PLAXIS implementation of hypoplasticity

David Masn
Plaxis bv, Delft
March 5, 2010
Contents
1 HYPOPLASTICITY 2
1.1 Hypoplastic model for granular materials . . . . . . . . . . . . . . . . 3
1.2 Hypoplastic model for clays . . . . . . . . . . . . . . . . . . . . . . . 6
1.3 Intergranular strain concept . . . . . . . . . . . . . . . . . . . . . . . 6
2 TIME INTEGRATION 9
3 PLAXIS INPUT 11
3.1 Hypoplastic model for granular materials . . . . . . . . . . . . . . . . 11
3.2 Hypoplastic model for clays . . . . . . . . . . . . . . . . . . . . . . . 13
4 TROUBLESHOOTING 18
5 EVALUATION 19
5.1 Drained triaxial test . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
5.2 Submerged construction of an excavation . . . . . . . . . . . . . . . . 19
5.3 Settlement due to tunnel construction . . . . . . . . . . . . . . . . . . 22
5.4 Building subjected to an earthquake . . . . . . . . . . . . . . . . . . 24
6 ACKNOWLEDGEMENT 26
1
Chapter 1
HYPOPLASTICITY
Hypoplasticity is a particular class of incrementally non-linear constitutive models,
developed specically to predict the behaviour of soils. The basic structure of the
hypoplastic models has been developed during 1990s at the University of Karlsruhe.
In hypoplasticity, unlike in elasto-plasticity, the strain rate is not decomposed into
elastic and plastic parts, and the models do not use explicitly the notions of the yield
surface and plastic potential surface. Still, the models are capable of predicting the
important features of the soil behaviour, such as the critical state, dependency of
the peak strength on soil density, non-linear behaviour in the small and large strain
range, dependency of the soil stiness on the loading direction, etc.
This is achieved by the hypoplastic equation being non-linear in the stretching tensor
D. The basic hypoplastic equation may be written as

T = L : D+ND, (1.1)
where

T is the objective (Jaumann) stress rate, D is the Eulers stretching tensor
and L and N are fourth- and second order constitutive tensors, respectively. The
early hypoplastic models were developed by trial and error, by choosing suitable
candidate functions from the most general form of isotropic tensor-valued functions
of two tensorial arguments (Kolymbas [6]). An important step forward in developing
the hypoplastic model was the implementation of the critical state concept. Gudehus
[3] proposed a modication of Equation (1.1) to include the inuence of the stress
level (barotropy) and the inuence of density (pyknotropy). The modied equation
reads

T = f
s
L : D+f
s
f
d
ND. (1.2)
Here f
s
and f
d
are scalar factors expressing the inuence of barotropy and py-
knotropy. The model by Gudehus [3] was later rened by von Wolersdor [11]
2
1.1. Hypoplastic model for granular materials Chapter 1. HYPOPLASTICITY
to incorporate Matsuoka-Nakai critical state stress condition. This model is nowa-
days considered as a standard hypoplastic model for granular materials, and this
version is also implemented in PLAXIS.
Later developments focused on the development of hypoplasticity for ne grained
soils. Herle and Kolymbas [5] modied the model by by von Wolersdor [11] to al-
low for lower friction angles and independent calibration of bulk and shear stinesses.
Based on this model, and the generalised hypoplasticity principle by Niemunis [9],
Masn [7] developed a model for clays characterised by a simple calibration procedure
and capability of correct predicting the very small strain behaviour (in combination
with the intergranular strain concept described later). This clay version is incor-
porated in PLAXIS.
The basic hypoplastic models characterised by Eq. (1.2) predict successfully the soil
behaviour in the medium to large strain range. However, in the small strain range
and upon cyclic loading they fail in predicting the high quasi-elastic soil stiness.
To overcome this problem, Niemunis and Herle [10] proposed an extension of the
hypoplastic equation by considering additional state variable intergranular strain
determine the direction of the previous loading. This modication, often denoted as
the intergranular strain concept, is implemented in PLAXIS and can be used with
both the model for granular materials and the model for clays.
The rate formulation of the enhanced model is given by

