317
A Comparison Between the Shear Strength Measured with Direct Shear and Triaxial
Devices on Undisturbed and Remolded Soils
Une comparaison entre la résistance au cisaillement mesurée avec appareils de cisaillement direct
et triaxiaux sur les sols non remaniés et remoulés
Castellanos B.A., Brandon T.L.
Virginia Polytechnic Institute and State University
ABSTRACT: A comparison is presented between the shear strength measured with direct shear and triaxial devices based on the
results from tests conducted on undisturbed and remolded soils. A series of consolidated-undrained (CU) triaxial tests and
consolidated-drained (CD) direct shear tests were performed on undisturbed samples from the New Orleans area. These tests were
conducted on soils ranging from low plasticity silts to organic fat clays. The results from the undisturbed samples showed that the
drained friction angles obtained from the CU triaxial tests were considerably higher than those obtained from the CD direct shear
tests. On the other hand, results obtained from remolded test specimens originating from a variety of locations showed almost no
difference in the effective stress shear strength. The difference in the results can be explained in part by the fact that natural soils can
have a preferred particle orientation or anisotropic fabric based on the deposition or formation of the soil. In the case of the New
Orleans soils, the undisturbed samples exhibited evidence of horizontal deposition. In remolded samples, the soil is more much more
homogeneous and isotropic, so every plane in the soil may be expected to have similar shear strength.
RÉSUMÉ : En utilisant des résultats de tests effectués sur des sols non perturbés et remoulés, une comparaison entre la résistance au
cisaillement mesurée par cisaillement direct et avec des appareils triaxiaux est présentée. Une série d’essais triaxiaux consolidés non
drainés et d’essais de cisaillement direct consolidés drainés ont été effectués sur des sols non remaniés de la région de la Nouvelle-
Orléans. Ces tests ont été effectués sur des sols allant de vases de plasticité faible à des argiles organiques de haute plasticité. Les
résultats des échantillons non perturbés ont démontré que les angles de friction drainés obtenus à partir d’essais triaxiaux consolidés
non drainés sont considérablement plus élevés que ceux obtenus à partir d’essais de cisaillement direct. D’autre part, les résultats
obtenus à partir de spécimens remoulés originaires de lieux divers n’ont montré pratiquement aucune différence dans la résistance
effective au cisaillement. La différence des résultats peut s’expliquer en partie par le fait que les sols naturels peuvent avoir une
orientation préférée des particules ou un tissu anisotrope basé sur le dépôt ou la formation du sol. Dans le cas des sols de la Nouvelle-
Orléans, les échantillons intacts ont présenté des signes de déposition horizontale. Dans les échantillons remoulés le sol est beaucoup
plus homogène et isotrope, de sorte que l’on peut s’attendre à ce que chaque plan du sol ait la même résistance au cisaillement.
KEYWORDS: direct shear, triaxial, shear strength, fully softened, remolded, undisturbed, shear strength measurement.
1 INTRODUCTION
The triaxial and direct shear devices have been historically used
successfully to measure the peak shear strength of soils. In
geotechnical projects, these tests are often used interchangeably
to determine effective stress or drained shear strength
parameters without regard to the potential difference in the
results. A series of consolidated-undrained (CU) triaxial tests
and consolidated-drained (CD) direct shear tests were
performed on undisturbed samples from the New Orleans area.
These tests were conducted on soils ranging from low plasticity
silts to organic fat clays. The bulk of these tests were conducted
as part of the reconstruction of the flood protection system,
which was severely damaged by Hurricane Katrina in 2005.
The city of New Orleans is located on alluvial soils that are
part of the delta formed by the sediments of the Mississippi
River. According to Dunbar and Britsch (2008), the surficial
soils were formed during the Holocene and consists of fine
grained soils extending to about 17 m to 25 m deep in most of
the area, while exceeding 46 m deep at some locations. Dunbar
and Britsch (2008) stated that this layer is characterized by
stacked and generally horizontal layering created by the
deposition mechanism.
Numerous CU triaxial compression tests and CD direct shear
tests were performed on undisturbed test specimens trimmed
from 125 mm diameter tube samples to characterize the
effective stress shear strength parameters of the soils in the
greater New Orleans area. These test results were assessed to
allow a comparison of the effective stress shear strength
parameters for the undisturbed samples based on soil type,
plasticity characteristics, etc. The tests conducted on the
undisturbed tube samples were assessed based on the peak shear
strength.
A series of direct shear tests and triaxial tests were also
conducted on remolded soil specimens to determine the
fully
softened
shear strength. The fully softened shear strength was
defined by Skempton (1970) as the drained peak shear strength
of a clay in its normally consolidated state. According to
Skempton (1977), the fully softened shear strength can be
measured on remolded normally consolidated specimens. Three
different soils from various locations in the USA were tested to
examine the difference in the fully softened strength parameters
obtained from triaxial and direct shear test apparatuses.
2 PREVIOUS WORK
The direct shear and triaxial devices have been used for over 70
years to determine drained shear strength parameters of soils
(Saada and Townsend 1981). These two devices have marked
differences in the stress condition that is developed in the test
specimen. Some of the biggest differences are: 1) strain
boundary conditions, 2) failure plane orientations, and 3)
principal stress orientation. In the triaxial device, the
intermediate and minor principal stresses are equal and are
normally specified at the beginning of the test. In the direct
