1175
Determination of soil-water retention curve for a young residual soil using a small
centrifuge
Détermination de la courbe de rétention d’eau pour un sol résiduel jeune à l’aide d´une petite
centrifugeuse
Reis R. M., Saboya F., Tibana S., Marciano C.R., Ribeiro A.B.
State University of Norte Fluminense Darcy Ribeiro
Sterck W.N.
Fugro – In Situ, Brazil
Avanzi E.D.
Federal University of Paraná, Curitiba, PR, Brazil
ABSTRACT: The soil-water retention curve (SWRC) determination usually involves time-consuming conventional methodologies
such as those based on soil-water plate extractors, tempe cells, suction plates and filter-paper method. In order to overcome the
considerable large time lag necessary for SWRC evaluation, an alternative methodology for direct determination of the SWRC of the
unsaturated soil was developed using a commercially available small centrifuge with a swinging type rotor assembly without in-flight
instrumentation. The testing consists of spinning up four initially saturated soil specimens until constant water content is achieved for
a given angular speed. The soil-suction relationship is determined by relating the respective water content to the suction magnitude
induced by the ceramic plate at the specimen´s base. This methodology was applied for evaluating the SWRC of a young residual soil
from gneiss using both, undisturbed and remolded soil specimens until 900 kPa matric suction magnitude. The testing results show
good agreement to the similar SWRC obtained by conventional methods and depicted using the Van Genuchten (1980) mathematical
model. Overall, it can be concluded that the methodology proposed ensures good agreement in determining the SWRC of studied soil.
RÉSUMÉ : Généralement on utilise, pour la détermination de la courbe de rétention du sol (SWRC), des méthodes classiques assez
fastidieuses telles que celles basées sur des extracteurs de plaques sol-eau, des cellules Tempe, des plaques d'aspiration et du papier-
filtre. Pour surmonter ces difficultés et réduire le temps nécessaire pour l'évaluation de la courbe de retention (SWRC), une autre
méthodologie a été développée pour la détermination directe du SWRC de sol non saturé en utilisant une petite centrifugeuse
disponible sur le marché, avec un assemblage de rotor de type oscillant sans instrumentation couplée. Le test consiste à faire tourner
quatre échantillons de sol initialement saturés jusqu'à ce qu´une teneur en eau constante soit obtenue pour une vitesse angulaire
donnée. La relation sol-aspiration est déterminée en fonction de la teneur en eau relative à la magnitude d'aspiration induite par la
plaque de céramique à la base de l'éprouvette. Cette méthode a été appliquée pour évaluer le SWRC d'un sol résiduel jeune de gneiss
en utilisant des échantillons de sol à la fois, non perturbés et reconstitués jusqu'à une magnitude d´aspiration matricielle de 900 kPa.
Les résultats des essais sont pareils à ceux obtenus, pour la SWRC, par les méthodes conventionnelles et en utilisant le modèle
mathématique de Genuchten Van (1980). Dans l'ensemble, on peut conclure que la méthode proposée montre une bonne concordance
dans la détermination du SWRC du sol étudié.
KEYWORDS: unsaturated soil, soil-water retention curve, centrifuge technique, soil suction.
1 INTRODUCTION
The application of a centrifuge for inducing an unsaturated state
within a soil sample is not new. Soil drainage tests using
centrifugal flow have long been recognized as a valid and
efficient way of determining the SWRC. Briggs and McLane
(1907) and Russell and Richards (1938) have proposed
methodologies for estimating the SWRC using a centrifuge. The
proposed methodologies essentially consist in draining an
initially saturated soil specimen under a certain induced gravity.
Different lower values of water contents can be rapidly
achieved by increasing the applied induced gravity. The
decrease in soil moisture is then related to the corresponding
increase in the soil suction magnitude. Gardner (1937),
interested in evaluating the effect of the induced gravity over
the matric potential of the soil, carried out several sets of filter
paper tests within a centrifuge apparatus, with soil specimens
placed over a free water surface. The matric potential of the soil
was evaluated after achieving hydraulic equilibrium and related
to the induced gravity applied. The work of Gardner (1937)
showed clearly that a relationship exists between the soil matric
potential and the induced gravity applied to the soil sample.
Other researchers (e.g. Hassler and Brunner, 1945; Croney et
al., 1952) have attempted inducing suction magnitudes at the
soil specimen´s boundaries through use of ceramic disks or
cylinders. Corey (1977), working with soil samples allowed to
free drainage due to centrifugal flow, observed that the negative
pressure gradient induced within the soil sample acts inward,
once the centrifugal force induces water flow on the outward
direction. Corey (1977) noted that the capillary pressure
induced within the soil sample varies along the length of the
sample, from zero magnitude at the outflow boundary (which is
open to the atmosphere), to a maximum value at the top
boundary of the sample. The author concluded that at each
angular velocity, the sample drains until the capillary force
equals the centrifugal force induced over the water molecules.
Recently, Khanzode et al. (2002) attempted to determine the
SWRC at a single induced gravity by simultaneously testing
several soil specimens with different ceramic disks attached.
The main idea was inducing different magnitudes of suction at
each specimen simultaneously. The results, however, exhibit a
pore matching when compared to the results obtained by using a
pressure plate apparatus. Although the authors suggest the need
of complementary studies for understanding the influence of the
induced gravity over the soil´s suction, a careful analysis of the
methodology applied by Khanzode et al. (2002) indicates that it
was not ensured that the specimens tested in the centrifuge had
identical initial densities and water contents as those tested in
