871
Technical Committee 104 /
Comité technique 104
stability of a slope that is under steady state seepage conditions,
with water emerging from close to the toe with the upper part of
the slope being unsaturated.
A numerical back-analysis with a kinematic hardening
plasticity constitutive model is performed, replicating many
features of the physical model observations. In particular, both
the physical and numerical models highlight how earthquake-
induced excess pore pressures can lead to a rise in the water
table and a loss of suction and strength in an unsaturated slope.
Moving to steeper slopes, Aklik and Wu (2013) describe a
study of geotextile-reinforced walls, standing at angles of up to
85
. Model tests were performed using a geotechnical
centrifuge to induce collapse of the slopes using the ‘gravity
turn-on’ method. These simple tests explored the failure
mechanisms within the slope and the embedded geotextile
layers. A simple camera system was used to record frequent
images as the slopes ‘grew in height’. The failure mechanisms
were quantified through particle image velocimetry analysis.
The failures were shown to occur above the toe of the slopes,
and controlled by the spacing of the geotextile layers. This
study illustrates how complex geotechnical systems can be
investigated using simple rapid experiments in a compact
geotechnical centrifuge, with image analysis providing detailed
quantification of the soil failure mechanisms.
Dashti et al (2013) present a series of centrifuge model tests
results exploring the performance of buried water reservoir
structures, made from concrete. The test arrangement includes a
novel transparent laminar shaking container which allows the
internal deformation to be visualised. In addition, the end walls
of the container have pressure pad sensors, to record the
distribution of pressure during shaking.
The tests are performed with and without a buried structure,
providing calibration data for 2D and 3D numerical models.
Time histories from historic earthquakes are used as the input
shaking motion, after filtering out frequency components that
are irrelevant or would damage the centrifuge. The data from
these tests is currently being used by the Los Angeles
Department of Water and Power, to assess the seismic
performance of existing and planned subsurface reservoirs.
3.2
Ground improvement
Two studies describe experimental work to determine
benchmarking data for the performance of ground improvement
techniques for enhanced seismic performance.
Bahadori et al. (2013) report a series of shaking table tests
used to evaluate the performance of tyre chips as soil
reinforcement to improve liquefaction resistance. The stiffness
and damping properties of tyre-chip – sand mixtures are
assessed through intensively instrumented physical model tests.
A level ground surface was modelled, and the results from an
array of accelerometers were used to derive stress-strain loops
at different elevations within the soil. These loops allow the
stiffness and damping ratio to be determined.
An alternative novel material to improve the seismic
performance of structures is expanded polystyrene – known as
geofoam. In the example application studied by Dave et al.
(2013), a layer of geofoam is used at the rear of a retaining wall.
Tests were performed on a
1m
3
sample using the shaking table
at IIT Bombay. A surcharge was applied at the top of the
retained soil to mimic field scale stress conditions. Varying
magnitude of shaking were applied, whilst the pressures and
acceleration within the backfill and at the wall facia were
recorded. This soft layer of geofoam served to reduce the
ground accelerations felt at the wall, and lowers the lateral
pressures generated within the backfill.
The final contribution on the topic of ground improvement,
to calibrate new analysis methods, is a paper describing a new
foundation system for embankments on soft soil. Detert et al.
(2013) describe a hybrid structure comprising two parallel sheet
pile walls connected by a tension membrane. The embankment
is constructed in top of the membrane, and undergoes reduced
settlement due to the support from the membrane.
This system acts to reduce embankment settlements through
a combined action. The sheet pile walls prevent spreading,
whilst the membrane generates tensile forces when distorted,
supporting the embankment and relieving the soft layer of load.
To validate this foundation system, parallel streams of
research using complementary physical and numerical
modelling have been undertaken. Physical modelling is
particularly important since large deformations and a complex
composite system are involved. However, once the mechanisms
of behaviour are clarified, suitably calibrated numerical
modelling is planned to allow the system to be optimised.
3.3
Vibration screening barriers
The final paper in this category is concerned with geotechnical
barriers to insulate sensitive areas from vibrations created by
railway traffic. Masoumi and Vanhonacker (2013) describe an
experimental programme studying the transmission of
vibrations through a bed of sand,
10 m
3
in volume. Careful
attention was given to bed uniformity and scaling laws, with the
vibration frequency being scaled up to reflect the reduced scale
of the experiment.
The impedance of the soil bed was first established by
impact testing, to determine stiffness and damping parameters
for subsequent numerical analysis. Vibration transmission tests
were then performed using a line of accelerometers at the soil
surface, and a surface foundation with a vibrating live load.
Tests were performed with and without concrete isolation walls,
which were shown to reduce the transmission of vibrations. The
results were compared with complementary numerical results.
4 PERFORMANCE DATA FOR DESIGN AND
OPERATION
4.1
Pipeline protection from anchor dragging
Offshore pipeline-soil interaction is a relatively new field of
investigations, which has greatly benefited from the
development of physical modelling techniques. The advantage
in terms of resources and timescale compared to field testing is
obvious. More importantly, the recent development of
sophisticated motion control has enabled modelling of the full
life cycle of a pipeline, from the 3D dynamic motion at the
touch down zone during laying, to the on-bottom stability under
storm loading and large ‘post-failure’ lateral sweeps under
thermal buckling. The large database developed from physical
modelling tests over the last five years has been used to develop
analytical models now currently used in design (Gaudin &
White 2009, White & Cathie 2011).
Pipelines that cross shipping lanes or lie in regions of
intensive fishing often require protection from anchor dragging.
If a ship anchor or fishing trawlgear snags a pipeline both may
be damaged. Pipelines can be protected by burial, often with
rockdump backfill.
To assess the required depth of burial, and the performance
of the protection scheme, the interaction between a passing
anchor and the backfill must be assessed. Physical modelling is
commonly used for this purpose, since the large deformations
and chain-anchor-backfill interactions must be properly
accounted for.
A model testing arrangement that can be used to determine
performance data for direct use in design is described by
Bezuijen et al. (2013). An elongated centrifuge strongbox is
used to allow significant lengths of anchor dragging to be
simulated. A faithful reproduction of a ship anchor,
manufactured by 3D printing, is attached to a miniature chain,
and dragged over a model seabed. A pipeline is buried in a
trench that has been back-filled with coarser material.
The paper compares results from tests performed at 1g and at
80g in the Delft centrifuge. The results can be compared using
