.
Computational geotechnics 4
466
Flieden bridge piled-raft foundation
Dilatancy angle ψ 0 - o
Permeability kx , k y , k z 0 - m/day
Interface strength Rinter 0.6 (Manual) Rigid -
Coefficient for initial lateral stress K0 Automatic Automatic -
Overconsolidation ratio OCR 1.3 - -
6. Drag and drop the clay material set from the Material sets window onto the borehole. The
mouse cursor changes shape when the material set can be dropped. After dropping the
borehole should get the colour of the material set. Now close the Material sets window
in order to return to the Modify soil layer window.
7. In the Modify soil layer window directly above the graphical representation of the bore-
hole it is possible to specifty a general phreatic level for this borehole by changing the
Head value. In this project the water level is 0.5 meters below ground level, therefore
change the Head to -0.5 m.
8. Press OK to close the Modify soil layers window and return to the drawing area. In the
drawing area there is now a block of soil with the horizontal dimensions specified in the
Project properties window and a depth according to the borehole.
We have now finished defining the subsoil and we will continue defining the foundation. Press
the Structures option ( ) on the mode toolbar to move to Structures mode.
The two bridge foundations are equal with exception of the load from the bridge acting on
the foundation. Therefore it’s sufficient to define 1 foundation and then make a copy of the
Computational geotechnics 5
467
Flieden bridge piled-raft foundation
foundation to get the second. Similarly, each foundation is supported by six equal piles, hence
it is sufficient to define 1 pile and make 5 copies to model all piles to model 1 foundation.
In the current version of PLAXIS 3D the only possibility to insert a pile is by inserting a cylin-
derical volume using the command line. The syntax for inserting a cylinder is:
cylinder <R> <L> <num planes> (<start_x> <start_y> <start_z>) (<dir_x> <dir_y> <dir_z>)
In short, one specifies the radius (R) and length (L) of the cylinder, a set of 3 coordinates to
indicate the starting point of the cylinder and a vector to indicate the direction of the cylinder.
Special attention should be given to <num planes>. In PLAXIS 3D a cylinder is modelled with
a polygon cross section, hence <num planes> gives the number of sides of the polygon. The
higher the number the more accurate the polygon will represent the circular cross section.
9. Insert the first pile at (x,y) = (-8.4, -1.8). Note that the piles have a 1.2m diameter (hence
a radius of 0.6m), are 18 meters long, start at z = -2.3m and go down vertically, that is
in the negative z-direction. The number of planes is set to 15 to accurately model the
cylinderical shape. This results in the following cylinder command:
Type this command on the command line and press <Enter>. The cylinder is now in-
serted in the model as a volume.
10. In order to assign interfaces around the pile, the pile has to be split into its separate
surfaces. To do so, right click on the pile and from the popup menu choose Decompose
into surfaces.
11. Now select the outer surfaces of the pile, right-click and select Create negative interface.
This will create a negative interface along the outside of the pile.
12. In order to create an interface below the foot of the pile, select the bottom circular surface
of the pile. It is probably necessary to rotate the model in order to see the foot of the pile
from below. Right-click again and select Create negative interface to create the interface
below the foot as well.
Hint: Interfaces are drawn as planes at a certain distance from the surface they
belong to. Therefore, if a project requires a lot of interfaces it may become
difficult to see the underlying structure as the interfaces are surrounding it.
This can be solved by either reducing the distance between interface and
structure or by making the interfaces invisible.
The distance between interface and surface can be reduced in the
Visualization settings that can be found under the Options menu. On the
View tabsheet the field Interface size controls the distance. By default this
value is set to 1. Reducing this value will reduce the distance between
interface and surface.
Alternatively, in the Object explorer it is possible to make the interfaces
invisible by clicking on the small eye in front of the branch Interfaces (to
make them all invisible) or in front of individual interfaces (to make only a
selection of interfaces invisible).
Computational geotechnics 6
468
Flieden bridge piled-raft foundation
We have now finished creating the first pile. The next step is to make 5 copies of the pile to
create the group of 6 piles of the first foundation slab.
13. Click the button Select rectangle ( ), ignore the suboptions that become available.
Now draw a rectangle that fits the whole pile so that all parts of the pile are selected.
14. Now click the Create array button ( ) to specify the locations of the copies of the pile.
The Create array window appears, see figure 3.
In x-direction we need 3 piles with an intermediate distance of 3.4 meters and in the y-direction
we only need 2 piles with a distance of 3.6 meters in between.
15. Set the Shape of the array to 2D, in xy plane as we want to copy the piles in both x and
y direction, keeping the z coordinate constant
16. Fill in 3 columns with a distance of x = 3.4m in between and 2 rows with a distance of y
= 3.6m in between.
We have now created the 6 piles for one of the bridge foundations.
After creating the 6 piles now the raft has to be modelled on top of the piles, including the
lower part of the column supporting the bridge:
Computational geotechnics 7
469
Flieden bridge piled-raft foundation
1. From the horizontal button bar with general options, click the Top view button ( ). This
will show the model seen along the z-axis.
2. In the Movement limitation window that appears, fix the z-coordinate to z = -0.8m by
filling in -0.8 in the z-value field and clicking the Set button.
3. Select the Create surface button ( ) and draw the surface representing the top side of
the raft from (x y) = (-9.6 -3.0) to (-9.6 3.0), (-0.4 3.0) and (-0.4 -3.0).
4. Select the surface that has just been created and click the Extrude button ( ). In the
window that opens fill in an extrusion vector of (x,y,z) = (0 0 -1.5) in order to create the
1.5m thick raft and click OK.
Now the raft has been created as volume, in order to assign interfaces to all sides of the raft,
the raft volume has to be decomposed into its surfaces.
5. From the button bar with general options, click the Perspective view button ( ). .
6. Right-click on one of the vertical sides of the raft and select the option Decompose into
surfaces. This will created surfaces for all sides of the volume.
7. For all 6 sides, right-click on the side and add an interface. Note that all sides need
a negative interface with exception of the vertical side at y = 3.0m; this side needs a
positive interface. Check if all created interfaces are on the outside of the raft!
8. In order to make the lower part of the supporting column, click again the Top view button
and fix the z-coordinate to ground level.
9. Create a surface from (x y) = (-6.0 -1.0) to (-6.0 1.0), (-4.0 1.0) and (-4.0 -1.0).
10. Extrude the surface 0.8 meters downwards, hence in the negative z-direction. This cre-
ates the lower part of the column from groundlevel down to the raft.
12. For all 4 vertical surfaces created, create an interface on the outside. That is, negative
interfaces for all vertical sides but the vertical side at y = 1.0m. The latter side needs a
positive interface.
The only part missing now is the load representing both the weight of the bridge and a passing
train
13. Right-click on the top plane of the column, that is the plane at ground level.
14. From the popup menu that opens, select the option Create surface load to add the load.
Computational geotechnics 8
470
Flieden bridge piled-raft foundation
The second raft is equal to the first raft, hence creating the second raft is simply making a
copy of the first raft:
1. Click the button Select rectangle ( ), ignore the suboptions that become available.
Now draw a rectangle that fits the whole structure of piles, raft and column so that all
parts are selected.
2. Now click the Create array button ( ) to specify the location of the copy of the founda-
tion structure in the Create array window.
3. Set the Shape of the array to 1D, in x direction as we want to copy the foundation just
one time in x direction, keeping the y and z coordinates constant
4. Fill in 2 columns with a distance of x = 10m in between and press OK. Now the second
raft is created as copy of the first raft.
In order to be able to refine the mesh in the area around the rafts it is needed to define a
volume of soil around the rafts where a mesh refinement can be applied. To do so, follow
these steps:
Computational geotechnics 9
471
Flieden bridge piled-raft foundation
2. Draw a rectangular surface from (x y) = (-10.0 -4.0) to (-10.0 4.0), (10.0 4.0) and (10.0
-4.0).
3. Select the Perspective view, select the newly created surface and extrude it 25m up,
hence in the positive z-direction.
We have now created a volume around the foundation structure that we can use for local mesh
refinement.
Mesh generation
In the Mesh mode we will specify global and local refinements and generate the mesh. In order
to generate more accurate results a refinement of the mesh around the foundation structures
will be applied.
1. In the geometry click somewhere close to the origin. This will select the body of soil that
encloses the foundation structures.
2. In the Selection explorer on the left the selected soil body appears, showing a mesh
refinement factor of 1.0. Change this mesh refinement factor to 0.30.
3. Select the Generate mesh button ( ) in order to generate the mesh. The Mesh options
window appears.
4. In the Mesh options window choose a Very coarse element distribution and click OK to
start the mesh generator.
5. After mesh generation has finished one can already see an indication of the amount of
elements and nodes generated in the command line box below the draw area. For this
project about 22,000 elements should be generated.
After inspecting the mesh the output window can be closed. Mesh generation has now been
finished and so creating all necessary input for defining the calculation phases has been
finished.
Computational geotechnics 10
472
Flieden bridge piled-raft foundation
Computational geotechnics 11
473
Flieden bridge piled-raft foundation
CALCULATION
The calculation consists of the initial phase and three additional phases. Since water levels
will remain constant the Water levels mode can be skipped. Therefore, click on the Staged
construction mode button to move to the defintion of the calculation phases.
Initial phase
By default the Initial phase is set to the K0 procedure, which is fine for this example. No further
changes have to be made.
1. Click on the Add phase button ( ) to add the first calculation phase.
2. As the foundations are surrounded by soil they cannot be accessed directly. In order to
change their properties the surrounding soil has to be made invisible. To do so, right-
click on the soil somewhere far away from the origin and from the menu that pops up
choose Hide to hide the outer soil. Now only the foundations and the refinement zone is
left. Make sure the soil is hidden, not deactivated!
