870
Proceedings of the 18
th
International Conference on Soil Mechanics and Geotechnical Engineering, Paris 2013
penetration, but the former is temporary whilst the latter may
have an irreversible effect on the properties of the surrounding
soil.
The model tests showed the variation in pile penetration
resistance into saturated sand with jetting flow rate. The
response is highly non-linear. There is only a modest reduction
in resistance at low flow rates, until a significant fall in
penetration resistance occurs. Full liquefaction appeared present
during penetration at the highest flow rate, with negligible
resistance encountered. Scaling of this behaviour is naturally
very challenging, particularly if the response is due to both
particulate (internal erosion) and continuum (effective stress
reduction) effects.
Arnold and Laue (2013) describe an experimental study into
the load distribution beneath surface foundations, and the
influence of the relative stiffness between the foundation and
the soil. Centrifuge model tests were performed with a vertical
point load applied at the central point of the foundation. Two
model foundations were used, with and without edge stiffening
to represent building walls. The stiffer foundation shows a more
even distribution of foundation-soil pressure. The same
response is evident in field measurements using pressure cells
built into building foundations.
2.3
Flow-induced migration through porous media
Truong et al (2013) describe a set of experiments using an
apparatus which is somewhat similar to Darcy’s (Figure 1).
They have studied clogging effects during one-dimensional
flow using a 2.5m long cylinder, 0.18 m in diameter. Unlike
Darcy, they used electrical transducers to record flow rates and
pressure, mercury manometers being general outlawed in
modern laboratories. The experimental results show the
inadequacy of assuming Darcian flow if the pore fluid
introduces fine particles that may accumulate and block pores.
The paper presents an example of steady flow through the tube,
with a Darcian distribution of pore pressure. On injection of a
bentonite slurry, the pore rapidly clog downstream of the
injection point, causing a sudden rise in the upstream pore
pressure.
These results are highly relevant to seepage through dams or
embankments, where careful control of drainage is important to
assure stability. The experiments illustrate that flow regimes
and pore pressure distributions can be quickly altered if fine
particles are transported within the pore fluid. The intrinsic
permeability within Darcy’s Law may be a material property,
but fines migration can rapidly change the composition of a
material, and therefore its properties.
Sarma and Sarma (2013) discuss the flow of cementitious
material into the zone surrounding a bored pile. They report a
detailed laboratory investigation which mimicked the bored pile
construction process to evaluate the parameters controlling the
thickness of the sol zone at a pile wall that is strengthened by
the inflow of cementing products during construction. The
distance of impregnation was identified by a novel staining
method, in which the carbonated cementation products are
highlighted. The subsequent measurements of impregnation
depth are extremely detailed, and have been elegantly
interpreted into a link between voids ratio and particle size.
2.4
Reactive effects on soil behaviour
Cardoso and Nogueira Santos (2013) describe a careful physical
model of electrokinetically-enhanced consolidation behaviour.
Their study is motivated by the potential use of electrokinetics
to improve the efficiency of soft ground improvement, in
combination with drains. They initially consider one-
dimensional consolidation of kaolin clay, using a modified cell
that allows a voltage to be applied across the sample.
Improvements in the rate of consolidation are identified, with c
v
increasing by a factor of 6, typically.
Subsequent tests use a larger Rowe cell, incorporating
vertical drains that encourage radial and well as vertical flow.
The electrical field is then applied between a central drain and
the outer circumference of the sample. In this case the beneficial
effect of the radial flow overshadows the benefit from the
electrokinetic effect. The conclusion is that electrokinetics can
enhance engineered consolidation of soft ground, but only in
certain circumstances.
A similar study using a physical model of an element of
cemented barrier cut-off wall is reported by Verástegui-Flores et
al. (2013). Laboratory apparatus was modified to provide long
term measurements of the permeability and small strain stiffness
of cement-stabilised bentonite clay. This is a widely used
material for barrier wall construction, but can suffer from
deterioration through chemical attack. Two novel pieces of
apparatus were developed to provide simple methods of
measuring the shear wave velocity and, in the second apparatus,
the permeability. The combined data of permeability and small
strain stiffness over >250 days shows the effect of sulphate
attack on the properties of the cement-stabilised soil. As the
cement hydrates, the pores become blocked by cementation,
which also raised the shear wave velocity. However, when
sulphate is added, these processes are halted, based on the
measurements. These new types of test allow the performance
of specific stabilised soil mixes to be determined via simple
laboratory tests that are more representative than conventional
methods.
3 PERFORMANCE DATA TO CALIBRATE MODELS
3.1
Seismic soil-structure interaction
Several papers contributed to this 18
th
ICSMGE focus on the
seismic response of slopes, and the estimation of pore pressure
build-up and lateral spreading. The use of highly instrumented
centrifuge models provides detailed evidence of the internal
accelerations and pore pressures within the slope. The models
are usually plane strain, with a transparent window allowing the
soil movement to be observed.
Due to the inability to undertake seismic field experiments,
physical modelling has been used extensively to investigate
seismic soil-structure interaction, using both 1g shaking tables
and on-board shakers in geotechnical centrifuges. The ability to
recreate accurate dynamic loading conditions and to measure
pore pressure generation and soil displacements have proven to
be essential to provide data to calibrate models.
Haeri et al. (2013) report a centrifuge model test of a slope
reinforced by a 3
3 pile group, with a surficial non-liquefiable
layer. During shaking, the soil liquefies and slides downslope,
applying passive loading to the pile. The maximum loads occur
at the start of the event, after the motion begins but before the
soil is fully liquefied. As the degree of liquefaction increases the
passive load reduces, although the load from the unliquefied
layer persists.
Back-analysis of the lateral pressures and the resulting
internal bending moment in the piles shows that the Japanese
Roads Authority (JRA) design code provides good predictions
of the maximum pile loading. However, the detailed distribution
of load within a pile group is not considered in this code. The
experimental data shows that significant shielding effects are
present. The most heavily loaded pile was actually the
downslope pile: the high load came not from the passive
pressure from the upslope soil, but from the loss of active
support on the downstream pile, as the downslope soil failed
and slid away.
Higo et al (2013) report a study of embankment stability
under seismic loading, using similar experimental techniques in
a geotechnical centrifuge. They used a compacted clayey silty
sand with the same composition as a material used to reinforce
river levees in the Kansai region. Their study is focused on the
