2744
Proceedings of the 18
th
International Conference on Soil Mechanics and Geotechnical Engineering, Paris 2013
measurement of representative deformation and strength
properties at this depth can be problematic.
2.1 Laboratory testing of core
Core samples subjected to laboratory testing are affected by
disturbance and stress relief and can give erroneous results
which usually represent a significant underestimate of the insitu
stiffness of the material. This leads to over-design of footings,
higher costs and in some cases the footings can be impractical
to design or construct.
2.2 In situ testing by SPT or cone tests
Two of the insitu tests commonly used in ground investigations;
standard penetration tests and cone penetration tests, are either
not appropriate for testing at significant depths or cannot
penetrate relatively competent founding materials. For example,
the results of SPTs at reasonable depth (say 30 m) must be
considered to be unreliable due to the rod weight and the
resulting ineffectiveness of the impact from the hammer. It is
also of very little value to report an ‘SPT’ value of 50 blows for
some nominal (say 50 mm penetration). Such a result cannot be
interpreted to give an estimate of ground stiffness.
Cone penetrometer tests are ineffective where they cannot
penetrate moderately competent ground. Predrilling to
overcome frictional resistance is not a solution since refusal
often occurs at the tip.
2.3 Pressuremeter and cross-hole seismic tests
High quality pressuremeter testing and cross-hole seismic
testing provide a practical method for obtaining estimates of the
deformation parameters of the rock at different strain levels.
The crosshole seismic test provides estimates of small
strain modulus which cannot be applied directly to analysis of
footings where strains in the ground under dead, live and wind
loading are significantly higher than those experienced during
seismic testing. As deformation parameters depend on the strain
level imposed in the test, this must be taken into account in the
test interpretation.
The pressuremeter on the other hand provides deformation
properties at strain levels which are commensurate with those of
the ground when subjected to service loading from the building.
On some sites however, for example in deep alluvial deposits,
pressuremeter testing may result in significant disturbance to the
ground and hence the results of such testing may not be of
benefit. Self-boring pressuremeter tests can overcome this
problem, however they may be impractical in relatively hard
materials such as discussed in Section 4.
2.4 Instrumented pile load tests
Deformation properties of the ground under load can be
obtained from an appropriately designed test on an instrumented
pile. The results can be used to supplement those obtained from
the tests described in Section 2.3 prior to final design of the
footing system.
Load cells (typically Osterberg cells) are located in the pile
at chosen depths, while displacement transducers can be located
below the tip. By placing one Osterberg cell close to the base of
the pile in conjunction with a displacement transducer, the load-
displacement performance of the base of the pile can be
measured. It is a relatively straight forward process to then back
calculate a representative modulus for the material immediately
below the pile toe.
By combining the results from pressuremeter and cross-hole
seismic tests (adjusted to take into account strain levels), a
reasonable level of confidence can generally be obtained to
undertake the footing design.
The overall pile load-displacement performance can also be
measured and provides a means of back-figuring pile and
ground properties for use in a group settlement analysis package
such as PLAXIS or FLAC.
2.5 Application of in situ testing to modulus estimates
The methods for estimating ground modulus described in
Sections 2.3 and 2.4 are demonstrated for the design of footing
systems for two towers. Section 3 describes the application to
the design of the proposed 1000 m Nakheel tower in Dubai
which is to be founded in a weak calcareous siltstone (UCS of
about 2 MPa). Section 4 considers design for a group of tall
towers (up to 300 m high) founded in deep alluvial deposits
comprising very dense silty sand and hard sandy silt.
3 NAKHEEL TOWER, DUBAI
3.1 The tower and ground conditions
The Nahkeel Tower in Dubai was designed to extend to a height
in excess of 1 km. With about 2,000,000 tonnes dead load, the
structure would have been one of the heaviest ever built. The
project was placed on hold in early 2009 at a stage when about
half of the foundations had been constructed.
The high bearing pressures applied to the ground coupled
with the soft calcareous rock ground conditions present at the
site provided a significant challenge to the design of the footing
system.
3.2 Foundation system
Based on prior but limited knowledge of the ground conditions
in Dubai, the foundation system concept adopted for the tower
was a piled raft. The raft design had a variable thickness, being
up to 8 m under the most heavily loaded structural elements.
Design founding depth was at about 20 m below ground level,
and at the base of a 120 m diameter excavation supported by a
circular, embedded diaphragm wall. Approximately 400
barrettes were proposed, for installation to depths of between
approximately 60 m and 80 m below ground level. The design
of the barrettes had to consider not only the control of ground
response to the tower loading, but also various regulatory
requirements and constructability issues.
3.3 Ground investigation
The ground investigation (Haberfield and Paul, 2011)
comprised an extensive laboratory testing program on core
samples together with pressuremeter and crosshole seismic
testing. The self-boring pressuremeter tests extended to depths
of up to 200 m below ground level. Cross-hole seismic testing
was undertaken in arrays of 3 boreholes with 3 m centre-to-
centre spacing between the boreholes.
Figure 2 shows the values of initial loading modulus (E
i
)
calculated from laboratory unconfined compression strength
(UCS) tests, pressuremeter tests and cross-hole seismic tests.
The small-strain cross-hole seismic tests gave estimates of
modulus which ranged between about 3 to 7 times those
measured in the pressuremeter tests at the same depths. This
difference is consistent with the effects of strain level on
modulus. To obtain a modulus value for engineering design
adopting the strain levels appropriate to field behaviour, the
cross-hole values were reduced by a factor of five.
The modulus values measured in the UCS tests showed a
wide scatter. An upper bound to the results over the depth of
interest is around 600 MPa, which is about half the value
estimated from the pressuremeter test results.
3.4 Instrumented pile load tests
The preliminary foundation design was based on the results of
the in situ tests. However, prior to the detailed design stage,
three test barrettes with cross-sectional dimensions of 1.2 m ×
