3467
Predicting Settlements of Shallow Footings on Granular Soil Using Nonlinear
Dynamic Soil Properties
Prédiction des tassements de fondations superficielles sur des sols granulaires en utilisant
des propriétés dynamiques non linéaires du sol.
Stokoe K.H., Kacar O.
University of Texas at Austin, Texas, USA
Van Pelt J.
CDM Smith, Inc. Dallas, Texas, USA
ABSTRACT: The governing design criterion for shallow foundations in freely draining granular soils is usually permissible
settlement. Due to difficulties in obtaining undisturbed samples of granular soils composed of mainly sands and gravels,
settlement predictions are generally based on correlations with in-situ penetration tests. In this study, field seismic
measurements are used to evaluate small-
strain (“elastic”) shear moduli of granular soils (G
max
). These small-strain moduli,
combined with nonlinear normalized shear modulus-shear strain (G/G
max
-
log γ) relationships, are used to predict settlements
of shallow foundations under working loads. The G/G
max
-
log γ relationships are based on models developed from dynamic
resonant column (RC) tests of reconstituted sands and gravels. The combination of field G
max
results and laboratory G/G
max
-
log γ relationships have been implemented in a finite element program (PLAXIS) via a subroutine.
Settlement predictions
with this approach are illustrated by comparison with a load settlement test using a 0.91-m diameter footing. At working
stresses, nonlinear footing settlements were predicted quite well, similar to predictions with traditional CPT and SPT
procedures.
RÉSUMÉ: Le critère de dimensionnement pour des fondations superficielles sur sols granulaires est, souvent, le tassement
admissible. A cause de la difficulté à obtenir des échantillons intacts de sols sableux et graveleux, les prévisions de
tassement sont basées sur des corrélations déduites des essais in-situ. Dans la présente étude, des mesures sismiques in situ
sont utilisées pour évaluer les modules de cisaillement elastique en petites déformations des sols granulaires (Gmax). Ces
modules, combinés avec des relations de variation non-lineaire module-distorsion, G/Gmax-
log γ, sont utilis
és pour prévoir
les tassements des fondations. Les relations G/Gmax-
log γ sont bas
ées sur des modèles développés à partir d’essais à la
colonne résonante (RC) sur des éprouvettes reconstituées de sables et graviers. La combinaison des mesures de Gmax in-situ
et des relations G/Gmax-
log γ obtenues en laboratoire sont introduites via une sous
-
routine dans un programme d’éléments
finis (PLAXIS). Les prédictions des tassements obtenues avec l’approche proposée sont présentés en les comparant aux
résultats d’un essai de charge
ment utilisant une fondation de 0.91m de diamètre. Sous contraintes de service, les previsions
de tassements non-lineaires sont bonnes. Elles sont similaires à celles déduites des procédures SPT et CPT traditionnelles.
KEYWORDS: in-situ seismic testing, nonlinear dynamic properties, granular soil, footing settlement
1 INTRODUCTION
In shallow foundation design, bearing capacity and
settlement are the two main criteria considered. For freely
draining granular soils, permissible settlement becomes the
governing factor in most cases. Laboratory tests to predict
the stress-strain behavior of soils generally require an
undisturbed sample which is nearly impossible and/or very
expensive to recover from the field for granular soils.
Therefore, settlements of shallow foundations on such soils
have traditionally been predicted using empirical
correlations that relate in-situ penetration test results with
load-settlement tests or case histories. In this article, an
approach based on field seismic evaluation of small-strain
(“elastic”) shear modulus (G
max
) combined with nonlinear
normalized shear modulus-shear strain (G/G
max
-
log γ)
relationships is presented. The effects of increasing
confining pressure and strain amplitude on soil stiffness
during loading of the footing are incorporated in this
formulation. The approach has several important benefits
including: (1) in-situ seismic testing which can readily be
performed in all types of granular soils, including gravels
and cobbles (2) continuous load-settlement curves that are
evaluated to stress states considerably above those expected
under working loads, and (3) a methodology that is
appropriate for all types of geotechnical materials, even
those where the effective stresses change with time.
2 TRADITIONAL AND RECENT SETTLEMENT-
PREDICTION METHODS
One of the first methods of predicting footing settlements on
granular soils was proposed by Terzaghi and Peck (1948).
They conducted plate-load tests on 300-mm square plates on
sand and then predicted the settlements of full-size footings
using an empirical relationship. Meyerhof (1965) proposed a
method where the settlements were predicted based on
standard penetration test (SPT) blow count, N
60
. One of the
most widely used methods today was originally proposed by
Schmertmann (1970). He used elastic theory, model load
tests, field cone penetration tests (CPT) and finite element
analysis to develop the approach. In Schmertmann’s method,
the soil stiffness is expressed as an equivalent elastic
modulus which is based on CPT results. Burland and
Burbidge (1985) reviewed a data set of case histories and
developed a method using corrected SPT results. In all
methods, a key parameter, the strain dependency of the soil
stiffness, is not directly considered.
One of the earliest methods to take the strain dependency
of the soil stiffness into account was proposed by Berardi
and Lancellotta (1991). They proposed an iterative scheme
1
redicti g Settleme t of Shallow Footings on Granular Soil Using Nonlinear
Dynamic Soil Properties
Prédiction des tassements de fondations superficielles sur des sols granulaires en utilisant
des propriétés dynamiques non linéaires du sol.