3. Right-click on the refinement zone volume and again choose the Hide option from the
popup menu. With the refinement zone hidden, only the foundations structures remain
visible.
4. Open the material sets database by clicking the Show materials button ( ). Drag and
drop the material set representing the concrete on all piles, the rafts and the two parts
of the column. When assigning the material set, the colour changes from the colour of
the material set representing the clay to the colour of the material set representing the
concrete.
5. In the Model explorer, activate all interfaces by clicking on the checkbox in front of the
interfaces branch so that a checkmark appears.
1. Click on the Add phase button ( ) to add the second calculation phase.
2. In the Model explorer open the Surface loads branch and change the value for the two
surface loads. Set the first surface load to a vertical stress of σz = −5000 kN/m2 (20 MN
dived by 4 m2 cross sectional area of the column) and set the second surface load to a
vertical stress of σz = −5500 kN/m2 .
3. Make sure the surface loads are activated, that is that they have a checkmark in the
Model explorer.
Computational geotechnics 12
474
Flieden bridge piled-raft foundation
1. Click on the Add phase button ( ) to add the third calculation phase.
2. In the Model explorer change the values of the Surface loads to σz = −10000 kN/m2 for
the first surface load and σz = −11000 kN/m2 for the second surface load.
Press the Calculate button ( ) to start the calculation. Ignore the message "No nodes or
stress points selected for curves" as we will not draw any load-displacement curves in this
example, and continue the calculation.
Computational geotechnics 13
475
Flieden bridge piled-raft foundation
OUTPUT RESULTS
Figure 6 demonstrates the calculated load settlement behaviour of the piled raft applying the
GAPR (Geotechnical Analysis of Piled Raft, El-Mossallamy 1996). Due to the non-linear re-
sponse of the foundation system the loads have been incrementally applied till the ultimate
limit state. Another aim of the analysis under working loads was to determine the pile/soil
stiffness and subgrade reaction distribution beneath the raft, which are necessary for the de-
sign of the foundation. However, within the framework of this exercise the subgrade reaction
distribution will not be checked. Figure 7 shows the measured settlements in comparison to
the calculated values
Figure 6: Load-settlement behaviour of the piled raft foundation (calculated by program GAPR,
El-Mossallamy)
Computational geotechnics 14
476
Finite Element Modelling of Tunnels and Tunnelling
William WL CHEANG
PhD (Geo) MSc PGDip BEng (Hons) (Civil)
Plaxis 2D & 3D
Contents
477
Modelling of Tunnels and Tunnelling in 3D
2. Construction stages
3. Modelling anchors
5. Conclusions
Geometric modelling issues
478
Geometric modelling issues (CM Line)
cylinder 4 100 48
• Decompose cylinder volume into surfaces
Geometric modelling issues
Hint: Draw cross section surface and use Extrude command to create shafts
PLAXIS 3D will automatically create intersections
479
Geometric modelling issues (CAD)
Geometric modelling issues (Import)
480
Construction stages (For sequence of events)
Modelling Anchors
Lineangles (x y z) 0 45 15
• Assign Beam feature to turn lines into anchors
• Alternatively, embedded piles can be used
481
Modelling Anchors
Modelling volume loss
482
Conclusions
2D & 3D MODELLING OF
TUNNELLING
Part 2
483
Part 2‐Outlines
A. 2D modelling of tunnelling
B. 3D modelling of tunnelling
Tunnelling Observations
484
Concept of Modelling Tunnelling in 2D
3D 2D 2D
Moller (2006)
Methods of Modelling Tunnelling in 2D
• Plaxis 2D provides
1. Lining Contraction Method
2. Stress Reduction Method (-method)
3. Applied Pressure Method (APM)
(from Grout Pressure Method by Moller & Vermeer, 2008)
485
Lining Contraction Method
1st 2nd
Phase Phase
Stress Reduction Method ()
1 Pk 1
Pk = initial ground radial
= pressure
Pk ΣMstage = 1 -
10
486
Applied Pressure Method (APM)
FE Prediction of Greenfield Surface Settlement
11
487
Modelling of Tunnelling in Hong Kong Soils
60
m 0
Fil -3
Marine l -6
Deposits
20m
40m
Completely Decomposed
6m Ø Granite (CDG)
tunnel
-
Rock 40
• Ground conditions: 3m Fill, 3m MD, 34m CDG & rock; GWT at surface
• Tunnel 6m diameter with axis at 20 mbgl; 2700 nos of 15-noded elements
Soils Modelled by Mohr Coulomb Model
E ν c' / cu '
Soil
(kN/m3) (MPa) [-] (kPa) (Deg)
Fill 19 20 0.3 0 30
MD
16 6 0.3 15 0
(Undrained)
CDG 20 39 0.3 5 35
Soils Modelled by HS & HS‐small Models
Fill
20 60 0.5 0 30 100 0.2 - -
(HS)
MD
(HS) 6 18 1 0 22 100 0.2 - -
(Und.)
CDG
5E
(HS- 39 117 0.5 5 35 200 0.2 200
-5
small)
12
488
Pre‐failure Stress‐strain Behaviour
1: Mohr Coulomb
Initial Stress Equilibrium
• K0 = 1 – sin'
' = drained effective friction angle (Fill=30°; MD=22°)
Soil K0 CDG
Fill 0.5
CDG 0.65
0.4 0.65 0.9
13
489
Details for Analyses
-2
Settlement (mm)
-4
Gaussian (K=0.5, VL 1%)
-6
Lining contraction - LC 1%, VL 0.32%
Mohr Coulomb Soil with Lining
Contraction
14
490
Comparison of MC and HS & HS‐small Models
-4
Gaussian (K=0.5, VL 1%)
-6
Lining contraction - LC 1%, VL 0.32%
-2
Settlement (mm)
-4
Gaussian (K=0.5, VL 1%)
-6
Lining contraction - LC 1%, VL 0.77%
Stress Reduction vs. Applied Pressure
Stress Applied Methods
Reduction Pressure
15
491
Summary of 2D Modelling of Tunnelling
Part 3.1
16
492
Presentation Outline
Three‐Stage Building Damage Assessment Due to Tunnelling
Burland (1995)
17
493
Tunnel‐pile‐soil Interaction
2. movement of piles in 3D
Zones of Influence
Pile
settlement
C BA
Depth
18
494
Analysis of Tunnel‐pile Interaction
Example of Tunnelling Below Piled Building
25m
25m
0 mbgl P4 P5 Rear P6
2m Pile cap
5 mbgl Fill 9m 10m
1m
4m
MD
10 mbgl P1 P2 Front P3
CDG 10m 6m Ø tunnel
1m
4m
Tunnel 2m Ø pile
6m Ø Pile design load 15MN (~5MPa)
30 mbgl
31.5 mbgl Rock 3m Ø bell-out
P1/P4 P2/P5 P3/P6
19
495
Information for Tunnel, Piles & Ground
Soil Small Strain Non‐linear Stiffness
0.01 0.1% 1%
%
20
496
CDG Small Strain Non‐linear Stiffness
1600
1400 Triaxial_Upper
Triaxial_Low er
1200
HSsmall_Upper
1000
Adopted
HSsmall_Low er
Gsec /p'
800 HSsmall_Baseline
600
400
line
200
0
0.0001 0.001 0.01 0.1 1 10
Shear strain (%)
3D Finite Element Model (Plaxis‐GiD)
Rear Load 15 MN
Building “Plate”
40m Pile cap
Bored pile
Front
Fill Tunnel face
Tunnel MD 149m
CDG
120m Rock TBM
length
Bell-out
43,000 elements Linings
21
497
Tunnel Confinement Pressure
A
PIII A. Face support pressure (PI to PII) =
PIV
PI hydrostatic pore pressure +
Rear Front overpressure
6m Ø TBM shield 9m B. Along TBM shield, tunnel support
pressures vary to consider
PII 1. Conical shape of TBM shield /
PVI over-cutting
2. Ground loss into tail void in rear
PV C. Any combination of support pressure
A
profiles can be modelled
PIII
Pressure
PV increases
Section A-A with
depth
Modelling of Tunnel Face Advance
TBM shield
(elements nulled) deactivated
2. Apply tunnel support pressure
profiles
1.5 1.5m 3. For each face advance, shift tunnel
support pressures forward &
correspondingly erect new lining
Lining
Lining
TBM shield
(elements nulled)
1.5 1.5m
22
498
Modelling of Structures
Prediction on Ground Surface Settlement
-8
VL 1.61%
-12
Mid-building
CDG
-16
Greenfield
Tunnel -20
Gaussian
-24
23
499
Prediction on Pile Transverse Displacement
Overpressure 20 kPa
Rear +2D
10
+2D
Depth (mbgl)
15
+10D Rear
20 1m P2
Front
25
30 -2D
35
Tunnel advance
Prediction on Pile Longitudinal Displacement
Overpressure 20 kPa
10 +2D
Depth (mbgl)
15 Rear
-2D 20 1m P2
Tunnel
Front Front
advance 25
Rear
+2D
30 -2D
+10D 35
Tunnel advance
24
500
Prediction on Pile Settlement & Axial Force
Overpressure 20 kPa
Settlement (mm) Increase in axial force (MN)
0 -1 -2 0 1 2 3 4
0 0
P2 -2D P2 -2D
5 Front 5 Front
Rear Rear
10 +2D 10 +2D
+10D
Depth (mbgl)
Depth (mbgl)
+10D
15 7
15
20 20
25 25
A B
30 30
C
35 35
Pile toe
Prediction on Pile Bending Moment
Overpressure 20 kPa
Transverse moment (kNm) Longitudinal moment (kNm)
1500 500 -500 -1500 1500 500 -500 -1500
0 0
P2 P2
5 5
10 -2D 10 -2D
Depth (mbgl)
Depth (mbgl)
Front Front
15 15
Rear Rear
+2D +2D
20 20
+10D Tunnel +10D
25 25advance
30 30
35 35
25
501
Check on Potential Structure Damage
OP 40kPa
25 OP 40kPa
_
Bldg. settlement (mm)
-0.4 0.2
/L (%)
