204
Proceedings of the 18
th
International Conference on Soil Mechanics and Geotechnical Engineering, Paris 2013
Proceedings of the 18
th
International Conference on Soil Mechanics and Geotechnical Engineering, Paris 2013
using methylene blue tracer in studying the shapes of clay
fracturing reported vertical cracks formed in overconsolidated
clays of K
0
>1. The measured K
0
values were higher than when
approximated by the established K
0
-OCR correlations (by
Mayne and Kulhawy, 1982). A similar conclusion was made by
Hamouche et al. (1995), who also found that the horizontal
pressure determined by fracturing corresponded well to the self
boring pressuremeter and Marchetti dilatometer results.
1.2
Indirect methods of K
0
determination
Skempton (1961) made use of four ways of determining the
capillary pressure of the undisturbed samples in the laboratory:
direct and indirect measurement of the load preventing swelling,
analysis of the undrained strength measured in the triaxial
device, and measurement of pore water suction in the triaxial
specimen. The averaged capillary pressure from the four
methods was used to compute the effective horizontal stress,
and the pore pressure coefficient was determined in the triaxial
apparatus.
Figure 1. The influence of creep on σ´
v max
. position at oedometer test.
Burland and Maswoswe (1982) used the method in
supporting the use of direct measurements of horizontal stresses
in London clay: Their suction based results agreed well with the
self boring pressuremeter and the push-in 'spade' by Tedd and
Charles (1981).
The current version of the Skempton's procedure makes use
of the „suction probe“ capable of direct measurement of
capillary suctions within undisturbed samples taken by a thin
walled samplers (e.g., Hight et al., 2003). Doran et al. (2000)
studied the changes of pore pressures and effective stresses in
the laboratory specimens on sampling and preparation. They
concluded that using isotropic elasticity in the 'suction method'
results in underestimating the K
0
. The only up-to-date
alternative in London clay projects seems to be to estimate K
0
on the basis of lift-off pressures measured in self-boring
pressuremeter
tests,
although
interpretation
remains
controversial (Hight et al, 2003).
The correlation methods for determination of K
0
are
represented by the Jáky formula for normally consolidated soils
and by its extensions to cope with the overconsolidated soils in
the form of K
0oc
=(1sinφ
c
')×(OCR)
α
. The most common
alternative for the exponent is α=sinφ (Mayne and Kulhawy,
1982), or α=0.5 (Meyerhof, 1976). Some studies indicated α ≈
1.0 (Lefebvre et al., 1991; Hamouche et al., 1995). Using such
correlations however neglects other effects than the stress
history (unloading), for example creep and cementation that
might have developed in the soil in situ, and may lead to
erroneous estimation of the values of K
0
. Creep moves the
position of the real maximal vertical stress to the position of an
apparent maximal vertical stress (Fig. 1). The oedometer test is
a common technique for evaluating σ´
v max.
An experimental determination using the advanced triaxial
instrumentation (stress path testing, local LVDT gauges
mounted on the specimens etc.) was suggested by Garga and
Khan (1991) and Sivakumar et al. (2009). The latter proposed
and experimentally confirmed a new expression K
0oc
=1/η(1-(1-
ηK
0nc
)OCR
(1-χ)
), which takes account of OCR (parameter χ is
determined by 1-D and isotropic compression tests on
undisturbed specimens) and of anisotropy (parameter η is
determined from a CIUP test). K
0nc
can be determined, for
example, by Jáky's formula.
Doležalová et al. (1975 in Feda, 1978) made use of the
displacements measured after unloading the massif by means of
a gallery. The deformation parameters of the rock were
determined by independent in situ testing and then the FEM was
used to simulate elimination of the monitored displacements of
the gallery. The stresses necessary for the simulation were
considered the in situ stresses in the massif. A similar approach
using an advanced hypoplastic model is presented further.
The review shows that in determining initial stresses in
overconsolidated clays a single method can hardly be sufficient.
The best way seems taking good quality samples (thin wall
sampler) and measuring suctions, and comparing the result with
a direct measurements, for which Marchetti dilatometer (DMT),
push-in spade-shaped pressure cells or self boring pressuremeter
seem most promising. If available, convergence measurements
of a underground cavity (gallery) evaluated using a numerical
model with an advanced anisotropic constitutive model is
believed the best method.
2
GEOLOGY AND CHARACTERISTICS OF CLAY
INVESTIGATED
Different methods were used to evaluate K
0
of clay from Brno,
Czech Republic. The investigated calcitic silty Brno Clay
(“Tegel”) of Miocene (lower Badenian) age belongs to the
Neogene of Carpathian foredeep, and reaches the depth of
several hundred metres. Sound Tegel has a greenish-grey
colour, which changes to yellow-brown to reddish-brown colour
at the zone of weathering closer to surface. According to X-Ray
analysis there is a substantial percentage of CaO (ca 20%) and
the main minerals are kaolinite (ca 23%) and illite (22%),
calcite (20%), quartz (17%), chlorite (up to 10%) and feldspar
(Boháč et al., 1995). Tegel exhibits stiff to very stiff
consistency. The clay is overconsolidated but the height of
eroded overburden is not known. Above the Miocene clay there
are Quartenary gravels overlain by loess loam. The clay is
tectonically faulted. The groundwater is mostly bound to
Quartenary fluvial sediments, and the collectors are typically
not continous. However the clay is fully water saturated.
In Tegel there is about 50% of clay fraction, w
L
is about
75%, I
P
about 43%, the soil plots just above the A-line at the
plasticity chart and its index of colloid activity is about 0.9.
3
MARCHETTI DILATOMETER MEASUREMENTS
. Due to creep
however the test produces a pseudo-overconsolidation value of
σ
vmax
*' instead of the present overconsolidation pressure σ
vmax
'.
Hence, both the OCR and K
0,OC
values determined by the
correlations and not considering creep (ageing) are
overestimated.
At the site the current phreatic water table was 4.7 metres under
the surface and top layer of about 5.5 metres was excavated
some 30 years ago. This generated negative pore water
pressures, which have not fully dissipated yet. At the current
depth of 11.7 metres the pore pressure of -32 kPa was measured
(after dissipation of excess pore pressures caused by the
sounding) by a push-in spade pressure cell. The present vertical
effective stress in the depth of 11.7 metres calculated from the
soil unit weight and pore water pressure was σ
v
' = 185 kPa.
The K
0
was measured using Marchetti (1980) flat
dilatometer. The measured K
D
according to Marchetti (1980)
was 8.0 and K
0
derived using the empirical equation
K
0
= (K
D
/1.5)
0,47
– 0.6 was K
0
= 1.6. This is substantially
