Sunteți pe pagina 1din 19

GAZETAS, G. (1981). GCorerhniqur 31, No.

2, 161-179

Strip foundations on a cross-anisotropic soil layer subjected to


dynamic loading

G. GAZETAS*

A rigorous semi-analytical formulation is presented to des modules d’ilasticite de Young horizontal/vertical est
study the static and dynamic response of rigid strip important, les d&placements statiques sont moindres et les
footings supported on the surface of a horizontally caracteristiques de response dynamique sont totalement
layered soil deposit. Each layer is modelled as a homo- differents de celles de sediments de sols isotropiques
geneous cross-anisotropic medium with a vertical axis of equivalents.
material symmetry while the loading consists of har-
monically time-varying horizontal or vertical forces and NOTATION
moments. The solution, based on an experimentally dimensionless frequency factor
verified relationship among the anisotropic soil para-
(= OB &PIE”))
meters that uncouples the wave equations in closed- half of the foundation width
B
form, is exact in that it properly accounts for the true
Dij elastic stiffness parameters of an aniso-
boundary conditions at the layer interfaces and the
surface. Comprehensive parametric studies are presented tropic material (equation (5))
in the form of normalized displacement-load or rota- Ev, En Young’s modulus in vertical and hori-
tion-moment ratios as functions of dimensionless geo- zontal directions
metric and material parameters. Simple and sufficiently t&n, %I shear modulus in vertical and hori-
accurate formulas ofdirect practical applicability are also zontal planes
given for static displacements and resonant frequency h WI@
factors. The results clearly demonstrate the significance of H thickness of soil stratum
soil anisotropy in determining undrained static and k
dynamic response of foundations; soils with a large ratio
WlB
L pseudo-distortional wave potential
of horizontal to vertical Young’s moduli experience
smaller static displacements and quite different dynamic MO moment on a rigid foundation (with
response characteristics from equivalent isotropic soil respect to axis Oy)
deposits. En/&
;; pseudo-dilatational wave potential
L’article presente une formulation semi-analytique horizontal force of a rigid foundation
P”
rigoureuse pour itudier la reponse dynamique et statique
P” vertical force of a rigid foundation
de semelles Mantes rigide reposant a la surface de
r dE,,dz
sediments de sob a couches horizontales. Chaque couche
est modelisee sous la forme dun milieu homogtne de U horizontal displacement
section anisotrope ayant un axe vertical de symetrie W vertical displacement
materielle, tandis que la charge est constitute de forces X horizontal co-ordinate
et moments verticaux ou horizontaux a variation Z vertical co-ordinate
harmonique dans le temps. La solution, bake sur une &WP)
relation verifite expkimentalement entre les parametres
JCGVHIP)
de sots anisotropes qui decouple les equations d’onde de ii H horizontal displacement of a rigid
forme fermte, est exacte en ce qu’elle tient bien compte des
foundation
conditions de hmite vraie aux interfaces des couches et a la
4 vertical displacement of a rigid founda-
surface. L’article presente des etudes parametriques
completes sous la forme de rapports rotation-moment ou tion
d&placement-charge normalises en tant que fonctions de normal and shear strain components
paramttres materiels et giomitriques adimensionnels. VVH Poisson’s ratio for effect of vertical
L’article prtsente egalement des formules simples et strain on horizontal strain
suffisamment precises ayant une application pratique vHH Poisson’s ratio for effect of horizontal
directe pour des facteurs de friquence de resonance et des strain on complementary horizontal
deplacements statiques. Les resultats montrent bien strain
I’importance que represente l’anisotropie des sols pour la soil damping ratio
determination de la reponse dynamique et statique non
angle of rotation of a rigid foundation
drainte de fondations; dans le cas de sob dont le rapport
(with respect to axis Oy)
Discussion on this Paper closes 1 September, 1981. For soil density
further details see inside back cover. normal and shear stress components
*Case Western Reserve University, Cleveland, Ohio. frequency of vibration (rad/s)
161
162 G. GAZETAS

INTRODUCTION rotational excitation.