15
5 -0.8
0.1 Cat.
-5 =0.14 2 3
-1.2
mm 1
-15 0.0 0
-15 -10 -5 0 5 10 15 0 0.1 0.2 0.3
Moment, M (MNm) -1.6 h (%)
Comparison with Closed Form Solution
Depth (mbgl)
15 CDG 15
CDG
20 20
25 25
Loganathan Loganathan
et al. (2001) 30 et al. (2001) 30
Rock
3D analysis Rock 3D analysis
35 35
26
502
3D FEA vs. Analytical Solution
3D FEA vs. Analytical Solution
27
503
TUNNELLING BENEATH A BUILDING
SUPPORTED BY PILES
Part 3.2
6m ø P8 2m
tunnel
P7
P4
P6
P3 P5 P6 P7 P8
P5 23m
P2 P1 P2 P3 P4
Bldg. footprint
P1 33×11m
y x
3m y
z 2m ø piles
31 mbgl
Tunnel advance direction z x
6m ø tunnel
28
504
3D Model by Plaxis-GID
Pile cap + “Plate”
45m
Building 23m
Pile
Settlement & Axial Force Pile
A B
P5
P5 C
Rear
cap
Front
cap Settlement, Uy (mm) Pile axial laod, N (MN)
0 -5 -10 -15 -20 -25 -30 0 2 4 6 8 10 12
0 0
-5 -5
Depth (mbgl)
Depth (mbgl)
-10 -10
-20 -20
-25 -25
Front Rear +2D +15D Front Rear +2D +15D
29
505
Channel Tunnel Rail Link (UK) Measurement
Selementas (2005)
2m
Approachin
Final Initia g
l
• Tunnel 8.15m dia. at 19m depth
• Driven cast-in situ pile 0.48m dia.
• Layered ground
Longitudinal Horizontal Displacement &
P5
Rear
Bending Moment Pile P5
cap
Front Longitudinal hor. disp., Uz (mm) Longitudinal moment, Mz (kNm)
cap -1 0 1 2 3 -400 -200 0 200 400 600
0 0
-5 -5
Depth (mbgl)
Depth (mbgl)
-10 -10
-15 -15
Tunnel
-20 -20
advanc
e
-25 -25
Front Rear +2D +15D Front Rear +2D +15D
30
506
Transverse Horizontal Displacement &
P6
Rear
Bending Moment Pile P6
cap
Front
cap Transverse hor. disp., Ux (mm) Transverse moment, Mx (kNm)
-6 -5 -4 -3 -2 -1 0 -2000 -1500 -1000 -500 0 500
0 0
-5 -5
Depth (mbgl)
-10 -10
Depth (mbgl)
-15 -15
-20 -20
-25 -25
Front Rear +2D +15D Front Rear +2D +15D
Tunne
l
Greenfield Surface & Building Settlements
-5
-10
-15
Greenfield
-20
Building
-25 VL 2.8%
-30 Building
31
507
Building Settlements > Greenfield Surface Settlements
Greenfield
surface
settlements
Part 3.3
32
508
Tunnelling Near a Group of 48 Piles
0 mPD
Pile cap 1.5m
13.6m
10m -20mPD
1m
0.6m Ø Franki piles
3D Model by Plaxis 3D
Bldg.
load
“Plate” modelling
superstructure EI &
Building EA
40m
Fill 1m 48 Franki piles
(Embedded
CDG Tunnel Piles)
advanc
Tunne 120m
e
140ml
6m Ø
tunnel
• Analysis by Plaxis 3D Release Candidate 2
• 69,951 nos. of 15-node wedge/13-node pyramid/10-node tetrahedral elements
33
509
Effect of 3m Thick Annulus Grout
Grouted
annulus
21m
3m
Tunnel
Effect of Fixed Pile Head Connections
Pile cap
“Plate” modelling
superstructure
Tunnel
34
510
Output of Results
Tunnel Tunnel
advance advance
Greenfield Surface & Building Settlements
Ho rizo
Horizontal ntal distance
distance fro m centreline
from tunnel centreline(m)
(m)
-60 -40 -20 0 20 40 60
0
-2
Settlement (mm)
-4
-6
-8
Greenfield
-10
B uilding
Bldg.
-12
• For pile toes above tunnel, building settlements are greater than greenfield
surface settlements due to undermining below pile toes
35
511
Building Settlements > Greenfield Surface Settlements
Greenfield
surface
settlements
Comparison of Building Settlements
Ho rizo
Horizontal ntal distance
distance fro m centreline
from tunnel centreline(m)
(m)
0 5 10 15 20 25
0
Building settlement (mm)
-2
-4
Annulus grout
-6
Fixed pile heads A 1- B aseline
-8
A 2 - A nnulus gro ut
-10
A 3 - Fixed heads
-12 Baseline
(Pinned pile heads)
36
512
Transverse Horizontal Displ. of Closest Pile
-4
Level (mPD)
A 1- B aseline
(Pinned pile
heads) -8
A 2 - A nnulus
Gro ut
A 3 - Fixed -12
Heads Annulus grout
-16
-20
Tunnel
Part 3.4
37
513
Tunnelling Intersecting Piles
17.8m
Building
footprint
50 nos 0.6m Ø 8.8 Plan View
Franki piles @ 3Ø m
spacing
Tunnel 6m Ø
advanc tunnel
0 mPD e
Pile 1.5m
cap
Front View
3D Model by Plaxis 3D
Upper half
Full annulus
annulus grout
grout
Bldg.
load
“Plate” models
superstructure
Building
Fill 40m
CDG
Tunne Grout
120m 3m
150m l 12m
38
514
Output of Results
6m Ø tunnel
Exaggeration scale 100x
Greenfield Surface & Building Settlements
-10
-20
-30 Greenfield
Building
-40
Bldg.
39
515
Effect of 3m Thick Annulus Grout on Bldg. Settlements
-10
Baseline No grout
-30
Half grout
Full grout
-40
Effect of Annulus Grout on Pile Settle. & Axial Force
Critical pile
Tunnel 6m Ø
advanc tunnel
Pile settlement (mm) e Pile axial force (MN)
-40 -30 -20 -10 0 -1 0 1 2 3
0
Tensio
-4 n
Level (mPD)
-8
Baseline -12
(No grout)
Initial
-16
Half grout
Baseline (No grout)
-20 Half grout
Full grout
-24 Full grout
40
516
Effect of Annulus Grout w.r.t Pile N‐M Capacity
Critical pile
Tunnel 6m Ø
advanc tunnel
2500 e Pile capacity
Baseline (No grout)
2000 Half grout
Full grout
Axial force (kN)
1500
1000
500
-500
-200 -150 -100 -50 0 50 100 150 200
Moment (kNm)
Numerical Modeling
• Plaxis 2D
• Plaxis 3D Tunnel
Slide 82
41
517
Why Numerical
Plaxis 2D
Plaxis 3D Modeling?
Slide 83
Because!
Empirical Method based on Close Form
Solution
Suitable for Circular Tunnel Profile/Geometry
Homogeneous Ground Condition
Isotropic Stress Regime
Slide 84
42
518
Why Numerical
Modeling?
Numerical Method not based on Close Form
Solution
Suitable for Any Tunnel Profile/Geometry
Inhomogeneous Ground Condition (Multilayers)
Anisotropic Stress Regime
Slide 85
Plotting Ground
Reaction Curve by
using Plaxis 2D
Slide 86
43
519
GROUND PROFILE
63m
Slide 87
CURVE CONSTRUCTION SEQUENCE
Slide 88
44
520
RELAXATION FACTOR ( ‐ FACTOR)
Slide 89
TUNNEL DEFORMATION PROFILE
0, 41.02
Slide 90
45
521
SELECTION OF POINT TO PLOT CURVE
0, 41.02
Slide 91
PLOTTING CURVE
Slide 92
46
522
GROUND REACTION CURVE
Slide 93
GROUND REACTION CURVE
Slide 94
47
523
Plotting Longitudinal
Deformation Curve
by using Empirical
Method
Slide 95
LONGITUDINAL DEFORMATION PROFILE (LDP)
This profile can be used to
Establish a distance‐convergence relationship for 2D
modeling or for analytical solutions
The following Equations are proposed by
Vlachopoulos and Diederichs (2009a) to estimate
LDP
Slide 96
48
524
LONGITUDINAL DEFORMATION PROFILE (LDP)
Input Parameters
• Radius of Plastic Zone (Rp)
• Tunnel Radius (Rt)
• Total Tunnel Deformation (Umax)
• Distance Interval Ahead and Behind of
Tunnel face (X)
Slide 97
TUNNEL AND PLASTIC ZONE RADIUS
(Final Stage)
Slide 98
49
525
TOTAL TUNNEL DEFORMATION
Slide 99
LONGITUDINAL DEFORMATION PROFILE (LDP)
Slide
100
50
526
Plotting Support
Reaction Curve
Slide
101
SUPPORT REACTION CURVE
Slide
102
51
527
SUPPORT REACTION CURVE
SF Psm/peq>1.0
psm
equilibrium
peq
Slide
103
DESIGN SPREAD SHEET
Slide
104
52
528
RELAXATION FACTOR DETERMINATION
Slide
105
COMPARISON
Slide
106
53
529
Verification by
Plaxis 3D
Slide
107
3D TUNNEL MODEL MESH
Slide
108
54
530
TUNNEL DEFORMATION
Slide
109
TUNNEL DEFORMATION
Slide
110
55
531
TUNNEL DEFORMATION
Comparison
3D
2D Slide
111
TUNNEL RADIAL DEFORMATION
DEFORMED MESH
Slide
112
56
532
TUNNEL RADIAL DEFORMATION
Slide
113
TUNNEL RADIAL DEFORMATION
Slide
114
57
533
TUNNEL RADIAL DEFORMATION
Slide
115
TUNNEL DEFORMATION
RADIAL FACE
Slide
116
58
534
RADIAL DEFORMATION
Slide
117
FACE DEFORMATION
Slide
118
59
535
DEFORMED PROFILE
Slide
119
DEFORMED PROFILE
Slide
120
60
536
DEFORMATION PROFILE
Slide
121
DEFORMATION PROFILE
Slide
122
61
537
Tunnel Face
PLAN VIEW
Tunnel Drive
SECTIONAL VIEW
Slide
123
With Support
Slide
124
62
538
DEFORMED PROFILE
With Support
Slide
125
DEFORMED PROFILE
With Support and Excavation
Slide
126
63
539
TUNNEL DEFORMATION PROFILE
With Support and Excavation
PLAN VIEW
Slide
127
TUNNEL DEFORMATION PROFILE
With Support and Excavation
SIDE VIEW
Slide
128
64
540
COMPARISON
Slide
129
THANK YOU
Slide
130
65
541
Summary
1. Details & results for 3D modelling of tunnel advance near a piled building are given:
a. pressure-controlled boundary on tunnel face & along TBM
b. integrated response of piles & building to tunnelling in 3D
c. bldg. on end-bearing piles: bldg. settlement < greenfield surface settle.