Stokoe K.H., Kacar O.
University of Texas at Austin, Texas,USA,
Van Pelt J.
CDM Smith, Inc. Dallas, Texas,USA
ABSTRACT: The governing design criterion for shallow foundations in freely draining granular soils is usually per issible
settlement. Due to difficulties in obtaining undisturbed samples of granular soils composed of mainly sands and gravels,
settlement predictions are generally based on correlations with in-situ penetration tests. In this study, field seismic
measurements are used to evaluate small-strain (“elastic”) shear moduli of granular soils (G
max
). These s all-strain moduli,
combined with nonlinear normalized shear modulus-shear strain (G/G
max
-log
γ
) relationships, are used to predict settlements of
shallow foundations under working loads. The G/G
max
-log
γ
relationships are based on models developed from dynamic
resonant column (RC) tests of reconstituted sands and gravels. The combination of field G
max
results and laboratory G/G
max
-log
γ
relationships have been implemented in a finite element program (PLAXIS) via a subroutine. Settlement predictions with this
approach are illustrated by comparison with a load settlement test using a 0.91-m diameter footing. At working stresses,
nonlinear footing settlements were predicted quite well, similar to predictions with traditional CPT and SPT procedures.
RÉSUMÉ: Le critère de dimensionnement pour des fondations superficielles sur sols granulaires est, souvent, le tassement
admissible. A cause de la difficulté à obtenir des échantillons intacts de sols sableux et graveleux, les prévisions de tassement
sont basées sur des corrélations déduites des essais in-situ. Dans la présente étude, des mesures sismiques in situ sont utilisées
pour évaluer les modules de cisaillement elastique en petites déformations des sols granulaires (Gmax). Ces modules,
combinés avec des relations de variation non-lineaire module-distorsion, G/Gmax-log
γ
, sont utilisés pour prévoir les
tassements des fondations. Les relations G/Gmax-log
γ
sont basées sur des modèles développés à partir d’essais à la colonne
résonante (RC) sur des éprouvettes reconstituées de sables et graviers. La combinaison des mesures de Gmax in-situ et des
relations G/Gmax-log
γ
obtenues en laboratoire sont introduites via une sous-routine dans un programme d’éléments finis
(PLAXIS). Les prédictions des tassements obtenues avec l’approche proposée sont présentés en les comparant aux résultats
d’un essai de chargement utilisant une fondation de 0.91m de diamètre. Sous contraintes de service, les previsions de
tassements non-lineaires sont bonnes. Elles sont similaires à celles déduites des procédures SPT et CPT traditionnelles.
KEYWORDS: in-situ seismic testing, nonlinear dynamic properties, granular soil, footing settlement
1 INTRODUCTION
In shallow foundati n design, be ring capacity and
settlem nt are the two main criteria considered. For freely
draining granular soils, permissible settlement becomes the
gov rning factor in most cases. Laboratory tests to predict
the stress- train behavior of soils generally require an
undisturbe sample which is nearly impossible and/or very
exp nsive to recover from the field for granular soils.
Therefore, s ttlements of shallow foundations on su h soils
have traditionally been predicted using empirical
correlations th t relate in-situ penetration test results with
l ad-settlement tests r ca e histories. In this article, an
approach based on field seismic evaluation small-strain
(“elastic”) shear modulu (G
max
) combined with nonli ar
no malized shear modulus-shear strain (G/G
max
-log
γ
)
relationships is presented. The effects f incr asing
confini pressure and train amplitude on soil stiffness
during loading of the footi g are incorporated in thi
formulation. The approach has s veral important benefits
incl ding: (1) in-situ seismic t sting which can readily be
performed in all types of granular soils, including gr vel
and cobbles (2) continuous load-settlement curves that are
evaluated to stress states considerably above those expected
under working loads, and (3) a methodology that is
appropriate for all types of geotechnical materials, even
those where the effective stresses change with time.
2 TRADITIONAL AND RECENT SETTLEMENT-
PREDICTION METHODS
One of the first methods of predicting footing settlements on
granular soils was proposed by Terzaghi and Peck (1948).
They conducted plate-load tests on 300-mm square plates on
sand and then predicted the settlements of full-size footings
using an empirical relationship. Meyerhof (1965) proposed a
method where the settlements were predicted based on
standard penetration test (SPT) blow count, N
60
. One of the
most widely used methods today was originally proposed by
Schmertmann (1970). He used elastic theory, model load
tests, field cone penetration tests (CPT) and finite element
analysis to develop the approach. In Schmertmann’s method,
the soil stiffness is expressed as an equivalent elastic
modulus which is based on CPT results. Burland and
Burbidge (1985) reviewed a data set of case histories and
developed a method using corrected SPT results. In all
methods, a key parameter, the strain dependency of the soil
stiffness, is not directly considered.
One of the earliest methods to take the strain dependency
of the soil stiffness into account was proposed by Berardi
and Lancellotta (1991). They proposed an iterative scheme
where the soil stiffness was evaluated based on the corrected
