904
Proceedings of the 18
th
International Conference on Soil Mechanics and Geotechnical Engineering, Paris 2013
Karpurapu and Bathurst (1992) used a non-linear finite element
analysis to simulate the controlled yielding concept for static
load and concluded that compressible inclusion with t=0.01h (t
– thickness of compressible inclusion, h – height of the wall)
would provide active stress conditions in the backfill, if the
stiffness of the compressible inclusion is sufficiently small.
Experimental investigations of the concept of reduction of
seismic load on the retaining wall in the presence of geofoam
inclusion were performed by several researchers on reduced
scale models tested on shaking table (Hazarika et al. 2002,
Bathurst et al. 2006, Zarnani and Bathurst 2007
)
. Hazarika et al.
(2002) showed reduction in the peak lateral loads in the range of
30% to 60% compared to that on an identical structure but with
no compressible inclusion. Zarnani and Bathurst (2007) noticed
that the magnitude of dynamic lateral earth force was reduced
with decreasing geofoam modulus. Horvath (2010) highlighted
compressive stiffness as the single most important behavioural
characteristic of any compressible inclusion influencing the
reduction. Athanasopoulos–Zekkos et al. (2012) observed that
EPS of 20 kg/m
3
density and relative thickness (t/h) of 15% to
20% can reduce the seismic pressure by up to 20%, and the
seismic displacement of the wall by up to 50%, depending on
shaking intensity and height of wall.
The available literature highlighted that with the use of EPS
geofoam, the earth pressures on the rigid retaining walls can
even be reduced below the active earth pressures. However,
behaviour of EPS geofoam and its influence on the earth
pressure reduction under seismic loading conditions are not well
understood, especially in the presence of realistic surcharge
loads, and need to be investigated further. Hence, the present
study is aimed at evaluation of earth pressure under combined
surcharge and seismic loading and to assess effectiveness of
EPS geofoam, through experimental investigations on small
scale models tested on 1-D shaking table facility.
3 EXPERIMENTAL PROGRAM
The physical tests described in this paper were carried out on
1.2 m × 1.2 m shaking table located at the Indian Institute of
Technology Bombay. The table has 10 kN payload capacity and
is driven by a 100 kN capacity Schenk hydraulic actuator with
ancillary controller and PC software. The table was driven in
the horizontal direction only, as it is noted that the horizontal
component of seismic induced dynamic earth loading is
typically the most important loading for the application under
investigation. The table can excite the rated payload at
frequencies up to 50 Hz and ± 5g. The maximum displacement
of the table is ±125 mm. The instrumented retaining wall
models were built in a stiff strong box (1.2 m long
0.31 m
wide and 0.7 m high) and bolted to the steel platform of the
shaking table. Detailed diagram and pictorial view of
experimental set up are illustrated in Figs. 1-2. The model
retaining wall was placed at a distance of 0.10 m from one of
the ends, allowing 1.1 m as backfill length behind retaining
wall. A 15 mm thick stainless steel plate was used as a model
retaining wall and was instrumented with 7 diaphragm type
earth pressure cells, attached flush with the surface of the wall.
The wall was restrained laterally using three universal load cells
rigidly connected to the other side of the retaining wall at 125,
325 and 555 mm elevations. One side of strong box was made-
up of Plexiglas and other sides of stainless steel. The inside
surface of the Plexiglas is covered by 120 mm wide and 60 µm
thick greased polyethylene sheet with 10 mm overlap with each
other. The combination of friction-reducing membrane and rigid
lateral bracing was adopted to ensure that the test models were
subjected to plane strain boundary conditions. A plywood sheet
was bolted to the bottom of strong box, and a layer of sand was
epoxied to the top surface of plywood to create a rough surface,
so as to simulate backfill continuity in vertical direction.
A series of experiments were carried out without geofoam
and with geofoam inclusion at wall-backfill interface. In all
experiments, the sand was backfilled at 68% relative density
using portable travelling pluviator (Dave and Dasaka, 2012) and
top surface was manually leveled. The actual relative densities
achieved in each test during the backfilling were monitored by
collecting samples in small cups of known volume placed at
different locations. Previous studies of the authors highlighted
that EPS panel of density of 10D (10 kg/m
3
) and 75 mm
thickness (t/H = 0.125) helps in maximum reduction in earth
pressure by mobilization of its elastic compression. Hence, EPS
panel of 10 kg/m
3
density and dimensions of 700 mm x 300 mm
and 75 mm thickness, prepared using hot-wire cutter, was
pasted to retaining wall using ABRO tape to have proper
contact of EPS panel with retaining wall during the test.
Uniaxial compression tests were carried out on EPS samples at
an axial strain rate of 10%/minute, and yield strength of the EPS
geofoam was found as 29.3 kPa, as shown in Fig. 3.
Figure 1. Detailed diagram of experimental setup
Figure 2. Picotrial view of experimental setup
To apply uniformly distributed surcharge on the backfill, a
rubber bellow was placed over an 8 mm thick rubber sheet
laying on the surface of the backfill. Specially designed
neoprene rubber bellow of 250 kPa capacity with non-return
pneumatic valve was connected to a compressor to apply
regulated pressure. A steel plate of 10 mm thickness with
attachments to measure surface settlement was placed between
rubber bellow and rubber sheet and a steel plate of 10 mm
thickness was placed on the rubber bellow such that when
inflated with compressed air, the plate moved upwards to
mobilize reaction from frame, which was rigidly connected to
the tank, thereby transferring pressure to the sand fill.
Three LVDTs were used to measure vertical settlement at top of
the backfill at 150 mm, 450 mm and 750 mm from retaining
wall. The LVDTs were firmly mounted on the reaction frame
with magnetic stand and were rested on angles welded on steel
plate. Four accelerometers (PCB Piezotronics) were used to
obtain acceleration-time excitation history. Out of these, three