A large body of experimental evidence suggests that In order to reach conclusions of the widest
soils and rocks in nature invariably exhibit some possible applicability with actually encountered
degree of anisotropy in their response to stresses. soil deposits, this Paper idealizes soil as a layered
Natural clay deposits, for instance, during their medium, with each layer being a linearly elastic
formation by sedimentation and subsequent one- cross-anisotropic continuum having a vertical axis
dimensional consolidation acquire a fabric that is ofsymmetry and exhibiting linear hysteretic-type of
characterized by particles or particle-units oriented damping when subjected to dynamic stresses. A
in a horizontal arrangement. This preferred semi-analytical solution is presented based on a
orientation and the resulting electrochemical transformation that uncouples the Navier-type
bonds among the clay particles are the cause of governing equations in terms of pseudo-distor-
cross-anisotropic deformational behaviour, as has tional and pseudo-dilatational wave potentials.
been demonstrated in numerous experimental Analytical expressions can, thereby, be obtained for
investigations (Ward, Samuels & Gutler, 1959, displacements and stresses in each layer and the
1965; Barden, 1971, 1972; Franklin & Mattson, correct boundary conditions at layer interfaces can
1972; Kirkpatric & Rennie, 1972; Saada & Ou, be enforced in continuous form. On the other hand,
1973; Gibson, 1974; Atkinson, 1975; Saada, in order to satisfy the mixed boundary conditions
Bianchini & Shook, 1978; Yong & Silvestri, 1979). at the surface, a numerical scheme has been devised
This anisotropy becomes marked in heavily over- involving discretization into a number of uniformly
consolidated clays, such as the London clay; an spaced nodal points and use of a so-called fast
extreme case is the eventual formation of strongly Fourier transform algorithm to perform the
anisotropic laminated shales, slates and mudstones. pertinent integrations. As a result, the method
Sands also exhibit deformational anisotropy, possesses the flexibility of the numerical techniques
which arises chiefly from the influence of gravity (like the finite element method) in properly hand-
and particle shape on the deposition process. ling any prescribed mechanical behaviour of the
Experimental investigations have revealed that soil-foundation interface; thus the two commonly
sand particles have a strong tendency to adopt assumed extreme cases of adhesive and frictionless
preferential orientation with the maximum dimen- contact can be studied almost as easily as the more
sion aligned in a horizontal plane (Willoughby, realistic case of a contact obeying Coulomb’s
1967, Parkin, Gerrard & Willoughby, 1968); thus friction law.
the number of contact points per unit horizontal Yet, owing to its analytical character, the pre-
area is smaller than that per unit vertical area sented method is devoid of a crucial limitation of
(Rowe, 1962; Barden, 1963). Such a fabric is condu- the finiteelement techniques, namely, their inability
cive to greater radial than vertical compressibilities, accurately to model the radiation of wave energy at
as was shown by Gerrard (1968) and Arthur & very large horizontal and vertical distances from
Menzies (1972). A set ofequal-diameter spheres in a the oscillating foundation. It is true, of course, that
hexagonal array have cross-anisotropic properties special energy-absorbing boundaries, such as those
(Gassman, 1953). developed recently by Valliapan, White & Lee
There is a growing awareness of the need to (1977) for a cross-anisotropic material, can provide
account for the influence of soil cross-anisotropy partial remedy and lead to a more or less acceptable
when estimating foundation settlements or dis- dynamic finite element formulation. Nonetheless,
tribution of stresses in the ground. This is well besides their greater accuracy, analytical solutions
reflected in a recent state-of-the-art presentation on offer a clear economic advantage in terms of both
foundation behaviour by Burland, Broms & de computer storage and time. This makes quite
Mello (1977) as well as a number of publications feasible the performance of comprehensive para-
dealing with the mathematical modelling of soil metric studies aimed at evaluating the relative
anisotropy and its implications on foundation importance of anisotropy with a variety of charac-
response. However, essentially all of these studies teristic soil profiles ranging from homogeneous
are limited to considering only static loading halfspace to shallow stratum underlain by rigid
conditions. The interest in designing foundations rock. The results of such a study are presented here
subjected to dynamic loads (such as those arising in the form of graphs and simple formulas of direct
from supported machinery, sea waves, earthquakes practical applicability, and are compared with
and ground-transmitted traffic or blast vibrations) other relevant solutions.
makes necessary the study of the dynamic inter-
action of foundations with cross-anisotropic soil. SUMMARY OF PREVIOUS WORK
Accordingly, the objective of this Paper is to assess Several studies related to the present problem
the effect of soil anisotropy on the response of strip have been published and it is useful to review
foundations to dynamic vertical, horizontal and them briefly before proceeding to our analyses.
STRIP FOUNDATIONS ON CROSS-ANISOTROPIC SOIL LAYER 163