d. bldg. on friction piles (toes above tunnel): bldg. settlement > greenfield surface settlement
b. considers layered ground, tunnel face advance, TBM support pressure, bldg. stiffness & combined piles-cap-bldg. behaviour
3. 3D analysis adds value to tunnel design & construction process, e.g. assessment of
requirement for protective measures
References
1. Atkinson, J. H. & Sallfors G. (1991). Experimental determination of soil properties. Proc. 10th ECSMFE, Florence, Vol.3, 915-956
2. Burland, J. B. (1995). Assessment of risk of damage to buildings due to tunnelling and excavation. 1st Int. Conf. on Earthquake Geotech. Engrg., IS Tokyo.
3. Geotechnical Control Office (GCO) (1985). Technical Note T4/85 - MTR Island Line: Effects of Construction on Adjacent Property. Civil Engrg. Services Dept.,
Hong Kong.
4. Hake, D. R. & Chau, I. P. W. (2008). Twin stacked tunnels - KDB200, Kowloon Southern Link, Hong Kong. Proc. 13rd Australian Tunnelling Conference, 445-
452.
5. Loganathan, N., Poulos, H. G. & Xu, K. J. (2001). Ground and pile-group responses due to tunnelling. Soils and Foundations, 41(1), 57-67.
6. Moller, S. (2006). Tunnel induced settlements and structural forces in linings. PhD thesis, University of Stuttgart.
7. Moller, S. & Vermeer, P. A. (2008). On numerical simulation of tunnel installation. Tunnelling & Underground Space Technology, 23, 461-475.
8. Ng, C. W. W., Sun, Y. F. & Lee, K. M. (1998). Laboratory measurements of small strain stiffness of granitic saprolites. Geotechnical Engineering, SEAGS,
29(2), 233-248.
9. Pang, C. H. (2006). The effects of tunnel construction on nearby pile foundation. PhD thesis, National University of Singapore.
10. Potts, D. M. & Addenbrooke, T. I. (1997). A structure’s influence on tunnelling-induced ground movements. Geotechnical Engineering, Proc. ICE, 125, 109-
125.
11. Schnaid, F., Ortigao, J. A. R., Mantaras, F. M., Cunha, R. P. & MacGregor, I. (2000). Analysis of self-boring pressuremeter (SBPM) and Marchetti dilatometer
(DMT) in granite saprolites. Canadian Geotechnical J., 37, 796-810.
12. Selementas, D. (2005). The response of full-scale piles and piled structures to tunnelling. PhD thesis, University of Cambridge.
13. Storry, R. B. & Stenning, A. S. (2001). Geotechnical design & contraction aspects of the Tsing Tsuen Tunnels – KCRC West Rail Phase; Contract DB320.
Proc. 14th SEAGC, Hong Kong, 443-448.
14. Storry, R. B., Stenning, A. S. & MacDonald, A. N. (2003). Geotechnical design and construction aspects of the Tsing Tsuen Tunnels – contract DB320 KCRC
West Rail Project. Proc. ITA World Tunnelling Congress, (Re)claiming the Underground Space, Saveur (ed.), 621-626.
15. Vermeer, P. A. & Brinkgreve, R. (1993). Plaxis Version 5 Manual. Rotterdam, a.a. Balkema edition.
66
542
Chennai 2007
Pieter A. Vermeer
Institute of Geotechnical Engineering, University of Stuttgart
Further reading : Möller (2006). Tunnel induced settlements and structural forces
in linings. PhD thesis, University of Stuttgart, Institute of
Geotechnical Engineering.
Partial Excavations
University of Stuttgart
67
543
Introduction: Conventional tunnelling on NATM IGS
Full excavation
Sequential excavation
d = round length
University of Stuttgart
(Kalotte)
(Ulme)
(Kern)
(Sohle)
University of Stuttgart
68
544
Introduction: top heading of excavation IGS
69
545
Introduction: Typical German railway tunnel IGS
Spritzbeton-
Spritzbeton-
Außenschale
Außenschale
1- oder 2-
1- oder 2-
lagige KDB mit
lagige KDB mit
Abschottung
Abschottung
Innenschale als
Innenschale als
WUB-KO
WUB-KO
University of Stuttgart
University of Stuttgart
70
546
Introduction: „undrained“ versus „drained“ type of tunnel IGS
Spritzbeton- Spritzbeton-
Außenschale Außenschale
1- bzw. 2-
1-lagige KDB
lagige KDB mit
Abschottung
Innenschale
Innenschale
Als WUB-KO
Ulmendränage
Sohldränage DN 200
University of Stuttgart
e.g. 50m
71
547
Introduction: Tunnel Boring Machine (TBM) without shield IGS
University of Stuttgart
University of Stuttgart
72
548
Introduction: The settlement trough IGS
Extension of
Settlement Trough
University of Stuttgart
University of Stuttgart
73
549
Introduction: Damage due to settlements IGS
University of Stuttgart
x z´ x´
z x
unit soil weight
z
z z
z K0 z
University of Stuttgart
74
550
FEM for NATM: On the use of the Mohr-Coulomb Model IGS
Most usually:
Anagnostou (1993):
... Drained conditions prevail if the coefficient of permeability is larger than
10-7 - 10-6 m/s and if the tunnelling speed is at the same time 2.5-25m
per day or less
University of Stuttgart
75
551
FEM for NATM: Intermediate support pressure IGS
0
0
University of Stuttgart
normalized lining
pressure, σ σ 0
σ
core V
stress
reduction
method
β
volume loss V
For shield tunneling, β is usually beyond 0.5. For NATM, β is usually below 0.5.
University of Stuttgart
76
552
FEM for NATM: Stress reduction method (-method) IGS
·0
Support pressure
77
553
Introduction into shield tunneling: Four types IGS
University of Stuttgart
University of Stuttgart
78
554
Introduction into shield tunneling: Causes of Settlements IGS
University of Stuttgart
University of Stuttgart
79
555
FEM for Shield-tunnelling: Grout pressure method IGS
In the grout pressure method the support pressure distribution is taken as a hydrostatical
grout pressure. No doubt this is better than the proportional reduction of the initial
stresses in the core. Ideally both the -method and the grout pressure method might be
extended with a displacement restriction of nodal point displacements towards the lining
University of Stuttgart
grout
University of Stuttgart
80
556
Case study: Heinenoord tunnel near Rotterdam IGS
University of Stuttgart
HS-small : h ≥ 1.5 D
HS-model : small strain stiffness starting at depth ± D/2 below tunnel
Mesh fineness around tunnel : element length in Table I < D/20
University of Stuttgart
81
557
Case study: Soil model parameters for MC-Model IGS
γsat E c´ φ´ K0
Layer
[kN/m³] [MPa] [kPa] [°] [-]
2 20 27 0 35 0.47
3 20 11 7 31 0.55
OCR = 1
University of Stuttgart
OCR = 1
University of Stuttgart
82
558
Case study: Site view of measuring IGS
University of Stuttgart
83
559
Case study: Horizontal displacements IGS
University of Stuttgart
University of Stuttgart
84
560
Case study: Computed bending moments IGS
- method
University of Stuttgart
Case study: Computed normal forces for Heinenoord shield tunnel IGS
University of Stuttgart
85
561
Case study: Measurement of structural forces using IGS
strain gauges
University of Stuttgart
M[kN/m]
M[kN/m]
University of Stuttgart
86
562
Modelling of Excavations in Plaxis
Dr William WL Cheang
Principal Geotechnical Consultant
Plaxis AsiaPac
Lecture notes are contributed by:
Dr Lee Siew Wei
Prof. Harry Tan
A.Prof. Ronald Brinkgreve
Dr Shen Rui Fu
Ir Dennis Waterman
Contents
1. Modelling of Excavations in PLAXIS
2. Influence of Soil Model
3. Validations on the use of PLAXIS
4. Influence of variation of water-table
563
AXI‐SYMMETRY, PLANE‐STRAIN & 3‐D MODEL
SECTION 1:
GEOMETRY‐ MODEL DISCRETIZATION
3-D MODEL
2-D Plane Strain
564
GEOMETRY‐ MODEL DISCRETIZATION
Axi-symmentry
565
3‐D MODELS
Piled building
Tower crane
Strut layout
Piled building
566
3‐D MODEL OF AN EXCAVATION
Top of PW (70/90)
N
Top of Grade III or Better
EXCAVATION ANALYSIS
A. Typical excavation sequences:
Berlin method:
• Soldiers installation
• Excavate (and dewater) one level
• Install wall panel
• Install (and prestress) anchor
567
WALLS – ONE CHAIN VS. MULTIPLE CHAINS
A. One chain
1. Continuous drawing of plate elements give one chain
2. One chain has one material set and is (de)activated in once piece
3. For walls that are fully installed prior to excavation
B. Mutiple chains
1. Draw wall piece → right click to end wall chain → draw next wall piece → etc.
2. Every piece can have its own material set and can be (de)activated individually
3. For walls that are installed following the excavation process
C. Sheet pile wall, slurry wall
1. One wall chain
D. Berlin wall
1. Generally the soldiers determine wall properties and not the panels → one wall chain
2. Multiple chains may be used if panels have significant contribution to the flexural rigidity of
the wall.