T = M: D (1.3)
where Mis the fourth-order tangent stiness tensor of the material. The total strain
can be thought of as the sum of a component related to the deformation of interface
layers at intergranular contacts, quantied by the intergranular strain tensor ; and
a component related to the rearrangement of the soil skeleton. For reverse loading
conditions and neutral loading conditions the observed overall strain is related only
to the deformation of the intergranular interface layer and the soil behaviour is
hypoelastic, whereas in continuous loading conditions the observed overall response
is also aected by particle rearrangement in the soil skeleton and the soil behaviour
is hypoplastic.
1.1 Hypoplastic model for granular materials
The hypoplastic model for granular materials has eight material parameters -
c
,
h
s
, n, e
d0
, e
c0
, e
i0
, and . Their calibration procedure has been detailed by
3
1.1. Hypoplastic model for granular materials Chapter 1. HYPOPLASTICITY
Figure 1.1: Inuence of n (a) and h
s
(b) on oedometric curves (Herle and Gudehus
[4]).
Herle and Gudehus [4]. A somewhat simplied calibration procedure is described
in the following. The critical state friction angle
c
can be obtained directly by
the measurement of the angle of repose. The next two parameters h
s
and n can be
directly computed from oedometric loading curves. The parameter n controls the
curvature of oedometric curve and h
s
controls the overall slope of oedometric curve
as is shown in Figs 1.1 (a) and (b). Having two states at the oedometric curve (Fig.
1.1), the parameter n can be calculated from
n =
ln(e
p1
C
c2
/e
p2
C
c1
)
ln(p
s2
/p
s1
)
(1.4)
where mean stresses p
s1
and p
s2
can be calculated from axial stresses using the Jaky
formula K
0
= 1 sin
c
, and e
p1
and e
p2
are the void ratios corresponding to the
stresses p
s1
and p
s2
. Tangent compression indices corresponding to the limit values of
the interval p
s1
and p
s2
(C
c1
and C
c2
) can be approximated by secant moduli between
loading steps preceding and following the steps p
s1
and p
s2
. The parameter h
s
can
then be obtained from
h
s
= 3p
s

ne
p
C
c

1/n
(1.5)
where C
c
is a secant compression index calculated from limit values of the calibration
interval p
s1
and p
s2
; p
s
and e
p
are averages of the limit values of p and e for this
interval.
The next three model parameters are the reference void ratios e
d0
, e
c0
and e
i0
, corre-
sponding to the densest, critical state and loosest particle packing at the zero mean
4
1.1. Hypoplastic model for granular materials Chapter 1. HYPOPLASTICITY
stress. The reference void ratios e
d
, e
c
and e
i
corresponding to the non-zero stress
depend on the mean stress by formula due to Bauer [1]:
e
c
= e
c0
exp

3p
h
s

(1.6)
The dependency of the reference void ratios on the mean stress is demonstrated in
Fig. 1.2.
Figure 1.2: The dependency of the reference void ratios e
d0
, e
c0
and e
i0
on the mean
stress (Herle and Gudehus [4]).
Following Herle and Gudehus [4], initial void ratio e
max
of a loose oedometric spec-
imen can be considered equal to the critical state void ratio at zero pressure e
c0
.
Void ratios e
d0
and e
i0
, which are the next two parameters, can approximately be
obtained from empirical relations. The physical meaning of e
d0
is void ratio at max-
imum density, void ratio e
i0
represents intercept of the isotropic normal compression
line with p = 0 axis. Void ratio e
i0
can be obtained by multiplication e
c0
by a fac-
tor 1.2. The ratio e
i0
/e
c0
1.2 was derived by Herle and Gudehus [4] considering
skeleton consisting of ideal spherical particles.
The minimum void ratio e
d0
should be obtained by densication of a granular material
by means of cyclic shearing with small amplitude under constant pressure. If such
a test is not available, it can be approximated using an empirical relation, with
e
d0
/e
c0
0.4.
The last two parameters and should be calibrated by means of single-element
simulations of the drained triaxial tests. The two parameters control independently
dierent aspects of soil behaviour, namely the parameter controls the shear stiness
and controls the peak friction angle.
5
1.2. Hypoplastic model for clays Chapter 1. HYPOPLASTICITY
1.2 Hypoplastic model for clays
The basic model [7] requires ve parameters (N,

