Escolar Documentos
Profissional Documentos
Cultura Documentos
Helmut F. Schweiger
Computational Geotechnics Group, Institute for Soil Mechanics und Foundation Engineering Graz
University of Technology, Austria
ABSTRACT: The influence of different modelling assumptions on the results of numerical analyses of a
deep excavation problem is discussed. Based on a reference solution a comprehensive parametric study is per-
formed, identifying modelling assumptions which may have a significant influence on the calculated dis-
placement behaviour and the bending moments in the wall. The parameters investigated include wall friction,
domain chosen for the analysis, constitutive models and modelling of the grout body. In a second example the
influence of the design approaches defined in Eurocode7 for ULS-design are investigated in connection with
numerical methods. It can be concluded from this study that care must be taken when setting up a numerical
model because the sum of various assumptions, not considered being of large importance when looked at it
individually, may significantly influence the outcome of the numerical calculation.
15
are introduced in order to increase this effect and to
take into account the high stiffness at low strains,
which will be prevailing in most of the deeper layers
of the domain analysed, at least in a very approxi-
mate way. Rinter in Table 1 determines the reduction
of strength parameters ϕ and c in the interface ele-
ments as compared to the surrounding soil (tanϕinter
= Rinter tanϕ, cinter = Rinter c). The stiffness of the in-
terface is reduced as well. A value of 1 kPa is intro-
duced for the cohesion which improves numerical
stability, this is however not strictly required.
16
2 RESULTS FOR REFERENCE SOLUTION the deformed mesh is shown and in Figure 6 the sur-
face settlements are plotted for the first and final ex-
cavation stage. Settlements increase from approxi-
mately 5 mm for the first stage to over 15 mm for
the final stage, which can be considered to be a very
plausible result. Figure 4 depicts the lateral dis-
placement of the wall together with the inclinometer
measurements, again for the first and final excava-
tion step. The measurements for the final stage have
been corrected for lateral movement of the base of
the wall which is not reflected in the inclinometer
measurement but most likely to occur (based on
measurements under similar conditions). Figure 5
shows calculated bending moments.
Deformed Mesh
Extreme total displacement 46.55*10-3 m
(displacements scaled up 100.00 times)
2
4
Fig 3. Deformed mesh (detail) – reference solution
6
-50 -45 -40 -35 -30 -25 -20 -15 -10 -5 0
0 8
2 10
4 12
8
16
10
18
12
depth below surface [m]
20
14
22
16 final stage
1. excavation stage 24
18
26
20
28
22
30
24
measurement 32
(final stage) -1000 -800 -600 -400 -200 0 200 400 600
measurement corrected 26
reference solution bending moments [kNm/m]
(final stage)
measurement
28 Fig 5. Bending moments – reference solution
(1. excavation stage)
reference solution 30
(1. excavation stage) distance from wall [m]
32 0 10 20 30 40 50 60 70 80 90 100
10
-50 -45 -40 -35 -30 -25 -20 -15 -10 -5 0
vertical displacement of surface [mm]
5
horizontal displacement [mm]
0
Fig 4. Wall deflection – reference solution -5
-10
-15
In the following the most relevant results obtained
-20
for the reference solution are presented. Unlike oth- final stage
-25
erwise stated the last construction stage is consid- 1. excavation stage
-30
ered. In addition a few results for the first excavation
step (no anchors installed) are shown. In Figure 3 Fig 6. Surface settlements – reference solution
17
3 INFLUENCE OF VARIOUS MODELLING
-70 -65 -60 -55 -50 -45 -40 -35 -30 -25 -20 -15 -10 -5 0
ASSUMPTIONS 0
2
In this section modelling assumptions such as the
dimensions of the domain analysed, modelling of 4
10
lows from Figures 7 to 9 that this parameter has a (1. excavation stage)
Rinter = 0.8 t_virt = 0.01 30
significant influence on the displacements. The hori- (1. excavation stage)
zontal displacement of the top of the wall increases -70 -65 -60 -55 -50 -45 -40 -35 -30 -25 -20 -15 -10 -5 0
32
18
distance from wall [m]
0 10 20 30 40 50 60 70 80 90 100 20
10
vertical displacement of surface [mm]
5 22
0
reference solution 24
-5
Rinter = 0.5
-10 26
-15
28
-20
-25 30
-30