Studies related to the static problem proportional to the applied normal pressure. With
Michell(1900), extending the work of Boussinesq very good accuracy their main result simplified to
and Cerruti, presented a solution for stresses and 4-n
displacements in a halfspace exhibiting hexagonal WC%Y) = &* Y) (2)
anisotropy’ and subjected to a vertical or hori- rC(4 - n) GvHIEv + 11
zontal point force. Much later, his solution was where, in this case, n and G&E, fully describe soil
simplified and made popular by Barden (1963) who cross-anisotropy, since for an incompressible
proposed the cross-anisotropic halfspace as an medium
improved mathematical model for natural soils, V VH = l/2
clays or sands. The major conclusion of Barden’s (3)
V HH = l-n/2
study is that, for a realistic range of Poisson’s ratios,
as the ratio n of the horizontal to vertical Young’s As a direct consequence, soil reactions against rigid
moduli increases so does the load-spreading capa- smooth foundations of any shape are uniform. This
city of the soil; hence, both stress concentration is also true with isotropic incompressible media
along the load axis and surface settlement decrease. obeying equation (1) (Gibson, 1974). It appears that
Extensive studies on static interaction of founda- this interesting conclusion can be generalized:
tions with anisotropic soils were presented by Sveklo (1970) has found that in the case of rigid
Gerrard & Harrison (1970a, b); they reported com- foundations having circular, elliptical or ellipti-
plete solutions (stresses, strains, displacements) for cal-paraboloidal contact with an anisotropic
cross-anisotropic and orthorhombic homogeneous homogeneous halfspace, soil reactions are indepen-
halfspaces carrying circular or strip foundations of dent of both type and degree of anisotropy-a
infinite or zero rigidity that are loaded by vertical, conclusion that does not apply to foundation
horizontal and moment forces. Their results, settlements.
although in a somewhat complicated mathematical Finally, Hooper (1975) studied with a finite
form, have contributed much to current under- element formulation the interaction of circular
standing of the behaviour of anisotropic soils. The flexible rafts in adhesive contact with a cross-
same authors (Harrison & Gerrard, 1972) estab- anisotropic layered stratum. Using material pro-
lished the equivalence between earth reinforced by perties appropriate for the overconsolidated Lon-
means of thin sheets (or bars) of a stiffer material, don clay (Ward et al., 1959; Ward, Marsland &
and a cross-anisotropic (or orthorhombic) homo- Samuels, 1965; Uriel& Kaiiizo, 1971; Gibson, 1974;
geneous medium. As a concept, reinforced earth Atkinson, 1975), he demonstrated that total and
had been proposed by Casagrande to model natur- differential foundation settlements may be reduced
ally stratified soil deposits; Westergaard (1938) by about 40% under undrained conditions if cross-
worked out the idea and presented solutions for a anisotropy is taken into account; under drained
halfspace so stiff in the horizontal direction, that no conditions this reduction is only 5-20%, depending
lateral strains could occur. Such a medium is, in on the type of soil profile assumed (homogeneous
fact, a cross-anisotropic material with n = cc and or linearly heterogeneous).
V VH = 0. Particular types of cross-anisotropy,
Studies related to the dynamic problem
characterized by three instead of five independent
parameters, were studied by Wolf (1935) and Kirkner (1979) presented an analytical study of
Milovic (1972). the steady-state response of a circular foundation
Analytical results for an incompressible cross- on a cross-anisotropic halfspace whose elastic
anisotropic halfspace whose modulus, E,, varies parameters satisfied a certain constraint relation-
with depth according to ship (equation (6)), originally suggested by Carrier
(19461.
\ I
The motivation in establishing such a
E,(z) = rz (1) relationship was one of convenience: the three
while m, n, vVH and vHHremain constant, have been equations of motion can uncouple and thereby be
published by Gibson (1974), Gibson & Kalsi (1974) solved analytically. Several other researchers have
and Gibson & Sills (1974). Such a continuum, also used equation (6) to restrict material aniso-
which is of direct practical interest when deter- tropy. For instance, Payton (1975) obtained solu-
mining undrained settlements of foundations on tions for dynamic displacements due to a con-
deep saturated clay deposits, was found to behave centrated force suddenly applied at a point within
as a ‘Winkle? medium; regardless of geometry of an infinite elastic space; and Valliapan et al. (1976)
the loaded area, surface settlement is directly derived dashpot constants of energy-absorbing
boundaries for a plain-strain finite-element for-
’ The terms hexagonal anisotropy as well as transverse mulation.
isotropy are used in the literature and are equivalent to the An interesting conclusion of Kirkner’s work is
term cross-anisotropy employed in this Paper. that anisotropy may have different effects at high
164 G. GAZETAS