WALLS – THIN WALL VS. THICK WALL
A. Thin wall
1. Wall thickness << wall length
2. Shearing not important
3. No end‐bearing, only friction
→ Plate element suffices
A. Thick wall
1. Wall thickness significant
2. Shearing important
3. End‐bearing capacity needed
→ Use soil elements with material set represen ng wall material
→ In order to obtain structural forces a plate may be inserted
568
Walls – thick wall
A. Soil elements with material set representing wall material
B. Difficult to obtain structural forces from soil elements, therefore introduce plate:
1. No influence on deformation: low stiffness, no weight
2. Located in on the neutral line (usually the middle) d
3. Tight bonding to the concrete elements: no interfaces
Soil elements: Esoil=Ewall, I = 1/12*d3 , d = wall thickness
WALLS – SUPPORT
A. Lagging
1. Use short additional section of plate perpendicular to the
wall
2. Create short section with increased stiffness using multiple
chains
B. Struts
1. Full excavation: node‐to‐node anchor
2. Half (symmetric) excavation: fixed‐end anchor
C. Anchors
1. Grout anchor: node‐to‐node anchor + geotextile for grout
body
2. Ground anchor: node‐to‐node anchor + perpendicular plate
element
569
GROUND ANCHORS
A. Combination of node‐to‐node anchor and geogrid
B. Node‐to‐node anchor represents anchor rod
(no interaction with surrounding soil)
C. Geogrid represents grout body (full interaction with surrounding soil)
D. No interface around grout body; interface would create unrealistic failure surface
E. Working load conditions only – no pullout
F. If pullout force is known this can be used by limiting anchor rod force
Ground anchors
Axial forces in ground anchors:
Input geometry
Nrod <> Ngrout due to shared node between anchor, geotextile and soil
570
INTERFACES
A. Soil‐structure interaction
1. Wall friction
2. Slip and gapping between soil and structure
B. Soil material properties
1. Taken from soil using reduction factor Rinter
Cinter = Rinter * Csoil
tan(inter) = Rinter * tan(soil)
ψinter = 0 for Rinter < 1, else ψinter = ψsoil
Ginter = (Rinter)2 * Gsoil
2. Some building codes prescribe soil‐wall friction angle δ:
Rinter = tan(δ)/ tan(soil)
3. Individual material set for interface
INTERFACES
Suggestions for Rinter:
1. Interaction sand/steel = Rinter ≈ 0.6 – 0.7
2. Interaction clay/steel = Rinter ≈ 0.5
3. Interaction sand/concrete = Rinter ≈ 1.0 – 0.8
4. Interaction clay/concrete = Rinter ≈ 1.0 – 0.7
5. Interaction soil/geogrid = Rinter≈ 1.0
(interface may not be required)
6. Interaction soil/geotextile = Rinter≈ 0.9 – 0.5 (foil, textile)
Stability should not be critical on Rinter !
571
MATERIAL BEHAVIOUR
A. Unloading due to excavation
1. Vertical unloading at excavation bottom
2. Horizontal unloading behind wall
B. Primary loading due to pre‐stressing
C. HS‐small model is preferred
1. Non‐linear elastic unloading/reloading behaviour
2. Shear plasticity due to horizontal unloading
3. High far‐field stiffness for better settlement trough prediction
MATERIAL BEHAVIOUR: STRESS PATHS
Construction phases: K
active K0
• I 1st excavation σ
v K=1
• II Pre-stressing anchor
Point A
•III Final excavation III
II
I
Point A
K
Point B passive
Point B
σ
h
572
MATERIAL BEHAVIOUR: STRESS PATHS
Eur ,,E50
Eoed
E0
Eur , E50
Eur
E0
MATERIAL BEHAVIOUR
A. Mohr‐Coulomb: unrealistic deformations
1. Overestimation over bottom heave
2. Often heave of soil behind the wall
3. Occasionally excavation widens spontaneously (even without anchors!)
B. Hardening Soil model: qualitative realistic deformations
1. Better bottom heave, but increases with model depth
2. Settlement trough behind wall, but often too shallow and too wide
C. HS‐small model: qualitative and quantitative realistic deformations
1. Good bottom heave independent of model depth
2. More realistic settlement trough behind the wall (narrower and deeper)
573
DEWATERING: WET EXCAVATION
A. Excavate without changing water conditions (in stages or at once)
B. Apply stabilising weight at the bottom
C. Set excavated area dry
1. Use “cluster dry” option or
2. Use “cluster phreatic line”
D. Pore pressures outside excavated area remain unchanged
DEWATERING: DRY EXCAVATION
Undrained excavation
A. For every excavation phase do
1. Excavate soil
2. Set excavated area dry
Phreatic level outside the excavation remains unchanged
→ Suitable for short‐term excavations in low permeability soils
574
DEWATERING: DRY EXCAVATION
Drained excavation
A. For every excavation phase do
1. Excavate soil
2. Define boundary conditions (heads)
3. Perform groundwater flow analysis.
Phreatic level outside the excavation lowers
→ Suitable for long‐term excavations in high permeability soils
Dewatering: dry excavation
Drained excavation
575
Dewatering
1 2 3
General
Z-shape phreatic level gives wrong results:
General
General
• No equilibrium in horizontal water pressures:
• Local peak stresses
6 7
INFLUENCE OF A LINEAR ELASTIC PERFECT PLASTIC AND HARDENING
MODEL
SECTION 2
576
CONSTITUTIVE MODELS
1: Mohr Coulomb
2: Hardening Soil
3:Hardening Soil + Small Strain Overlay
SURFACE HEAVE IN INITIAL EXC./CANTILEVER WALL
3 m deep excavation with cantilever wall
20kPa
5m
3m
7m
• 3 analyses with Mohr Coulomb, Hardening Soil & Hardening Soil-Small models
using equivalent soil input parameters
• Compare ground movements, wall displacements & wall stability
577
SOIL INPUT PARAMETERS FOR 3 ANALYSES
Parameters for soil strength & initial stress state
Analyses Material c' ' Rinter
Model (or ur)
3
(kN/m ) (kPa) (Deg) [-] [-]
1 MC 20 5 35 0.3 0.426 0.67
2 HS 20 5 35 0.2 0.426 0.67
3 HSsmall 20 5 35 0.2 0.426 0.67
PREDICTED SURFACE SETTLEMENT BEHIND WALL
Distance behind wall (m)
0 5 10 15 20 25 30
0.006
0.004 Heave
0.002
Settlement (m)
0.000
-0.002
-0.004
-0.006 Settlement
MC
-0.008 HS
HSsmall
-0.010
578
PREDICTED HEAVE AT EXC. LEVEL IN COFFERDAM
Distance in front of wall (m)
-5 -4 -3 -2 -1 0 1 2 3
0.025
MC
HS
0.020
Wall HSsmall
0.015
Heave (m)
0.010
0.005
0.000
-0.005
PREDICTED WALL RESULTANT DISPLACEMENT
MC
Ux=6mm HS HSsmall
Ux=11mm Ux=10mm Ux: wall horizontal
displacement
579
PREDICTED STABILITY OF WALL
3 Sum-Msf = FOS FOS=2.8
2.5
MC Rotation mechanism
2 with FOS 2.8
1.5
Sum-Msf = FOS
3 FOS=2.8
2.5
2
HS
1.5
• “Phi-c' reduction” for predicting FOS
• FSP III sheetpile properties:
3 Sum-Msf = FOS FOS=2.8 EI=34440 kNm2/m; EA=3.92×106kN/m
2.5 Mp=369 kNm/m; Np=3575 kN/m
2 HSsmall
1.5
SUMMARY OF PREDICTIONS
Analyses Surface settlement Heave at Wall horizontal FOS for wall
behind wall excavation level displacement stability
MC Heave 4 mm Heave 20 mm 6 mm 2.8
(not OK)
HS Settle 9 mm Heave 11 mm 11 mm 2.8
HSsmall Settle 9 mm Heave 8 mm 10 mm 2.8
580
VARIATION OF SOIL STIFFNESS IN EXCAVATION
1. Soil stiffness is not constant and varies with
a. stress-level. Higher stress, higher stiffness
b. strain-level. Higher strain (or displacement), lower stiffness
c. stress-path (recent soil stress history).
d. Rotation of stress path, higher soil stiffness
2. During excavation, soil elements at different locations experience
different changes in stress, strain & stress-path direction
SOIL STRESS PATHS NEAR EXCAVATION
GCO No.1/90
581
SOIL STRESS PATHS NEAR EXCAVATION
20kPa 25 20kPa
Failure line
20
3m A K0
15 A
Exc. A
B 10
B
Exc.
t (kPa)
K0 20kPa
7m 5
B
5m
-5
-10
Failure line
A: unloading compression
B: unloading extension -15
0 10 20 30 40 50 60
s' (kPa)
STRESS PATH DEPENDENT SOIL STIFFNESS
Stress path rotation,
Shear modulus, 3G’ (MPa)
t
°
=0°
=180°
K0 °
=90°
s' °
582
STRESS PATH DEPENDENT CDG STIFFNESS
Stress-level Test series
Extension
Compress
Compression
Extension
=90°
WHY MC PREDICTS INCORRECT SURFACE HEAVE?