,
c
and r). The parameters
have the same physical interpretation as parameters of the Modied Cam clay model,
and they are thus easy to calibrate based on standard laboratory experiments. The
model parameters N and

dene the position and the slope of the isotropic normal

compression line in the ln(1 + e) vs. ln p plane:
ln(1 +e) = N

ln
p
p
r
. (1.7)
where p
r
= 1 kPa is a reference stress; parameter

denes the slope of the isotropic

unloading line in the same plane. Their denition is demonstrated in Fig. 1.3
and their calibration using isotropic loading and unloading tests on reconstituted
London Clay specimens by Gasparre [2] is demonstrated in Fig. 1.4 [8]. The last two
Figure 1.3: Denition of parameters N,

and

.
parameters are the critical state friction angle
c
and the parameter r that controls
the shear stiness. Due to the non-linear nature of the model, the parameter r needs
to be calibrated by simulation of the laboratory experiments. With decreasing value
of r the shear stiness is increasing (Fig. 1.4b).
1.3 The intergranular strain concept (small strain
behaviour)
The intergranular strain concept requires ve additional parameters: R controlling
the size of the elastic range,
r
and controlling the rate of stiness degradation,
6
1.3. Intergranular strain concept Chapter 1. HYPOPLASTICITY
Figure 1.4: (a) Calibration of parameters N,

and

using isotropic tests on

reconstituted London clay, (b) calibration of the parameter r using undrained shear
test on reconstituted London clay (exp. data from Gasparre [2]).
m
R
controlling the initial shear stiness for the initial and reverse loading conditions
and m
T
controlling the stiness upon neutral loading conditions. When combined
with the hypoplastic model for clays, the initial very-small-strain shear stiness G
0
may be calculated from
G
0

m
R
r

p (1.8)
The parameters should be calibrated by simulating the laboratory experiments with
measurements of the small strain stiness using local strain transducers and with
measurements of the very-small-strain stiness using dynamic methods (such as ben-
der elements). The inuence of the parameter m
R
and
r
on the predicted small-
strain shear stiness curves is demonstrated in Fig. 1.5.
7
1.3. Intergranular strain concept Chapter 1. HYPOPLASTICITY
(a) (b)
Figure 1.5: The inuence of the parameters m
R
and
r
on the predicted small-strain-
stiness behaviour (experimental data from Gasparre [2]).
8
Chapter 2
TIME INTEGRATION
Constitutive models are integrated using explicit adaptive integration scheme with
local substepping. The constitutive model forms an ordinary dierential equation of
the form
dy
dt
= f(t, y)
The equation is for nite time step size t solved using the Runge-Kutta method.
Solutions that correspond to the second- and third- order accuracy of Taylor series
expansion are given by
y
(2)
(t+t)
= y
(t)
+k
2
y
(3)
(t+t)
= y
(t)
+
1
6
(k
1
+ 4k
2
+k
3
)
where
k
1
= t f

t, y
(t)

k
2
= t f

t +
t
2
, y
(t)
+
k
1
2

k
3
= t f

t + t, y
(t)
k
1
+ 2k
2

The accuracy of the solution is estimated following Fehlberg as the dierence between
the second- and third- order solutions. The time step size t is accepted, if
err =

y
(3)
(t+t)
y
(2)
(t+t)

< TOL
9
Chapter 2. TIME INTEGRATION
where TOL is a prescribed error tolerance. If the step-size t is accepted, y
(3)
(t+t)
is
considered as a solution for the given time step and the new time step size t
n
is
estimated according to Hull
t
n
= min

4t, 0.9t

TOL
err

1/3

If the step-size t is not accepted, the step is re-computed with new time step size
t
n
= max

t
4
, 0.9t

TOL
err

1/3

In the case the prescribed minimum time step size or the prescribed maximum num-
ber of time substeps is reached, the nite element program is asked to reject the
current step and to decrease the size of the global time step.
10
Chapter 3
INPUT OF PARAMETERS AND
STATE VARIABLES IN PLAXIS
3.1 Hypoplastic model for granular materials
Parameters are specied in the PLAXIS input in the following order:
Parameter 1 critical state friction angle
c
Parameter 2 p
t
shift of the mean stress due to cohesion. For the basic
hypoplastic model set p
t
= 0, but non-zero value of p
t
is needed to overcome
problems with stress-free state.
Parameters 3-9 parameters of the basic hypoplastic model for granular ma-
terials h
s
, n, e
d0
, e
c0
, e
i0
, , .
Parameters 10-14 the intergranular strain concept parameters (m
R
, m
T
, R,