Poisson’s ratios vvu and vHH (Lekhnitskii, 1963;


Pickering, 1970). With the vertical axis z being the
axis of elastic symmetry (Fig. 1), the stress-strain
relationships appropriate for plane-strain condi-
‘- LO tions take the form
. _-w,-- - --,__,
^+ - _-_ _-_

Ldy--_-L-_:_- -_--
. . . . . . .. . . . . . . . .
ox = Di,e,+D~e,
::;. ‘., .. . .- .,..,.... . .‘:_... .:::. .‘.
. . . ,‘,.
. . . . . . .
.::..
:*
.:..:.
. .:*...
. . . . . . ....
.:. . ‘..
. ..-. ,‘.... .‘:;.,
.. ey = D,,a,+D,,s,
(4)
Cross-anisotropic ez = D,,e,+&s,
layer E,, E,, G,,r vVH, vHHj <
r XZ= Gvn Y,,
,/ I’ I I,, f / , i, /,/ “,‘,
, I’ ’ f , where the four elastic coefficients D, ,, D,,, D,, and
,,,I< / 1, / / / 1 I ‘/ ,,/‘f’, D are related to the Young’s moduli and
.‘. . . . * . . . . .. . . ~ . 9 . .
PZsson’s ratios

Dii = (E&)(1-nvvn2)

.,: . 1 . z . . . . ..-.. . :‘.


Di2 = (-%la)(nvvn2 +v&
: (5)
Di3 = (-%la)vvn(l +vnn)
Fig. 1. Soil profile and footing diagram
D,, = (&/a)(1 -vnu2)
and low frequencies of vibration. For example, on in which
an incompressible medium a circular footing
n = E,/E,
experiences static horizontal displacements that (54
decrease as the lateral soil stiffness increases rela- a = (1 + vu”) (1 - vnn - 2nvvn’)
tive to the vertical, i.e. as n = E,/E, increases; at Thermodynamic considerations require that the
high frequencies, however, larger dynamic displace- strain energy in an elastic material due to all
ments are associated with larger n values. Such a possible stress fields be non-negative. This imposes
conclusion is of practical significance since the
on the acceptable range of the elastic parameters
constraint relationship on which the analysis was
certain restrictions (Hearmon, 1961; Pickering,
based has been shown to be satisfied with sufficient 1970; Gibson, 1974; Hooper, 1975).
accuracy by many soils, as demonstrated in this
Paper. Constraint relationship
The constraint relationship which has been
DEFORMATIONAL ELASTIC SOIL
proposed by Carrier (1946) defines a subset of all
ANISOTROPY
cross-anisotropic materials by making the shear
Mathematical modelling
modulus Gv, a function of the other elastic
The use of an elastic model for soils at low parameters
working stresses has received much attention (e.g.
2
Poulos & Davis, 1974). It is generally accepted that D,,D,,-D,3
GVH =
for stress changes imposed by shallow foundations D,,+~D,,+~s
linear elastic theory provides an adequate engineer-
ing model enabling the prediction of settlements to Thus the number of independent elastic parameters
a sufficient degree of accuracy. This is especially reduces to four, and the equations of motion can
true e.g. for heavily overconsolidated clays (Wroth, uncouple and be solved in closed form--a con-
1971) and brittle sensitive clays (Yong & Silvestri, venient simplification which motivated the adop-
1979) stressed below their yield limit. Moreover, tion of equation (6). For a material with
elastic models lead to satisfactory answers in V”” = V”“ZV and E, = E,= E, equation (6)
problems of dynamic soil-foundation interaction; reduces to the well-known relationship between
for small dynamic strains such as those developed moduli and Poisson’s ratio of isotropic media:
beneath machine foundations, soils exhibit an 2G,, = 2G = E( 1+ v); that is, isotropy can be re-
approximately elastic behaviour (Richart, Woods covered from the proposed constraint relationship.
& Hall, 1970) while the local soil non-linearities From a physical standpoint, significant experi-
arising during strong earthquake or wave excita- mental evidence has come to support the use of
tion ‘do not significantly affect the response of the equation (6) with many types of soil (Gazetas,
structure’ (Kausel, Roesset & Christian, 1976). 1980b). This rather unexpected but most welcome
A cross-anisotropic elastic material is charac- conclusion was reached by testing the validity of
terized by five independent constants: two Young’s equation (6) against numerous published experi-
moduli E, and En, a shear modulus Gvn, and two mental data which are summarized in Table 1.
STRIP FOUNDATIONS ON CROSS-ANISOTROPIC SOIL LAYER 165