1. MC models a constant soil stiffness prior to failure – not realistic
2. In reality, stiffness of soil elements near excavation varies according to
a. stress-level
b. strain-level
c. direction of stress-path
3. Realistic prediction of wall deflections & ground settlements in all excavation
stages requires a constitutive model that considers above factors, e.g. HS &
HSsmall models
4. HS & HSsmall consider factors (1), (2) & (3) in determining the operational
soil stiffness (E), i.e. E is changing during excavation
583
INFLUENCE OF SMALL STRAINS AT FAR FIELD AREAS
MODELLING OF DEEP EXCAVATIONS
SECTION 2.1:EXAMPLES
584
585
586
587
588
589
MODELLING OF DEEP EXCAVATIONS
SECTION 3: VALIDATIONS
Plaxis vs. SAP2000
25m
56
590
Plaxis 3D Foundation SAP2000
Plaxis 3D Foundation
SAP2000
58
591
Validation 3 – Deformed Mesh
Plaxis 3D Foundation SAP2000
Validation 3 – DWall Deflection
592
Validation 3 – Strut Axial Force
61
Validation 3 – DWall Bending Moment
62
593
MODELLING OF DEEP EXCAVATIONS
SECTION 4: INFLUENCE OF WATER‐
TABLE VARIATION
Effects of GWT Changes
Lectures of Prof Harry Tan
National University of Singapore
594
CASES STUDIED
A. Case 1 ‐ Lower GWT in all Soils above and below Soft Clay
B. Case 2 ‐ Lower Phreatic Level in Soil below Soft Clay
C. Case 3 ‐ Lower Phreatic Level in Soil above Soft Clay
CASE 1 – LOWER GWT IN ALL SOILS ABOVE AND
BELOW SOFT CLAY
-35m
Stiff OA, MC UNDrained, E’=50 MPa,
nu’=0.3, c’=20 kPa, phi’=30 deg, k=0.001
m/day
595
ANALYSIS STAGES
Initial GWT GWT lower to -8m
RESULTS OF SETTLEMENT OF PT A
596
RESULTS OF EXCESS PP OF PT K
RESULTS OF SETTLEMENT PROFILES
30 yr 50yr
20 yr
10 yr
5 yr
1 yr
597
RESULTS OF EXCESS PORE PRESSURE PROFILES
50yr
30 yr
20 yr
10 yr
5 yr
1 yr
COMPARE PP AND EFFECTIVE STRESS CHANGES
EXC PP
Eff Stress
dissipating with
increasing with time
time
598
CASE 2 – LOWER GWT IN SOIL BELOW SOFT
CLAY
WATER CONDITIONS SETTINGS
Analysis Stages
For Soft Clay: Interpolate between
top and bottom soil Clusters
599
RESULTS OF SETTLEMENT CASE 2
RESULTS OF EXCESS PP CASE 2
600
RESULTS OF SETTLEMENT PROFILES
50yr
30 yr
20 yr
10 yr
5 yr
1 yr
RESULTS OF EXCESS PORE PRESSURE PROFILES
50yr
30 yr
20 yr
10 yr
5 yr
1 yr
601
COMPARE PP AND EFFECTIVE STRESS CHANGES
EXC PP
dissipating with Eff Stress
time increasing with time
CASE 3 – LOWER GWT IN SOIL ABOVE SOFT
CLAY
602
WATER CONDITIONS SETTINGS
Analysis Stages
For Soft Clay: Interpolate between
top and bottom soil Clusters
RESULTS OF SETTLEMENT CASE 3
603
RESULTS OF EXCESS PP CASE 3
RESULTS OF SETTLEMENT PROFILES
50yr
30 yr
20 yr
10 yr
5 yr
1 yr
604
RESULTS OF EXCESS PORE PRESSURE PROFILES
50yr
30 yr
20 yr
10 yr
5 yr
1 yr
COMPARE PP AND EFFECTIVE STRESS CHANGES
EXC PP
dissipating with
Eff Stress
time
increasing with time
605
CONCLUSIONS
A. GWT changes in soils of higher permeability above or below a thick soft clay layer
will induce excess pore pressures which will cause long‐term consolidation
settlements
B. The amount and rate of settlement can be computed correctly by using an
appropriate FEM code with suitable soft soil models that have the compression
properties of Cc, Cs and Pc values build into the model (eg HS or SS model in
Plaxis)
End
606
3D Excavation Supported By Struts
Three dimensional finite element modelling using
Plaxis 3D
William Cheang, Shen Rui Fu & Tan Siew Ann
11/20/2011
Three dimensional finite element modelling of a 50 x 10 excavation supported by two levels of struts.
Excavation depth is 8.5m and the toe of the retaining structure is located at 12 m below ground level. Struts
spaced at 10m centre‐to‐centre spacing.
607
INTRO
ODUCT
TION
A. Ge
eometry
y
1. The
e geometryy of the exe
ercise is sh
hown in Ta
able 1.
2. The
e excavatio
on is 50m in 10m in width and 8.5m deep.
i length, 1
3. The
e toe of the
e diaphragm
m wall is lo
ocated at 12m
1 below
w the grounnd.
4. The
e excavatio
on is done in three sttages and supported
d by two leevels of struts spaced
d
at 1
10m centre
e-to-centre distance a
along the le
ength.
each level there are 5 numberss of struts.
5. At e
6. The
e ground co
ondition is simple an
nd it is a sin
ngle layere
ed problem
m.
608
B. Dime
ensions
s
1. The
e size of th
he models and
a dimen
nsions of th
he excavation is indiccated in Ta
able 2.
2. The
e size of th
he finite ele
ement mod
del in Plax
xis 3D is se
et at 100 x 60 x 22m
m in length,
width and dep
pth respecttively.
Figure
F 2: M
Model size in
i Plaxis 3D
609
C. Flow
w of worrk
The
e flow of wo
ork consistt the follow
wing steps. In genera
al the stepss will be:
1. Settting up of the
t projectt informatio
on and model size in “Project S
Settings”.
2. In M
Mode-Soil the depth of the mod
del is set and
a the initial positionn of the wa
ater table iss
set..
3. In M
Mode-Structures the dimension
ns of the excavation
e n and excaavation lev
vels will be
e
con
nstructed. This
T will be
b followed
d by the creation of the retain ing wall and supporrt
sysstem.
4. In M
Mode-Mesh
h the mode
el is discre
etised. Refinement will
w be madee on imporrtant zoness
to e
enhance th
he analysis
s and resul ts.
Mode-Watter Levels the wate r condition
5. In M n at each stages oof excavatiion will be
e
seq
quentially lo
owered using the ‘Clu
uster Dry’ option.
6. In Mode-Stag
ged Cons
struction th
he excava
ation proc
cess and constructiion of the
e
reta
aining syste
em will be made. Fin
nally the ca
alculation will
w be madde.
7. Insp
pect the ca
alculated re
esults in O
Output and construct displacemeent chart versus
v time
e
usin
ng the Cha
art option.
ANAL
LYSIS
A. PR
ROJECT
T SETTINGS
Fiigure 3: Qu
uick Selectt option (Se
elect “Starrt a new project” for thhis exercis
se)
610
Figure 4: S
Setting the size
s of a Pla
axis 3D mo
odel via “Mo
odel Tab”. The
T point off origin is automaticallyy
set at the ccentre of the
e model (x-o
origin = 0, yy-origin = 0) when using
g the abovee stated valu
ues.
1. The
e size in x-y space is set using the “Conto
our” input box.
b
2. The
e dimensio
ons in x-y space
s are 100 x 60 m in length
h and widtth. The min
nimum and
d
maxximum vallues are chosen as such that the point of origin w
will be loca
ated at the
e
cen
ntre in x-p space.
s
3. The
e units and
d general quantities fo
or the anallysis are given in the above figu
ure.
4. The
e dimensio
ons in ‘z-direction is sset using the
t ‘Borehole’ input ooption in ‘M
Modify Soiil
Layyers’ box.
5. Oncce the inpu
ut is complete click ‘O
OK’.
611
B.MODE: SOIL
L
Plaxis 3D has 5 main mo
odes to deffine the prroject, that is Soil, Sttructures, Mesh,
M
Wate
er Levels and
a Stage
ed Constru
uction. Eve
ery mode has a deedicated ve
ertical
toolbar located between the explorrer boxes and
a drawin
ng area. Inn Soil mod
de we
define the soil layers for projects. In
n this curre
ent case th
here is onlly a single layer
of so
oil. Informa
ation of this soil laye
er and loc
cation of th
he water ttable is en
ntered
using
g a single borehole. To define
e use the borehole
b option
o to deefine the single
s
soil la
ayer, follow
w these ste
eps:
1. Selecct the Bore
ehole tool ffrom the ge
eometry toolbar.
at represents the soil.. It is suggested
2. Clickk at a location in the ccluster tha
to cliick on (x; y) = (-50; -30) but any
a other position
p woould be fin
ne as
well. This place
es a boreh
hole at loc
cation (x; y)
y = (-50; --30) and opens
o
the Modify
M soil layers win
ndow (Figure 5). The Modify soiil layers wiindow
provides a grap
phical reprresentation
n of the cu
urrent boreehole on th
he left
and a spreads
sheet for d
data inputt. However when crreating the
e first
hole in an
boreh a analyssis both the
t graph
hical repreesentation and
sprea
adsheet is empty.
