267
Proceedings of the 18
th
International Conference on Soil Mechanics and Geotechnical Engineering, Paris 2013
1
Comparison of Stress-Strain Behaviour of Carbonate and Silicate Sediments
Comparaison de la réponse contrainte-déformation de sédiments carbonatés et siliceux
Safinus S., Hossain M.S., Randolph M.F.
Centre for Offshore Foundation Systems, The University of Western Australia, Perth, Australia
ABSTRACT: Compared to silica sand, carbonate sand has considerably higher angularity, lower grain hardness and higher intra-
particle porosity, which result in high friction angles and compressibility. The corresponding dilatancy is affected strongly by the
confining stress. Thus, even for low relative densities, dilation occurs at low confining stresses, reflecting the greater particle
interlocking compared to silica sand. However, with the increase of confining stress, the dilatancy is suppressed quickly, and finally
diminishes completely at a relatively low stress level, due to particle degradation. This distinctive characteristic significantly
influences the behaviour of continuously penetrating spudcan foundations in calcareous sediments. Centrifuge tests were carried out
on spudcan foundations penetrating multi-layer soils with an interbedded strong layer composed with either carbonate or silica sand.
All measures of spudcan punch-through severity were significantly lower for interbedded carbonate sand despite its higher friction
angle (
crit
= 40
) compared to silica sand (
crit
= 34
). For the spudcan penetration through the sand layer to the lower clay layer, the
soil failure mechanisms quantified by particle image velocimetry (PIV) analysis allowed for identifying the differences in the
evolution of sand frustum beneath the advancing spudcan. The spreading angle of the frustum, which determines the size of the
projected bearing area, was found to be proportional to the mobilised dilatancy.
RÉSUMÉ : Comparativement au sable siliceux, le sable carbonaté a une angularité considérablement plus élevée, une plus faible
dureté de grain et une porosité intra-particulaire plus élevée, ce qui a pour effet de produire un angle de frottement et une
compressibilité élevés. La dilatance de ce dernier est fortement affectée par la contrainte de confinement. Ainsi, même pour de faibles
densités relatives, le comportement dilatant peut se produire pour des contraintes de confinement faibles, reflétant une tendance à
l’imbriquement des particules plus
élevée par rapport au sable siliceux. Cependant, la dilatance est rapidement réprimée lorsque la
contrainte de confinement augmente, et finalement disparaît complètement pour des niveaux de contrainte relativement faibles, du fait
de la dégradation des particules. Cette caractéristique particulière influence de manière significative le comportement des fondations
‘
spudcan
’ lors de leur
pénétration dans des couches de sédiments calcaires. Des essais en centrifugeuse ont été réalisés sur des
fondations
‘
spudcan
’
pénétrant des sols multi-couches comprenant une couche intermédiaire composée soit de sable carbonaté, soit de
sable siliceux. Toutes les mesures de sévérité du risque de pénétration du « spudcan » étaient significativement plus faibles pour le cas
d’une couche intermédiaire de
sable carbonaté, en dépit
du fait que l’
angle de frottement soit plus élevé (
crit
= 40
), par rapport au
sable siliceux
(
crit
= 34
). Pour la pénétration du « spudcan » à travers la couche de sable jusqu'à la couche sous-
jacente d’argile, les
mécanismes de rupture du sol quantifiés par vélocimé
trie d’image de particule (PIV) ont révélé des différences d’évolution du
tronc
de sable en dessous du « spudcan ». L
’angle d’ouverture
du tronc de sable, qui détermine la taille de la surface portante projetée,
s’est
révélé être proportionnel à la dilatance mobilisée.
KEYWORDS: carbonate, silicate, dilation, spudcan foundations.
1 INTRODUCTION
Carbonate sediments are prevalent in Australian waters and in
the Caspian Sea, Arabian Gulf, South China Sea, offshore Qatar
and offshore Florida. Standard geotechnical analysis models
were generally developed for silica sediment. Extreme care
should be exercised when applying those models for carbonate
sediments and indeed predictions using routine bearing capacity
methods linked to the friction angle have been shown to be
inappropriate. This is exacerbated for continuous penetration of
spudcan foundations due to the gradually rising stress levels
(SNAME 2008, InSafeJIP 2010). Discrepancies between the
predicted and measured behaviour can be significant, especially
in cases involving loose sand or high stresses. This results
mainly because of the critical characteristics of calcareous
sediments such as crushable particles, high in-situ void ratios
and compressibility. With increasing stress level, grain particles
are crushed, which alters the stress-strain behaviour.
Many studies have been undertaken in the last decades to
improve understanding of the stress-strain behaviour of
carbonate sediments (Datta et al. 1980, Evans 1987, Golightly
and Hyde 1988, Semple 1988, Coop 1990, Al-Dhouri and
Poulos 1992, Randolph et al. 1999, Desrosiers and Silva 2002).
Bioclastic carbonate sediments comprising skeletal and shell
fragments usually have very angular grains, and hence high
friction angles and low particle crushing strength parameter,
Q
(see Table 1). The use of friction angle as the sole strength
indicator for sand often results in excessive overestimation of
bearing capacity and underestimation of penetration depth
(Overy 2012). Dutt et al. (1985) reported a much lower apparent
friction angle (19°), through back analysis of the measured
spudcan penetration response, compared to the value obtained
from a direct shear test (
crit
= 50°). Semple (1988) recorded
relatively large settlements of offshore jack-up footings in
carbonate sediments, which was attributed to the high
compressibility of the soil. Current offshore design guidelines
SNAME (2008) and InSafeJIP (2010) recommend using a
reduced design friction angle (by as much as 25
) and a
mobilisation (reduction) factor of ~0.25, respectively, for
assessing spudcan penetration resistance in carbonate sands.
In stratified sediments, with interbedded sand layers, the
problem is even more complex. The likelihood and severity of a
