712
Proceedings of the 18
th
International Conference on Soil Mechanics and Geotechnical Engineering, Paris 2013
NEN 9997-1 (2011) or equivalent. The parameter R
inter
is the
ratio tan(
’)/tan(
’). Smear of bentonite should be considered.
The advantages of this Method 1 are: the bending moments,
shear and normal forces and deformations can directly be read
from the beam, structural connections to floors and struts can
easily be defined and the method is suitable for strength
analyses of the wall.
Disadvantages are: possible numerical problems caused by
the mesh in the area around the tip of the beam and fixed-end-
anchor and the unrealistic stress distribution below the tip of the
beam element. There is a work around for the first disadvantage
by extending the interface into underlying strata. The second is
important where group effects are significant. Here application
of Method 2 may be considered or a crossbeam could be
introduced at the beam tip. The vertical spring stiffness of the
beam/crossbeam should again be fitted to NEN 9997-1 (2011).
Application of Method 2 comprises linear elastic or Mohr-
Coulomb volume elements. The elements are modelled with
realistic dimensions (thickness and height). The required input
consists of parameters such as
(kN/m
3
), E
uncracked/cracked
(MPa,
like Method 1) and R
inter
(-). When using Mohr Coulomb,
additional strength parameters as c’ and φ’ are required.
The advantages of Method 2 are: better visualisation of
behaviour, proper calculation of stresses and deformations in
the soil, more stable numerical calculation process (especially
where walls have a bearing function) and a more realistic
vertical deformation behaviour at the tip (especially when
interaction with the environment is considered at tip level) and
of the wall itself (especially when the thickness is not constant).
Disadvantages are: load-settlement behaviour not in
accordance with NEN 9997-1 (2011), bending moments and
forces can not easily be extracted from the volume element and
structural connections to the diaphragm walls are difficult to
model. When using Mohr Coulomb for mixed or injected walls,
information of soil strength and stiffness is required for the
determination of strength and stiffness of the D-wall by using
empirical relations (Van der Stoel, 2001). Concerning the first
disadvantage, the spring stiffness can be fitted to standard load-
settlement curves by introducing a thin dummy volume element
below the diaphragm wall. A work around for the second
disadvantage is modelling a beam inside the linear elastic
volume element. This beam should not contribute to the
strength and stiffness of the diaphragm wall. Where struts are
required, or other structural connections, a dummy plate may be
introduced to the model having EI ≈ 0 kNm
2
.
It should be noted that installation effects and uncertainties
at the soil-wall interface (smear) make bearing capacity and
vertical stiffness hard to predict. It is common practice to apply
the design approach of bored piles to situations where cast in-
situ concrete walls are considered.
3 CASE INTRODUCTION
3.1
Railway tunnel Delft
The Delft railway tunnel project comprises the design and
construction of a 2.4 km long, four track double railway tunnel
in the historical city centre. The excavation level is
approximately 10 m below ground surface. Nearby buildings
are supported by shallow foundations at very close distances
from excavations. Therefore, a top-down multi-propped
construction sequence, using diaphragm walls was adopted.
Construction of the diaphragm walls near critical buildings
require additional measures to limit deformations of the
diaphragm walls in order to meet the criteria for angular
distortion and horizontal strain of buildings along the tunnel
alignment. The deformations of foundations of contiguities are
an accumulation of deformations, as follows:
1. Earthworks for underground infrastructure (pipes and
cables) in the narrow area between the buildings and the
diaphragm wall. At some places the distance is less then
4.0 m and the excavation depths over 2.5 m.
2. Removing obstacles of the historic town defense walls at
the proposed route of the diaphragm walls (excluded from
the analyses, impact is negligible).
3. Trench deformations during excavation with the ground
supported by bentonite mud or similar. Once the
reinforcement cage has been lowered into place, concrete
is tremmied into the slot, displacing the mud.
4. Deformations as a result of staged excavation of the
strutted tunnel trench.
Finite element models (Plaxis 2D and Plaxis 3D) were used
to assess the deformations of the tunnel system. The diaphragm
walls have typical thicknesses of 1.0 m and have standard
widths of 7.3 m. Standard excavation stages consider two strut
levels; the first just below surface level and the second at 50%
of the final excavation level. The model does not take account
of interaction of soil and foundation slabs. It assesses green
field deformations outside the tunnel trench. The deformations
at foundation level can be extracted from the model.
The design approach outlined below was adopted for the
prediction of deformations:
1. The ground deformations are assessed (SLS) as a result of
the construction of the diaphragm walls for panel widths of
3.8 m and 7.3 m (Plaxis 3D)
2. The required dimensions of the diaphragm wall are
determined with an elastic beam model using bi-linear
ground springs in (ULS and SLS) in combination with
structural analyses (ESA PT).
3. The ground deformations are assessed (SLS) as a result of
cable and pipe trenching.
4. The ground deformations are assessed (SLS) as a result of
the tunnel trench excavation taking account of detailed
construction stages (Plaxis 2D). This model continues from
step 3 and uses the input from step 2.
5. Finally the results of step 1 and 4 are combined. Where the
deformation requirements were not met additional
measures have to be taken, as described below.
Additional efforts to meet the deformation criteria of
buildings focus on further limiting the deformations of the
diaphragm walls by:
Excavation in stages, were the groundwater in the building
pit also is lowered in stages.
The panel width can be reduced to 3.8 m.
The struts could be pre-stressed to reduce elastic shortening
of the steel cross section and to pre-stress the ground at the
active side of the retraining walls.
3.2
Drents Museum Assen
The Drents Museum is located on a historical rich site in the
city centre of Assen, the provincial capital of Drenthe. As a
result of further development and growth of the museum, a new
large underground exhibition hall is realised. The expansion
provides an underground connection of the exhibition hall with
the monumental main building. To realise this connection, an
underground excavation right underneath the monumental
Bailiff’s House is executed.
The excavation, to a level of about 8 m below ground
surface, is realised in two separate building pits: the main
excavation for the exhibition hall and the indoor excavation
(Figure 1) below the monumental Bailiff’s House. The indoor
wet deep excavation is retained by jet grout walls (VHP-
grouting). These walls also support and reinforce the existing
shallow foundations (Figure 2). To achieve the required wall
thickness of about 1.0 m up to 1.5 m two rows of columns are
installed in a triangular mesh of 0.6 m to 0.7 m. Each column