Using the reported values of n, vVH, vHH and E,, tested in triaxial compression, extraneous bending
shear moduli GvH were computed from the con- and shear end effects are generated; as a result the
straint relationship; they are also depicted in Table recorded E,, underestimates the actual modulus
1 for comparison with the experimental values. by an amount which increases with the degree of
The performance of equation (6) appears to be anisotropy. Consequently, Gv, may in reality be
quite satisfactory. In several cases calculated and somewhat larger than the reported values indicate,
measured moduli are nearly identical and in no case especially for the heavily overconsolidated London
do they differ by more than 20%. This is believed to clay. This might further improve the agreement
be within the range of possible error of the reported with values derived from the constraint relation-
G VH values, which are frequently derived from ship.
Youngs moduh E,, E, and E,, (Gibson, 1974). In conclusion, equation (6) appears to be quite a
Because, as pointed out by Pagan0 8c Halpin (1968) realistic assumption for a number of clays and its
and Saada & Bianchini (1977), when samples cut use in the theory that is presented here is thus fully
with their axis inclined at 45” to the vertical are justified.

Table 1. Anisotropic elastic constants of clays and evaluation of the constraint relationship

Description of soil

Heavily overconsolidated
T Reference

Ward et al. (1959)


T
n VVH ‘HH GvH/&
T(equation
Computed

Gv&v
(6))

London clay (Ashford) Ward et al. (1965)


(undrained loading) Gibson (1974)
Depth: 30 ft 1.35 0.50 0.325 0.35* 0.355
Depth: 50 ft 1.59 0.50 0.205 0.37* 0.41
Average of all 1.80 0.50 0.08 0.38* 0.46
samples
Heavily overconsolidated Atkinson (1975)
London clay (Barbican 200 0.19 OQO 0.5367 0.553
Arts Centre)
(drained loading)
Lightly overconsolidated Saada et al. (1978)
kaolinite clay
(Florida Edgar plastic
kaolin)
(undrained loading)
uF’ = 40 lb/in2 1.25 0.50 0.375 0.356 0.364
water content
= 40.7%
crc’= 60 lb/in’ 1.355 0.50 0.322 0.362 0.378
water content
= 38.7%
Normally consolidated Bianchini (1980)
illite clay (Grundite)
(undrained loading)
uc’ = 70 lb/in* 1.17 0.50 0.415 0.355 0.353
w = 29.5%
oc’ = 60 lb/in’ 1.13 0.50 0.436 0.322 0.310
w = 38.1%
Colorado clay shale Kaarsberg (1968) 1.38 0.197 0.266 0.423 0524
(drained loading)
Sensitive, naturally Yong & Silvestri
cemented Champlain (1979)
sea clay (Canada) 0.62 0.35 0.20 0.205” 0.187
(drained loading)

*Estimated from E,,, using equation (21) of Gibson (1974).