r
, ). If m
R
= 0 the intergranular strain concept is switched o and the
problem is simulated using the basic hypoplastic model.
Parameter 15 not used.
Parameter 16 initial void ratio corresponding to the zero mean stress e
0
or
initial void ratio e. If Par(16) < 10, then e is calculated from the mean stress
p and from e
0
= Par(16) using Bauer [1] formula. If Par(16) > 10, then
e = Par(16) 10.
11
3.1. Hypoplastic model for granular materials Chapter 3. PLAXIS INPUT
Parameters 17-22 initial values of the intergranular strain tensor in Voigt
notation (
11
,
22
,
33
, 2
12
, 2
13
, 2
23
).
State variables:
The routine uses 14 state variables:
State v. 1-6 intergranular strain tensor in Voigt notation (
11
,
22
,
33
, 2
12
,
2
13
, 2
23
).
State v. 7 void ratio e.
State v. 8 not used.
State v. 9 Eective mean stress.
State v. 10 Number of evaluation of the constitutive model in one global
time step (for postprocessing only).
State v. 11 Mobilised friction angle
mob
in degrees (for postprocessing only).
State v. 12 Normalised length of the intergranular strain tensor (for
postprocessing only).
State v. 13 Suggested size of the rst time substep (for calculation controll).
State v. 14 free.
The hypoplastic model for granular materials is implemented via user dened sub-
routine usermod.dll. To use the model in PLAXIS, copy the le UsrMod.dll into the
PLAXIS installation directory. Then, select user-dened model from the Material
model combo box in the General tab sheet (Fig. 3.1). After selecting the user-dened
model, correct user-dened dynamic library (typically UsrMod.dll) needs to be se-
lected in the Available DLLs combo box under Parameters tab sheet. In the
Models in DLL combo box, model with ID 1 (Hypoplas. - sand) must be selected.
The parameters can then be input into the parameter table (Fig. 3.2).
Parameters of the sand hypoplastic model for dierent soils have been evaluated by
Herle and Gudehus [4]. They are given in Table 3.1. Parameters of the intergranular
strain concept for granular materials are in Tab. 3.2.
12
3.2. Hypoplastic model for clays Chapter 3. PLAXIS INPUT
Figure 3.1: Selecting user dened model in the Material model combo box.
3.2 Hypoplastic model for clays
Parameters are specied in the PLAXIS input in the following order:
Parameters:
Parameter 1 critical state friction angle
c
Parameter 2 p
t
shift of the mean stress due to cohesion. For the basic
hypoplastic model set p
t
= 0, but non-zero value of p
t
is needed to overcome
problems with stress-free state.
Parameters 3-6 parameters of the basic hypoplastic model for clays

,
N, r.
Parameters 7-9 free.
Parameters 10-14 the intergranular strain concept parameters (m
R
, m
T
, R,

r
, ). If m
R
= 0 the intergranular strain concept is switched o and the
problem is simulated using the basic hypoplastic model.
13
3.2. Hypoplastic model for clays Chapter 3. PLAXIS INPUT
Figure 3.2: Selecting sand hypoplasticity model in the Parameters tab sheet.
Parameter 15 bulk modulus of water K
w
for undrained analysis using the
penalty approach with user-dened value of K
w
. In drained analysis, consol-
idation analysis, and undrained analysis using PLAXIS option undrained K
w
should be set to 0.
Parameter 16 initial void ratio e or overconsolidation ratio OCR. If Par(16) <
10, then e = Par(16). If Par(16) > 10, then OCR = Par(16) 10.
Parameters 17-22 initial values of the intergranular strain tensor in Voigt
notation (
11
,
22
,
33
, 2
12
, 2
13
, 2
23
).
State variables:
The routine uses 14 state variables:
State v. 1-6 intergranular strain tensor in Voigt notation (
11
,
22
,
33
, 2
12
,
2
13
, 2
23
).
State v. 7 void ratio e.
14
3.2. Hypoplastic model for clays Chapter 3. PLAXIS INPUT