Figure
e 5: Modify soil
s layers window
w
612
3. Presss Add button to defin
ne a new la
ayer in the borehole. By default both
the top
t and bo
ottom bou
undaries off the bore
ehole are set to z = 0.0.
Chan
nge the bo
ottom boun
ndary to -2
22m. This action in effect is setting
s
the depth
d of the
e finite elem
ment mode
el (see Figure 6)
4. To define
d the material properties
s for this layer of soil, press the
Mate
erial button. The Mate
erial sets window
w will appear
Fig
gure 6: Defin
ning a single
e layered soil using a single
s borehhole
Table
e 1: Material Properties
s for soil
Param
meter Symbol Soil nit
Un
Soil M
Model Name Mohr-Coulomb -
Drain
nage type Undrained (B) -
Unsaturated soill weight unsat 18.0 kN//m3
Saturrated soil we
eight sat 19.0 kN//m3
Young’s moduluss E-ref 15000 kN//m3
Poissson’s ratio Nu 0.25 -
Cohe
esion Su 30 kN//m2
Frictio
on angle 0 -
Interfa
faces Rigid -
Coeffficient of earth pressure
e at rest Ko Man
nual = 0.55 -
613
6. Creatte a new material set ffor the soil material ac
ccording to the data given in
Table
e 1. Any parrameters no
ot mentioned should be
e left at thei r default va
alue.
7. After entering the material p
properties for
f the soil close
c the w
window by clicking
OK in
n order to re
eturn to the Material se
ets window.
8. The material properties a d into the model byy ‘drag-and
are entered d-drop’
oach onto the boreho
appro ole layer (s
see Figure 7). Once this is don
ne the
desig
gnated soil layer in th
he borehole
e should be
e indicatedd with the colour
chose
en for this material
m set..
Figure
re 7: Boreho
ole with matterial properrties properrly in place.
9. By de
efault the water-table
w is set at 0 m. In this example thhe water ta
able is
ed at ground level and therefore no
locate n alteration
n is neededd. Therefore
e keep
the ground level and there
efore no alteration is needed.
n Th erefore kee
ep the
d at 0 m.
Head
We havve now fin
nished de
efining the subsoil and
a we w
will continu
ue to
Structure
es Mode to
o define the
e geometrry of the ex
xcavation.
C. MO
ODE: ST
TRUCTU
URES
Press tthe Structu
ures option
n on the m
mode toolb
bar to mov
ve to Strucctures mod
de. We willl
create the excavvation size by constru
ucting a vo
olume. This volume w
will be loca
ated at the
e
point o
of origin of the
t model..
614
To construct the diaphragm wall we will decompose this volume into surfaces. We
will change the surfaces located at the four sides of the volume to plate elements. Also
interface elements will be introduced by using the same approach. Finally we will
construct 3 excavation levels by creating 3 horizontal surfaces at 3 locations along the
z-axis.
The 5 individual struts located at a given level are constructed using Beam elements.
One single strut is drawn and set with the corresponding material properties. This will be
copied and replicated using the Array tool along the x-direction at 10 m centre-to-centre
distance. The waling system located along the perimeter of the retaining wall will be
constructed using Beam elements. The struts and walers located at the first level will be
grouped together to form ‘support system level 1’. This is to allow easy identification and
activation during the staged construction process.
Again by using the Array tool all the struts and waler system (support system level 1)
located at the first level will be copied to the second level and this therefore forms
‘support system level 2’. Create the following material sets and properties listed in Table
2, 3 and 4.
Table 3: Strut
Element type Beam Unit
Identification Name Strut -
Area A 1.225E-1 m2
Weight 78.5 kN/m3
Behaviour Linear -
Modulus of Elasticity E 210.0E6 kN/m2
615
Second moment
m are
ea I1=
=I2 1.800E-2
1 m4
Table 4: Waling
Element type Beam Unit
Identifica
ation Na
ame Waling -
Area A 8.682E-3
8 m2
Weight 78.5 kN/m33
Behaviou
ur Linear -
Modulus of Elasticity
y E 210.0E6
2 kN/m22
Second moment
m are
ea I1=
=I2 1.045E-2
1 m4
eate the ex
C1.Cre xcavation volume
Figure
e 8: Switch
h to Top Viiew (Plan vview) and movement
m limitation w
window ap
ppears
To expedite the
t constru
uction of th
he foot prin
nt of the excavation we switch
h to the top
p
view
w mode byy clicking the
t tool loccated on the menu. The viewppoint will be switched
d
to p
plan view. A ‘Movem
ment limitatiion’ window
w appears
s (see Figuure 8). Usin
ng this too
ol
it iss possible to
t fix the drawing
d pla
ane at a sp
pecific elev
vation. In tthis case th
he drawing
g
plan
ne is fixed at z = 0 m).
m This is to
o facilitate drawing of
o the excavvation foottprint
616
e volume is created by firstly d
The drawing a surface in x-y planee at z= 0 m and then
n
extrruding dow
wnwards by 12 m ussing the ex
xtrusion too
ol. From thhe vertical tool bar in
n
Stru
uctures Mo
ode select the Create
e surface option.
o
1. Construct a surface of 50 (L) a
and 10 (W)) m in size by locatinng the four points of a
rectangula
ar at the fo
ollowing c oordinates
s. A clockw
wise or annti-clockwis
se scheme
e
can be cho
osen (see Figure 9 a
and 10)
Table 5: Location
L of the surfa ce points at
a z=0m
Poiint 1 Poiint 2 Poin
nt 3 Poin
nt 4
x y x y x y x y
-25 -5 -25 5 25 5 25 -5
Figure 9
9: Construcct surface sttarting from
m point 1 to 3 ( and finallly 4) in clocckwise direc
ction
617
Figure 10: Constru
uction of surrface at z = 0 m (comp
pleted). The position off the points can
c still be
fine tun
ne using the
e table inputt seen here at the top right
r corner..
After inserrting the surface rep
presenting the cross section off the excavvation in plan
p we willl
extrude this surface to form the
e size (dep
pth) of the retaining structure
s (D
Diaphragm wall).
Figure
re 11: Extrussion along z-axis
z using
g a vector le
ength of 12 (vector z = -1)
618
5. Create a volume
v by filling an e
extrusion le
ength of 12
2m. To extrrude down
nwards that
is beneath
h the groun
nd surface the extrusion vector is z = -12 (see Figurre 11).
6. Click Applly to confirm
m the extru
usion process.
C2.Cre
eate the re
etaining sttructure w
with interfa
aces
1. Ensure tha
at the view
w is set as P
Perspectiv
ve
2. Select Vollume 1 by clicking th
he volume
e in the dra
awing areaa. Once se
elected the
e
volume is marked in red and in
n the Selec
ction explorer Volumee 1 is indic
cated.
3. Using the ‘Right Mo
ouse Butto
on (RMB)’’ gesture turn the vvolume into surfacess
using ‘Deccompose in
nto surface
es’ tool (se
ee Figure 12).
4. Six surface
es will be formed
f aro
ound the original volu
ume (in thiss case Volume 1) but
we will ne
eed on the
e 4 vertica
al surfaces
s to form the
t diaphrragm wall by turning
g
these 4 surfaces
s in
nto plates. One may
y choose to delete the top and
a bottom
m
surfaces to
o get a cle
ean model or mainta
ain these knowing thaat we can modify the
e
elevation of these 2 horizonttal surface
es into the
e excavatioon levels. The latter
option is straight
s forw
ward for exxperienced
d user.
5. In this exe
ercise we will
w use the
e first apprroach that is to deletee the top and
a bottom
m
surface. Select
S the to
op and botttom surfac
ce and delete them.
619
6. Select the
e 4 vertical surfaces located along the perimeter
p oof the volume. To do
o
this, click each of the
t vertica
al surfaces
s (while prressing thee ‘ctrl’ buttton on the
e
keyboard).
7. Perform the RMB gesture a
and select ‘Create plate’.
p Youu should obtain the
e
following schematic
c as show
wn in Fig
gure 13 showing
s thhe selecte
ed vertica
al
surfaces.
620
Figure 14: The plate elements
e rep
presenting the
t dimensiions of the D
D-wall.
621
Figure 15a Figure
e 15b Figure 15c
Figure
re 15: Locall axis on a plate
p (z axiss which is in
ndicated in blue, Fig.1 5a). A positive value iss
the ooutward direection. A neegative valuue is the innward direction. By seelecting ‘cre
eate positive e
interfa
face’ interface elementts will be crreated on thhe positive side (Fig.155b). Similarrly to create
e
an inwward interfa
ace the ‘crea
ate negative e interface’ option is ch
hosen (see Fig.15.c)
Figure
re 16: Com mplete interfface elemen nts along the
t perimetter of the pplate using the ‘create
e
positiive interfacces’ and ‘c create neg gative interffaces’ apprroach. Thee diagram shows the e
directtion of ‘local blue axis’ along the p
plate.
eate the ex
C3.Cre xcavation levels
To constru
uct the stag
ged excava
ation proce
ess 3 horiz
zontal surfa
aces locateed at (z = -3m),
- (z = -
6m) and (zz = -8.5m)) will be co
onstructed.. The first surface is drawn butt the subse
equent two
o
surfaces a t ‘Array’ tool.
are formed by using the
1. Construct a surface
e using the
e same po
oints as in
ndicated inn Table 5 but with z
et to -3 m.
position se
2. Change to
o ‘Top view
w’ and the ‘Movemen
nt limitation tool’ apppear again. Set the z
value to -3
3.0 m.
622
ate surface
3. Use ‘Crea e’ tool loca
ated in the
e vertical toolbar annd draw a horizonta
al
surface ussing the co
oordinates given in Ta
able 5. This is the firsst excavation level to
o
3 m below
w ground le
evel (see F igure 17)
4. To constru
uct the exc
cavation sta
age 2 the position off the seconnd surface will be at z
= -6m. The
e firs surfa
ace will be ccopied and
d place at z = -6m. T
The verticall difference
e
between th
he first and
d second ssurface is 3m.