t Estimated by the Author from undrained tests on the basis of the theoretical formulas of Uriel & Kaiiizo (1971), as
explained in Appendix 1.
166 G. GAZETAS

PROBLEM FORMULATION AND SOLUTION: to IAand w as follows


CROSS-ANISOTROPIC LAYERED HALFSPACE
aN aH
Within each layer, dynamic equilibrium under u = z+bz
plane-strain conditions, as is appropriate for strip
loading, requires that the stresses satisfy the equa- (12)
dN aH
tions w = b--__
aZ ax
Substituting (12) in (9) while accounting for (10) and
.
(7) (11) leads, after some simple operations, to

(13)
where the displacements a@, z) and w(x,z) are a2H d2H 2a2H
related to the strain components E,(x, z), E,(x, z) and Jp+==F -gF
yXZ(x,z) as follows
au The general harmonic solution of (13) is
N = (A! eihmr + AU e - ihmr) ei(wl - hlx)
% = a;

aw H = (B’ eikV + B” e-‘k42) eiCM-kpx)


(14)
Ez = x (8)
where i = 4 - 1, provided that the parameters 1, m,
au aw pand 4, being in general complex numbers, satisfy
YXL =5+x
1212+m2 = 1 (15)
Equations (4), (7) and (8) describe completely the p2+q2 = 1 (16)
time and space variation of stresses, strains and
displacements. Upon eliminating stresses and In the above equations o is the frequency of
strains, the following Navier-type coupled equa- vibration (in rad/s), h = W/OL and k = co//?. It is easy
tions of motion are obtained to check by direct substitution that equations (14)

a2 u 2 2 azw
(along with equations (1.5) and (16)) constitute a
solution of equation (13). A’, A”, B’, B” in (14) are
~,t,= D,,~+GYH$+(D~~+CFH)- axaZ arbitrary constants of integration, to be determined
(9) from the boundary conditions of the problem.
2
Using equations (14) (12), (8), (10) and (4)
Pg=G”” ~~+D,,!$+Vh,+G,.~& expressions for u, w, err and ~~~ can be derived in
terms of the same constants ofintegration A’, A”, B’,
For computational convenience define B” (shown in matrix form in Table 2).
Having expressed stresses (or, r,,) and
a2+3, p+, q=;
displacements (u, w) in terms of four constants for
each layer, a total of 4n equations are needed to
(10)
determine all the unknown quantities, if n is the
j2+l, p+, b+$
number of soil layers in the deposit. These
33 33
equations are provided by the boundary conditions
It can then readily be verified that, due to the at each layer interface and at the loaded surface.
constraint relationship (equation (6)) The analysis is identical with the one described by
Gazetas (1980a) in connection with a soil deposit
A’-(p+q2)/b = ‘I’ (11) consisting of heterogeneous isotropic layers,
To uncouple equations (9) two potential although the basic solutions for u, w, a, and rxr
functions N(x, z) and H(x, z) are introduced, related within each layer are different in the two cases.

Table 2. Expressions for u, w, cr2 and r,,; f(z) = e”““= and g(z) = e*q=

10
-h*W* +bm*)D,,f(4 k*pdbv- 1)4, s(z) -h2(d2+bm*P,,lf@) -k2pq@v- I)D&dz) A’
h*ml(I+ b) GVH/(4 k*(p* - bq’) Gw g(z) -h* ml( 1 + b) GYH/f(z) k2 (P* -bq’) G&g(Z) B e,(w_h,x,
- ihlf(z) ikqbg(d - ihl/f(z) -ikqb/g(z) A”
ihmbf(z) ikpg(4 - ihmb/f(z) ikplg(4 B”
STRIP FOUNDATIONS ON CROSS-ANISOTROPIC SOIL LAYER 167

PH 6,
--+I4

b-&Z
..se _, ,....._ --7.
2 n=1/3 :‘I :.H, ‘..,,:: .‘_. I.;:. .‘...‘.,.. :::,.
,.
1 t :_:_ ,’ ; : ...- _.._.
I
I 1 (Isotropic)
1' I
/
l___,'
2
//‘_,
I /’ I
I / \
,
\ / _.---
4 / I __----
0 /
\ /k
\ _/ \ /C
:
, /

0 1 2 3 4
Frequencyfactor,
A,

03

Fig. 2. Undrained dynamic horizontal displacement-load ratio (H/B= 1): (a) in-phase component; (b) WA-out-of-phase
component

PRESENTATION OF RESULTS cross-anisotropic soils.