c
h
s
n e
d0
e
c0
e
i0

Hochstetten gravel 36

32 10
6
kPa 0.18 0.26 0.45 0.5 0.1 1.9
Hochstetten sand 33

1.5 10
6
kPa 0.28 0.55 0.95 1.05 0.25 1.5
Hostun sand 31

1.0 10
6
kPa 0.29 0.61 0.96 1.09 0.13 2
Karlsruhe sand 30

5.8 10
6
kPa 0.28 0.53 0.84 1 0.13 1
Lausitz sand 33

1.6 10
6
kPa 0.19 0.44 0.85 1 0.25 1
Toyoura sand 30

2.6 10
6
kPa 0.27 0.61 0.98 1.1 0.18 1.1
Table 3.1: Typical parameters of the hypoplastic model for granular materials (Herle
and Gudehus [4]
R m
R
m
T

r

Hochstetten sand 1.e-4 5.0 2.0 0.5 6
Table 3.2: Parameters of the intergranular strain concept for sandy soils (Niemunis
and Herle [10])
State v. 8 Excess pore pressure u for undrained analysis using user-dened
value of K
w
. In undrained analysis using PLAXIS option undrained this vari-
able is equal to 0 end excess pore pressure may be found in standard PLAXIS
menu.
State v. 9 Eective mean stress.
State v. 10 Number of evaluation of the constitutive model in one global
time step (for postprocessing only).
State v. 11 Mobilised friction angle
mob
in degrees (for postprocessing only).
State v. 12 Normalised length of the intergranular strain tensor (for
postprocessing only).
State v. 13 Suggested size of the rst time substep (for calculation controll).
State v. 14 free.
The hypoplastic model for granular materials is implemented via user dened sub-
routine usermod.dll. To use the model in PLAXIS, copy the le UsrMod.dll into the
PLAXIS installation directory. Then, select user-dened model from the Material
model combo box in the General tab sheet (Fig. 3.1).
15
3.2. Hypoplastic model for clays Chapter 3. PLAXIS INPUT
After selecting the user-dened model, correct user-dened dynamic library (typi-
cally UsrMod.dll) needs to be selected in the Available DLLs combo box under
Parameters tab sheet. In the Models in DLL combo box, model with ID 2
(Hypoplas. - clay) must be selected. The parameters can then be input into the
parameter table (Fig. 3.3).
Figure 3.3: Selecting clay hypoplasticity model in the Parameters tab sheet.
Parameters of the clay hypoplastic model for dierent soils have been evaluated by
Masn and co-workers. They are given in Table 3.3. Typical parameters of the
intergranular strain concept for ne-grained soils are in Tab. 3.4.
16
3.2. Hypoplastic model for clays Chapter 3. PLAXIS INPUT

