1559
Liquefaction characteristics of crushable pumice sand
Caractéristiques de liquéfaction des sables de pierre ponce sensibles à l’écrasement
Orense R.P., Pender M.J.
University of Auckland, New Zealand
ABSTRACT: Because of their highly crushable nature, there are concerns on whether liquefaction evaluation procedures which were
empirically derived from hard-grained sands are applicable to pumice sands. To understand their liquefaction characteristics, several
series of undrained cyclic triaxial tests were performed on both reconstituted and undisturbed pumice specimens. Results showed that
the effect of relative density on the liquefaction resistance of pumice sands was not as significant when compared to that of hard-
grained sands. Correlations were also made between the liquefaction resistance from the cyclic tests on undisturbed specimens and
those from field tests conducted at sites where the samples were obtained. CPT and sDMT-based empirical methods of estimating the
cyclic resistance ratio did not agree well with the laboratory-obtained values, while that based on shear wave velocity produced better
correlation. The results obtained can be used to evaluate the in-situ liquefaction potential of pumiceous deposits.
RÉSUMÉ : En raison de leur forte sensibilité à l’écrasement, il y a des doutes quant à savoir si les méthodes d'évaluation de
liquéfaction, obtenues de façon empirique avec des sables à grains durs, sont applicables à la pierre ponce. Pour comprendre les
caractéristiques de liquéfaction, plusieurs séries de tests triaxiaux cycliques non drainés ont été effectués sur des échantillons de pierre
ponce remaniés et intacts. Les résultats ont montré que l'effet de la densité relative sur la résistance à la liquéfaction des sables de
pierre ponce n'était pas aussi important comparé à celui sur sables à grains durs. Des corrélations ont également été faites entre la
résistance à la liquéfaction obtenue par les essais cycliques sur échantillons intacts et les résultats des essais in-situ menés sur les sites
d’échantillonnage. Les méthodes empiriques d’estimation du rapport de résistance cyclique, basées sur des essais CPT et SDMT, ont
sous-estimé les valeurs obtenues en laboratoire, tandis que celles basées sur la vitesse des ondes de cisaillement ont produit de
meilleures corrélations. Les résultats obtenus peuvent être utilisés in-situ pour évaluer le potentiel de liquéfaction des dépôts de pierre
ponce.
KEYWORDS: Pumice deposits; undrained cyclic test; liquefaction; cyclic resistance ratio; particle crushing.
1 INTRODUCTION
The recent earthquakes in Christchurch have demonstrated the
impact of soil liquefaction to the built environment (e.g., Orense
et al. 2011). With the central government, local councils and
community residents in New Zealand now fully aware of the
devastating effects of earthquakes in general and of soil
liquefaction in particular, attention has shifted to the seismic
performance of local soils, i.e., whether soils in certain localities
will undergo the same degree of liquefaction as the
Christchurch soils did.
Pumice deposits, which originated from a series of volcanic
eruptions centred in the Taupo and Rotorua regions, are found
in several areas of the North Island. They are frequently
encountered in engineering projects and their evaluation is a
matter of considerable geotechnical interest. Because of their
lightweight, highly crushable and compressible nature, they are
problematic from engineering and construction viewpoint.
Moreover, no information is available as to whether empirical
correlations and liquefaction procedures derived for hard
grained soils are applicable to pumice deposits because there
has been very little research done on their characteristics.
The authors presented preliminary results of an experimental
programme they conducted to investigate the undrained cyclic
characteristics of undisturbed and reconstituted pumiceous soils
through cyclic triaxial testing (Orense et al. 2012). This paper
further discusses more undrained cyclic triaxial results.
Moreover, geotechnical investigations, including cone
penetration testing (CPT) and seismic dilatometer testing
(sDMT), were conducted at the sites where the undisturbed
pumice samples were obtained. Finally, the validity of the
conventional methods of evaluating the liquefaction resistance
developed for hard-grained sands was examined to see if they
are applicable to crushable soils like pumice.
2 MATERIALS USED
Two sets of materials were used in the triaxial tests. The first set
consisted of undisturbed pumiceous soils obtained through push
tube sampling at two sites in Waikato in central North Island:
(1) at Carrs Rd in Hamilton; and (2) Mikkelsen Rd in Waihou.
The samples from Carrs Rd site, which were obtained at depths
between 8.0-8.5 m using 60 mm push tubes, were completely to
heavily weathered ignimbrite. The closest SPT N-value was 18
(at depth=15m). The undisturbed Mikkelsen Rd samples were
sourced at three depths: 3.0-3.3m, 6.0-6.6m and 12.0-12.4m,
using 60mm push tubes. The SPT N-values were 11 and 13 at
depth=4.5m and 6.6m, respectively. Because of its loose nature
(some cores were lost), the push tubes were placed in a freezer
for 1-2 days before the samples were extracted. Although the
two sets of samples thus taken may have been “disturbed” one
way or the other by the sampling, the degree of disturbance may
be considered insignificant. In addition, care was taken during
handling and transport; thus they are referred to as
‘undisturbed” in this paper.
The other set of materials used was commercially-available
pumice sand. This is not a natural deposit but was derived by
processing sand from the Waikato River. The particles were
centrifugally separated from the other river sand particles so
that the samples consist essentially of pumice grains.