reference solution
-35 Rinter = 0.5 32
-1000 -800 -600 -400 -200 0 200 400 600
-40
bending moments [kNm/m]
Fig 7. Surface settlements – influence of wall friction Fig 9. Bending moments – influence of wall friction
18
-1000 -800 -600 -400 -200 0 200 400 600
3.2 Influence of domain analysed 0
2
For the reference solution the domain analysed was
chosen as 150 x 100 m for width (W) and depth (D) 4
of the mesh respectively. In order to study the influ-
ence of the discretized domain chosen the following 6
8
0
1. excavation stage
10
-5
12
depth below surface [m]
-10
-30
20
Fig 12. Surface settlements – influenced of domain analysed
reference solution 22
D=100 W=150
D=100 W=100
D=70 W=100 24 3.3 Influence of modelling ground anchors
D=70 W=150
D=150 W=200
reference solution
26
When using the code Plaxis the load transfer from
D=100 W=150
D=100 W=100 28 the free length of the ground anchors into the ground
D=70 W=100 can be conveniently modelled with membrane ele-
D=70 W=150
D=150 W=200
30 ments. These elements, which have no bending stiff-
32
ness but axial stiffness only, allow a continuous load
-50 -45 -40 -35 -30 -25 -20 -15 -10 -5 0 transfer from the membrane element to the ground
horizontal displacement [mm] along its entire length and avoid a concentrated point
load at the end of the free anchor length. Of course
Fig 10. Wall deflection – influenced of domain analysed this modelling technique is only applicable for
working load conditions because the limiting pull
19
out force cannot be taken into account correctly with To emphasize the importance of a continuous load
this simple model. transfer along the grout body two analyses without
-140 -120 -100 -80 -60 -40 -20 0
membrane elements have been performed. In the
0 first one the free anchor length has been kept the
2
same as in the reference solution and in the second
one the free anchor length has been increased by
4 half of the length of the grout body in order to com-
6
pensate for not modelling the load transfer in more
detail. Figures 13 to 15 clearly show that care must
8
be taken when choosing the model representing the
10 ground anchor and grout body.
12
18 -10
-20
20
-30
22
-40
20
-60 -55 -50 -45 -40 -35 -30 -25 -20 -15 -10 -5 0
distance from wall [m]
0
0 10 20 30 40 50 60 70 80 90 100
30
2
4 20
15
6 10
5
8 0
-5
10 -10
-15 reference solution
12 MC_3 (E-loading)
16
Fig 18. Surface settlements - MC-models 3 and 4
18
4.1 Parameter variation with MC-model 1
20
In this section a limited parametric study, similar to
22 the one presented in section 3, is performed with the
parameter set of MC_1 (Table 3) as basic analysis. It
24
is interesting to see that with certain, however not
26 very realistic, assumptions the Mohr-Coulomb
model calculates a similar lateral deflection of the
28 wall as the Hardening Soil model (Figure 19). A
reference solution
MC_3 (E-loading) 30
match of surface settlements however cannot be
MC_4 (E-unloading) achieved (Figure 21). Bending moments are not in-
32 fluenced so much (Figure 20). Again the strong in-
-60 -55 -50 -45 -40 -35 -30 -25 -20 -15 -10 -5 0
fluence on the results of the assumptions made for
horizontal displacement [mm]
Fig 16. Wall deflection – MC-models 3 and 4 wall friction is obvious.
-60 -55 -50 -45 -40 -35 -30 -25 -20 -15 -10 -5 0
0
-1000 -800 -600 -400 -200 0 200 400 600
0
2
2
4
4
6
reference solution 6
MC_3 (E-loading) 8
MC_4 (E-unloading)
8
10
10
12
depth below surface [m]
12
depth below surface [m]
14
14
16
16
18
18
20
20
22
22
24
24
MC_1 26
Poisson = 0.2
26 Poisson = 0.4
Rinter = 0.5 28
28 reference solution
(Hardening Soil) 30
30
32
-60 -55 -50 -45 -40 -35 -30 -25 -20 -15 -10 -5 0
32
-1000 -800 -600 -400 -200 0 200 400 600 horizontal displacement [mm]
Fig 19. Wall deflection – MC-variations
bending moments [kNm/m]
Fig 17. Bending moments - MC-models 3 and 4
21
5 PROBLEM DEFINITION – EXAMPLE 2
-1000 -800 -600 -400 -200 0 200 400 600
0
2
The second, simpler example briefly addresses the
influence of the design approach according to Euro-
MC_1
Poisson = 0.2
4 code7 when performing ULS-design with finite ele-
Poisson = 0.4 ments. Eurocode7 allows for three different design
6
Rinter = 0.5
reference solution approaches DA1 to DA3 which differ in the applica-
(Hardening Soil)
8 tion of the partial factors of safety on actions, soil
properties and resistances.