Results are obtained for vertical, sliding and The results are displayed in the form of
rocking vibrations of rigid massless footings normalized displacement-load amplitude ratios
subjected to vertical, horizontal and moment (hereafter called compliances) as functions of
loading varying harmonically with time. By dimensionless groups of key material and
allowing the frequency of vibration to vanish, geometric parameters
general solutions for the static displacements are
Ev 4 Ev 6, Ev BZ 4,
also readily derived. Only two extreme types of
p, Or -P, Or M0
mechanical behaviour of the soil-footing interface,
corresponding to adhesive or frictionless contact,
are discussed in the Paper since several analyses
have confirmed that a contact obeying Coulomb’s
=f(H
-,4v”H,vH”A”,r
B >
(17)

in which a,, 6, and 4, are the amplitudes of the


friction law leads to intermediate response
vertical, horizontal displacements and the angle of
amplitudes. Also, since layering conditions vary
rotation of the footing caused bv harmonic forces of
from site to site, only two simple, characteristic soil
amplitudes (per unit length) Pv, P, and M,
models, namely, a homogeneous cross-anisotropic
respectively; B is half the foundation width and H
halfspace and a homogeneous cross-anisotropic
the thickness of the soil deposit; Ev, n, vYHand vHH
stratum underlain by rigid rock, are examined here.
are the independent anisotropic soil parameters; A,
These models represent extreme categories of
is a dimensionless frequency factor defined as
actually encountered soil profiles and their study
can offer considerable insight into the dynamics of A, = ~W&/P) (18)
168 G. GAZETAS

I
2
Frequency factor, A,

K4

Fig. 3. Undrained dynamic horizontal displacement-load ratio (H/B = 3): (a) in-phase component; (b) W%-out-of-phase
component

and 5 is the critical damping ratio which reflects the initial displacements, caused by static loads during
internal energy dissipation in soil due to hysteresis or immediately after construction, take place due to
and friction. undrained shearing deformations of water-
Each displacement consists of two components; saturated clays that may exist in the soil deposit.
one in phase and one 90” out of phase with the Moreover, dynamic loads involve short time
applied harmonic load. The first represents the intervals between imposed stress changes and thus
recoverable, elastic component of deformation undrained conditions prevail in most saturated
while the second expresses the dissipation of energy soils (see, for example, Richart et al., 1970).
by waves propagating away from the foundation Under undrained conditions saturated soil
(radiation or geometric damping) and by hysteresis behaves as an incompressible solid and the two
and friction in the soil (internal damping). independent Poisson’s ratios, vvHand vHH,are given
by equations (3). Thus, the degree of material
anisotropy is uniquely described with the ratio
PARAMETRIC STUDY: UNDRAINED RESPONSE n = E,/E,, since the shear modulus ratio
Study of the undrained foundation response is of tn = G,,/E, can be computed in terms of n from the
particular geotechnical engineering interest since constraint relationship (6).
STRIP FOUNDATIONS ON CROSS-ANISOTROPIC SOIL LAYER 169

I
\
\
1.5- \
\ n=1/3
\
\

l-

1
:
& 0.5-
d
;o
m
z O-
P

“1
0 1 2
I
3
I
4

Frequency factor,
A,

04

Fig. 4. Undrained dynamic horizontal displacement-load ratio (H/B= co):


(a) in-phase component; (h) 9O%-out-of-phase

Horizontal loading
Figures 2, 3 and 4 portray the horizontal
displacementtload ratio Ev 6,/P, (horizontal
compliance) as a function of the frequency factor A,
and the degree of cross-anisotropy n, for three
homogeneous soil deposits. A wide range of
possible thicknesses is covered, from H/B = 1 (very
shallow deposit) to H/B = co (very deep deposit).
Both in-phase and !W-out-of-phase displacement
components are shown for a critical damping ratio
5 = 0.05. Adhesive contact is assumed between
footing and soil. Several conclusions can be drawn
from these figures.
Layer thickness. Regarding the effect of layer
thickness, for a given degree of anisotropy, it is n= E,tE,
evident that the existence of rigid bedrock at
relatively shallow depths drastically reduces the Fig. 5 Undrained static horizontal displacement-load
static and low-frequency foundation displace- ratio as a function of the degree of ansiotropy n, for various
ments. This is better illustrated in Fig. 5 depicting values of the H/B ratio
170 G. GAZETAS