3
5. Select the first surfac
ce (after se
election it will
w be marrked red).
6. Select the
e ‘Array tool’ from th
he vertical toolbar. The
T ‘Creatte array window’
w willl
appear as shown in Figure 18. By default the array
y pattern is ‘rectangullar’.
7. In the ‘Co
onfiguration’ section
n choose ‘1D,
‘ in z direction to copy th
he chosen
n
surface an
nd place th
he second horizontal surface at
a z = -6m
m. To perfo
orm this we
e
set the ‘Diistance between colu
umns’ as -3.
- The ‘ne
egative’ vallue is need
ded to shifft
the copied
d surface in the dirrection opp
posite to the
t global z-directio
on (upward
d
positive). A value off 3 is the d
difference in distance
e between the first and
a second
d
surface
8. The create
ed second surface re
epresenting
g the second excavaation level is shown in
n
Figure 19.
623
Figure 18: The ‘Create
e array’ tooll.
Figure 19
9: Horizonta
al surface 1 and 2 for sstaged exca
avation
624
Figure 20: The excava
ation levels represented by 3 horiz
zontal surfaaces located
d at
(z1= -3m, z2
z = -6m annd z3 = 8.5mm)
eate struts
C4.Cre s and wale
ers for sup
pport systtem at level 1(z = -1 .5 m)
The suppo m for the firrst excavattion stage is made of
ort system o a line oof ‘waling’ transecting
t g
along the perimeter of the wall at z1 = -1 .5m and 5 numbers of struts loocated at a centre-to-
centre spa
acing of 10
0m along x--direction.
1. The
e walers will be mode
elled using
g ‘beam’ ele
ements.
2. . Tw
wo approacches can be
b used to
o construct the walers
s using thee ‘beam’ ellements. In
n
the first, the beams
b can
n be drawn
n using the ‘beam’ too
ol found inn the vertic
cal tool barr.
The
e coordinattes in x-y plane
p is ag
gain the same in Table 5 but thee elevation
n (z) will be
e
-1.5
5m below ground.
g Th
his approa ch is named ‘Cad in
nput’ approoach. We will
w use the
e
seccond appro
oach whe
ere the ‘b
beam’ elem
ments rep
presenting the wale
ers will be
e
con
nstructed using the ‘C
Command line’ appro
oach.
3. The
e ‘Comman
nd line’ box n in Figure 21 and the
x is shown e coordinaates for the
e beam line
e
is sshown. In
n this exa
ample bra
ackets are
e used to
o indicate the five blocks of
coo
ordinates. The
T bracke
ets are nott necessary
y but used here for cclarity reaso
ons.
Figu
ure 21: A be
eam line is drawn using
g 5 points. The last point is similarr to the firstt point
4. The
e constructed waling
g system iss shown in
n Figure 22.
2 The maaterial for the waling
g
can
n be inserrted using the same
e ‘drag-and
d-drop’ ap
pproach orr by desig
gnating the
e
matterial set th
hrough the
e ‘Model exxplorer’ box
x (see Figu
ure 22).
625
Figu
ure 22: Inpu
ut of materia
al set ‘Walin
ng’ to ‘Beam
ms’ represen
nting first leevel of walin
ng.
Figure
e 23: First le
evel waling system
s with
h material set placed in
n.
5. The
e 5 nos. of strut will be
b construccted. The first
f strut will
w be placeed using tw
wo points.
6. The
e centre-to
o-centre sp
pacing is 1
10m and elevation
e of
o the struut is z= -1.5m below
w
ground level. The first point is lo
ocated at (x1,
( y1, z1) = (-20, --5, -1.5) and
a second
d
poin
nt located at (x2, y2, z3) = (-20 , 5, -1.5).
7. The
ere are two
o possible options. In
n the first option
o the strut can bbe constru
ucted using
g
the Cad inputt approach using the coordinate
es given above. Thiss is done by choosing
g
626
the ‘beam’ ele
ement option from th
he vertical toolbar and drawing the two points using
g
the coordinate
e set.
8. The second option
o gain using the ‘Comm
is ag mand line’ approach..
9. Forr the secon
nd approac
ch type “be
eam -20 -5 -1.5 -20 5 -1.5”.
10. Pla
aced the material
m set for the strut eith
her using the
t ‘Drag--and-drop’ or ‘Mode
el
exp
plorer inputt’ approach
h.
11. The
e complete
ed model at
a this with one single
e strut with
h the walingg system is shown in
n
Figu
ure 24
Figu
ure 24: Loca
ation of firstt level ‘walin
ng’ with one
e strut. Firstt excavationn level and interface
i are
e
swittched off.
12. The
e remaining 4 nos. of
o struts a
are replicatted using the ‘Array tool’. The
e important
parrameters when
w using
g this too l are settiing the sh
hape = 1D
D, x directtion, no of
o
umns = 5 and
colu a the dis
stance bettween the columns = 10 (note ‘positive’). The strutss
will be replica
ated along x direction
n. The outtcome of this processs is shown
n in Figure
e
25.
627
Figu
ure 25: The
e complete support
s sys tem for leve
el 1 (z=-1.5m
m)
C5.Cre
eat struts and walerrs for supp
port syste
em at level 2 (z = -4. 5 m)
628
ments loccated here
elem ein are grrouped tog
gether. We
W can renname this
s group to
o
‘Gro
oup_1_Sup
pport_Leve
el_1’ unde
er the ‘Mod
del explorerr’ box (seee Figure 28
8)
6. The
e second level su
upport sysstem is also
a ‘Grou
uped’ andd then re
ename ass
‘Gro
oup_2_Sup
pport_Leve
el_2’ unde
er ‘Model ex
xplorer’ bo
ox (see Figgure 28)
629
C. MO
ODE: MESH
Figure 29: M
Mesh option
n using the default ‘Ele
ement Distri
ribution’ optiion of ‘Veryy Course’ mesh
Mesh
h> Mesh Quality
Q
630
Figure 31:: Mesh quallity check
Click the ‘U
Update’ bu
utton locate
ed at the to
op of the window
w and
d return to ‘Mode: Me
esh’.
4. We
e will now proceed
p to ‘Mode: Wa
ater Levels
s” to design the channge of the water-table
w e
with
hin the exccavated zone.
D. MO
ODE: WATER
W LEVELS
L S
1. In ‘Mode: Wa
ater Levels
s’ the conffiguration of
o the initial groundw
water cond
dition is ass
succh that the ‘General Phreatic
P Su
urface’ is set
s right at ground levvel as seen
n in Fig.32.
Thiss is the inittial groundwater leve
el that was set at ‘Borrehole Inpuut via Head
d’.
Figure 32: M
Mode: Wate
er Levels. Visual
V in deffault Fig
gure 33: Distribution of hydrostatic
c porewater
settings pressuure
631
2. Clicck ‘Preview
w’ button lo
ocated at th
he ‘Vertica
al Toolbar’.. The initiaal porewate
er pressure
e
disttribution fo
or this sta
age (Initia l phase) is shown in Fig 333. This is the input
porrewater pre
essure thatt we want tto adopt fo
or this phas
se.
In summa
ary the su
urrounding soils and
d soil bene
eath zone 2 is refe rred to the ‘Genera
al
Phreatic S
Surface’. Zone 1 is set
s to ‘Dry’ and Zone
e 2 is set to
t ‘Interpollate’ betwe
een zone 1
eral Phreattic Surface’.
and ‘Gene
632
Phase 2: S
Support System
S 1 (Mode
( Wa ter Levels
s)
Phase 3: E
Excavatio
on Stage 2 (Mode W
Water Levels)
1. Thiss phase re
epresent ex
xcavation o
of zone 2 to
t 6m below
w ground llevel.
2. The
e ground water
w is low
wered by selecting zone
z 2 and setting tthe condition to ‘Dryy.
The
e procedurre is similarr to the me
ethod state
ed in Phase
e 1.
3. The
e porewate
er pressure
e within zon
ne3 that is beneath zone
z 2 is s et to ‘Interpolate’
Phase 4: S
Support System
S 2 (Mode
( Wa ter Levels
s)
1. Add
d phase 4 and name
ed this sta
age as ‘Support System 2’ Thiss phase will
w be used
d
late
er.
2. The
e porewate
er pressure
e is kept the same as
s in Phase 3.
Phase 5: E
Excavatio
on to Finall Level (Mo
odel Wate
er Levels)
633
Figure 35b:: Changing the water condition
c to ‘Dry’ for ex
xcvation zon
ne 1
Figure 36
6b: Select ‘In
nterpolate’ option
o for zzone 2 and a preview of the input h ydrostatic porewater
cond
dition for Phase 1
634
E. MODE: STAGED CONSTRUCTION
Initial Phase
By default ‘Initial Phase’ is already included. All the structural elements are switched off.
In this stage:
In this stage:
In this stage:
635
Proceed to
o ‘Calculatte’. Prior to
o calculatio
on we nee
ed to selec
ct two poinnts. These two pointss
are used to plot the ‘Displacem
‘ ment vs. tim
me’ curves.
Table 6: Co
oordinates
s for ‘A’ and
d ‘B’
Point x y z
A 0 5 0
B -24.5
5 5 0
Figure 37: S
Selection of point ‘A’ and ‘B’ for pllotting
636
F. OU
UTPUT OF
O RES
SULTS
Figure 38: P
Plot: Displa
acement usiing iso-surfa
aces
Figure 39: A
Axial forcess for struts at
a final exca
avation leve
el.
Figure 40: D
Deformation
n of retainin
ng structure
e
637
Figure 41: Wall displaccements at point A and
d B with tim
me.
638