N r
London clay 22.6

0.11 0.016 1.375 0.4

Brno clay 19.9

0.13 0.01 1.51 0.45

Fujinomori clay 34

0.045 0.011 0.887 1.3

Bothkennar clay 35

0.12 0.01 1.34 0.07

Pisa clay 21.9

0.14 0.01 1.56 0.3

Beaucaire clay 33

0.06 0.01 0.85 0.4

Kaolin 27.5

0.11 0.01 1.32 0.45

London clay. (data Gasparre) 21.9

0.1 0.02 1.26 0.5

Kaolin 27.5

0.07 0.01 0.92 0.67

Trmice clay 18.7

0.09 0.01 1.09 0.18

Table 3.3: Typical parameters of the hypoplastic model for clays
R m
R
m
T

r

London clay [7] 1.e-4 4.5 4.5 0.2 6
London clay (data Gasparre) 5.e-5 9 9 0.1 1
Brno clay (nat.) 1e-4 16.75 16.75 0.2 0.8
Table 3.4: Typical parameters of the intergranular strain concept for clays
17
Chapter 4
TROUBLESHOOTING
Although the implementation has been tested to be accurate and robust, the hy-
poplastic models are highly non-linear which may cause problems during solving
complex boundary value problems. When encountering problems, the following steps
may help to improve the overall performance:
1. In the Iterative procedure settings under Parameters Tab sheet, set Man-
ual settings and do not use the arc-length control.
2. Try to modify the iterative procedure by decreasing the Desired minimum
and Desired maximum number of iterations (for example, 3 and 5 respec-
tively).
3. When used in combination with structural elements using interface elements,
try to decrease the interface stiness in the Interface Tab sheet under Ma-
terial models window.
4. The hypoplastic models are undened in the tensile stress region, which can
cause integration problems in the vicinity of the free surface and in the case of
staged construction starting from the stress-free state. For this reason, articial
cohesion is introduced in the model implementation through the parameter p
t
.
The user can specify any p
t
value; in the case p
t
=0 kPa, however, the program
replaces it by a default value p
t
=10 kPa, which is sucient to overcome most
problems due to the zero stress state.
18
Chapter 5
EVALUATION
5.1 Drained triaxial test
PLAXIS implementation of the hypoplastic models has been evaluated by means
of simulations of a drained triaxial test. Fig. 5.1 shows the stress-strain curves of
the drained triaxial test, compared with exact results obtained by accurate time
integration of the model using a single-element program. The curves generated by
the PLAXIS implementation practically coincide with the exact solution.
5.2 Submerged construction of an excavation
In this case, submerged construction of an excavation close to the river is simulated.
The upper 20 m of the subsoil consist of soft soil layers, which are modelled as a
single homogeneous clay layer using the clay hypoplastic model. Underneath this
clay layer there is a stier sand layer, modelled using the sand hypoplasticity model.
As in this case the displacement eld is signicantly inuenced by the small-strain
behaviour, hypoplastic models with the intergranular strain concept are used.
Soil parameters of the sand layer correspond to the Hochstetten sand from Tab. 3.1;
the intergranular strain parameters used are in Tab. 3.2. The sand is in a medium
dense state with e
0
=0.7. The clay layer is simulated with London clay parameters
(Tab. 3.3 for parameters of the basic model; Tab 3.4. for the intergranular strain
parameters), with the initial value of OCR equal to 2.
Figure 5.2 shows the displacement eld and Fig. 5.3 shows the normalised length
19
5.2. Submerged construction of an excavation Chapter 5. EVALUATION
Figure 5.1: Stress-strain curves of the drained triaxial test for the sand and clay
hypoplastic models with the intergranular strain concept compared with the exact
solution.
of the intergranular strain tensor. The normalised length of the intergranular strain
tensor varies between 0, which indicates the soil being inside the elastic range, and 1,
corresponding to the state swept-out of the small-strain memory. In the case of the
normalised length of the intergranular strain being equal to 1 the soil behaviour is
governed by the basic hypoplastic model. Indeed, Fig. 5.3 shows that this is the case
of the submerged construction of an excavation, where the soil is below the bottom
of the excavation and behind the wall outside the small-strain-stiness range.
20
5.2. Submerged construction of an excavation Chapter 5. EVALUATION
Figure 5.2: Submerged construction of an excavation - total displacement.
Figure 5.3: Submerged construction of an excavation - normalised length of the
intergranular strain tensor.
21
5.3. Settlement due to tunnel construction Chapter 5. EVALUATION
5.3 Settlement due to tunnel construction
In the next example, a shield tunnel excavated partly in soft clay and partly in
medium dense sand is simulated, and its inuence on a pile foundation is evaluated.
The tunnel has a diameter of 5 m and is located at an average depth of 20 m. The soil
prole indicates four distinct layers: The upper 13 m consists of soft clay type soil.
Under the clay layer there is a 2 m thick sand layer, which is used as a foundation
layer for the piles. Below this sand layer, there is another clay layer followed by a
deep sand deposit.
As in the previous example, soil parameters of the sand layer correspond to the