10
EC7 states: "It is to be verified that a limit state of
12 rupture or excessive deformation will not occur with
25
20
15 5.1 Geometry, parameters and computational steps
10
5 The geometry of the problem follows from Figure
0
22. The following construction steps have been con-
-5
-10 MC_1
sidered in the analysis:
-15 Poisson = 0.2 - initial phase (K0 = 0.5)
Poisson = 0.4
-20
Rinter = 0.5 - activation of diaphragm wall (wished-in-place)
-25
-30
reference solution
(Hardening Soil)
- activation of surcharge loads
- excavation step 1 to level -2.0 m
Fig 21. Surface settlements – MC-variations
- activation of strut at level -1.50 m, excavation step
2 to level -4.0 m,
- groundwater lowering inside excavation to level -
6.0 m
- excavation step 3 to level -6.0 m
22
was shortened in 0.5 m intervals and for each wall
length a new analysis was performed. Horizontal
displacements of the base of the wall, bending mo-
ments, strut forces and the factor of safety, obtained
by means of a strength reduction technique, were
evaluated. Figure 23 shows the increase of horizon-
tal deformation of the base of the wall when decreas-
ing the length of the wall for DA2, DA3 and for an
analysis with characteristic parameters. The charac-
teristic analysis and DA2 is almost the same because
the only difference is the factor of 1.11 for the vari-
able load in DA2. Numerical convergence could not
be achieved for an embedment depth of 1.5 m for
DA2 and characteristic parameters and a depth of
3.0 m for DA3 respectively.
Fig 22. Geometry and excavation stages for example 2
40.0
no convergence
The surcharge of 10 kPa is a permanent load, the 35.0
E = 30 MPa 10.0
ν = 0.3 characteristic
5.0 DA2
ϕ = 27.5° DA3
c = 10 kPa 0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
γsaturated = 20 kN/m3 Embedment depth [m]
γabove water table = 19 kN/m3 Fig 23. Horizontal wall displacement vs. embedment depth
Diaphragm wall:
E = 3.0E7 kN/m2 Figure 24 plots the safety factor obtained from a
ν = 0.18 strength reduction technique for different embed-
γ = 24 kN/m3 ment depths. Again values for DA2 and characteris-
d = 0.8m tic parameters do not vary much and values for DA3
are much smaller because the soil strength is already
factored at the beginning of the analysis. Figure 24
Strut:
features an additional line, namely the value ob-
EA = 1.5E6 kN/m tained for DA2 divided by the partial factor which
has been put on the strength parameters in DA3.
They compare well for embedment depths between
3.5 and 4.5 m but for factors around or below 1 they
6 RESULTS – EXAMPLE 2 differ. To some extent this can be attributed to de-
tails of the iteration procedure and convergence set-
The following results were evaluated applying DA2 tings which become more sensitive for states at or
and DA3 as described above: required embedment very near to failure. No attempt has been made to
depth, design strut force, design bending moment. It achieve a closer matching by tightening tolerance
is acknowledged that in general the required em- factors because this was not the main goal of this in-
bedment would not be determined be means of finite vestigation.
element analyses but by employing a more conven-
tional approach. However, it is done here for high-
lighting differences in design approaches.
In order to determine the embedment depth
analyses were performed with different wall lengths.
Starting with an embedment depth of 5 m the wall
23
1.6
minimum embedment depth of approximately 3.5 m
characteristic
is required for a factor of safety > 1.0 (Figure 24).
1.5
DA2 However this result is considered to be quite accept-
DA3
1.4 DA2 / 1.25
able given the various modelling assumptions in-
volved in the two approaches.
Safety factor [-]
1.3
Discussion on the merits of numerical analyses in
1.2
combination with Eurocode7 design approaches can
also be found in Schweiger (2005), Bauduin, De Vos
1.1
& Frank (2003) and Simpson (2000).
1.0
400.0
0.9
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
350.0
characteristic
Embedment depth [m] DA2
300.0
Fig 24. Factor of safety vs. embedment depth DA3
200.0
400.0
150.0
350.0
characteristic 100.0
Bending moments [kNm/m]
DA2
300.0
DA3 50.0
DA2*1.35
250.0
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
200.0
Embedment depth [m]
150.0
Fig 26. Strut forces vs. embedment depth
100.0
50.0
7 SUMMARY AND CONCLUSION
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
24
on the relative stiffness between support system
(wall and struts/anchors) and soil.
Finally it is emphasized that numerical methods
provide a valuable tool in assessing serviceability
limit states (SLS) because it is possible to model soil
structure interaction adequately. However, care must
be taken in setting up the model because a number
of assumptions which have to be made by the mod-
eller will not be found in any code of practice or
geotechnical report. As far as Ultimate Limit State
(ULS) design is concerned, numerical methods can
be used but more experience is still required with re-
spect to model soil structure interaction in the ulti-
mate limit state in order to guarantee a consistent
level of safety factor in structural elements and soil.
REFERENCES
25
26