Hochstetten sand from Tab. 3.1; the intergranular strain parameters used are in
Tab. 3.2. The sand is in a medium dense state with e
0
=0.7. The clay layer is
simulated with London clay parameters (Tab. 3.3 for parameters of the basic model;
Tab 3.4. for the intergranular strain parameters), with the initial value of OCR equal
to 2.
Figure 5.2 shows the displacement eld and Fig. 5.3 shows the normalised length
of the intergranular strain tensor. Again, the normalised length of the intergranular
strain tensor in Fig. 5.3 demonstrates how the small-strain stiness is activated in
dierent parts of the modelled geometry. The soil is in hypoplastic state in the
vicinity of the tunnel, above the tunnel and also next to the pile foundations, which
bear the building weight. Further from the tunnel, the value of the normalised length
of the intergranular strain tensor is low and at these places the soil remains elastic
with high shear and bulk stinesses.
22
5.3. Settlement due to tunnel construction Chapter 5. EVALUATION
Figure 5.4: Settlement due to tunnel construction - total displacement.
Figure 5.5: Settlement due to tunnel construction - normalised length of the inter-
granular strain tensor.
23
5.4. Building subjected to an earthquake Chapter 5. EVALUATION
5.4 Building subjected to an earthquake
This example demonstrates capabilities of a hypoplastic model in dynamic analy-
sis of the earthquake impact on existing infrastructure. A real accelerogram of an
earthquake recorded by USGS in 1989 is used for the analysis.
The building consists of 4 oors and a basement. It is 6 m wide and 25 m high. The
subsoil consists of a sand with water level reaching the surface. The soil behaviour
during the earthquake is considered as undrained. Two cases were simulated. In one
case, the soil is in a loose state (e
0
= e
c0
), in the second case the soil is in a dense
state (e
0
is close to e
d0
). Hypoplastic model parameters of the Hochstetten sand
from Tab. 3.1 and the intergranular strain parameters from Tab. 3.2 are adopted.
Overall displacements of the top of the building are shown in Fig. 5.6. The soil
response to the earthquake depends signicantly on the soil state. The loose soil
liquees after 3-4 s of the earthquake, leading to the failure. The analysis cannot
continue and fails. The displacements are much lower in the case of dense soil.
Although some displacements occur also in this case, the soil retains some bearing
capacity sucient to overcome the failure.
Figure 5.6: Displacement of the top of the building during the earthquake.
Figure 5.7 shows total displacements after approx. 4 s of earthquake for the loose
24
5.4. Building subjected to an earthquake Chapter 5. EVALUATION
soil case, whereas Figure 5.8 shows the displacements at the same time (and in the
same scale) for the dense soil. The gures indicate foundation failure for the loose
soil case and relatively low displacements for the dense soil case.
Figure 5.7: Total displacements after 4 s of earthquake - loose soil.
Figure 5.8: Total displacements after 4 s of earthquake - dense soil.
25
Chapter 6
ACKNOWLEDGEMENT
The RKF-23 time integration scheme was programmed and kindly provided by Prof.
C. Tamagnini and E. Sellari. The original umat - PLAXIS interface, which served as
a basis for the present developments, was programmed and kindly provided by Dr.
P.-A. von Wolersdor. The implementation was optimised during the stay of the
author at Plaxis bv in March 2010. Plaxis support is greatly appreciated.
26
REFERENCES
[1] E. Bauer. Calibration of a comprehensive constitutive equation for granular
materials. Soils and Foundations, 36(1):1326, 1996.
[2] A. Gasparre. Advanced laboratory characterisation of London Clay. PhD thesis,
University of London, Imperial College of Science, Technology and Medicine,
2005.
[3] G. Gudehus. A comprehensive constitutive equation for granular materials. Soils
and Foundations, 36(1):112, 1996.
[4] I. Herle and G. Gudehus. Determination of parameters of a hypoplastic constitu-
tive model from properties of grain assemblies. Mechanics of Cohesive-Frictional
Materials, 4:461486, 1999.
[5] I. Herle and D. Kolymbas. Hypoplasticity for soils with low friction angles.
Computers and Geotechnics, 31(5):365373, 2004.
[6] D. Kolymbas. Computer-aided design of constitutive laws. International Journal
for Numerical and Analytical Methods in Geomechanics, 15:593604, 1991.
[7] D. Masn. A hypoplastic constitutive model for clays. International Journal for
Numerical and Analytical Methods in Geomechanics, 29(4):311336, 2005.
[8] D. Masn. 3D modelling of a NATM tunnel in high K
0
clay using two dierent
constitutive models. Journal of Geotechnical and Geoenvironmental Engineering
ASCE, 135(9):13261335, 2009.
[9] A. Niemunis. Extended Hypoplastic Models for Soils. Habilitation thesis, Ruhr-
University, Bochum, 2002.
[10] A. Niemunis and I. Herle. Hypoplastic model for cohesionless soils with elastic
strain range. Mechanics of Cohesive-Frictional Materials, 2:279299, 1997.
27
REFERENCES REFERENCES
[11] P. A. von Wolersdor. A hypoplastic relation for granular materials with
a predened limit state surface. Mechanics of Cohesive-Frictional Materials,
1:251271, 1996.
28