I I I I
0 1 2 3 4
Frequency factor, A,

UN
Fig. 6. Undrained dynamic vertical displacement-load ratio (H/B = 1): (a) in-ph ase component; (b) !M”%-out-of-phase
component

0.5

0 1 2 3 4
Frequency factor, A,

Fig. 7. Undrained dynamic vertical displacement-load ratio (H/B = 3):(a) in-phase component; (b) 90x-out-of-phase
component
STRIP FOUNDATIONS ON CROSS-ANISOTROPIC SOIL LAYER 171

1 2 3 4
Frequency factor, A,

Fig. 8. Undrained dynamic vertical displacement-load ratio (H/B = CC):(a) in-phase component; (b) 90x-out-of-phase
component

the horizontal static compliance as a function of the gibly small, especially when compared with the
degree of anisotropy n for several values of the H/B corresponding halfspace displacements. This is due
ratio. With an infinitely thick deposit (i.e. halfspace) to the fact that no surface waves can be created in a
6, tends to infinity as A, tends to zero, in agreement soil stratum at these frequencies; thus no radiation
with classic theory of elasticity (e.g. Poulos & damping is present and the said displacement
Davis, 1974). component reflects only the internal damping in the
The variation of dynamic displacements with soil.
frequency reveals an equally strong dependence on Cross-anisotropy. It is evident from Figs 2 to 4
H/B. On a stratum, both in-phase and 90”-out-of- that, relative to a corresponding isotropic deposit,
phase components of displacement are not smooth the following effects can be ascribed to cross-
and monotonically decreasing functions of fre- anisotropy.
quency, as on a halfspace, but exhibit several peaks As the ratio n of horizontal to vertical modulus
which are the product of resonance phenomena: increases, the static horizontal compliance de-
waves propagating away from the foundation creases-an anticipated phenomenon that is more
reflect at the soil-bedrock interface and return back clearly illustrated in Fig. 5, for several H/B ratios. In
to the surface. As a result, foundation motion the extreme case of n = 4 the medium behaves as
significantly increases at specific frequencies of irrotational, in addition to being incompressible,
vibration which, as shown later, are close to the and consequently no deformation occurs, as was
natural frequencies of the soil deposit. first pointed out by Gibson (1974).
In the low frequency range, below the first The importance of anisotropy increases as the
resonant frequency, as long as bedrock does exist at deposit becomes thicker (Fig. 5). In other words, the
some depth below the surface (i.e. H # m), the 90”- larger H/B is, the faster the static 6, decreases with
out-of-phase component of displacement is negli- increasing n. For H/B between 1 and 6 and n
172 G. GAZETAS

Fig. 9. Undrained static vertical displacement-load ratio 01 I I I 1 1


I 2 3 4 5 6 7 8 9 10
as a function of tbe degree of anisotropy n, for various
values of tbe H/B ratio Overconsolidation ratio, dCldO

Fig. 10. Settlement correction for three-dimensional ef-


fects on initial excess pore pressure distribution (adapted
between 0.50 and 2, the simple expression from Highway Research Board, 1973)
5P” 4.10-n(H/B)O”
6H static = 8E (19)
” -1 f(5/3)(BIH) determine the response of foundations and
for structures to such dynamic loads are well
established and can be found in numerous
l<H/B<6 publications (Richart et al., 1970; Ratay, 1971;
O.SGnG2.5 Yoshimi et al., 1977; American Society of Civil
Engineers, 1979).
fits the numerical data with very good accuracy
(error less than 4%). Vertical loading
The second effect of increasing n ratio is to The vertical settlement-load ratio E,&/P, is
increase the resonant frequencies and decrease the displayed in Figs 68 as a function of A, and n, for
resonant amplitudes for all soil profiles but the three homogeneous soil deposits having H/B = 1,3
halfspace. For each deposit the first resonant and cc, respectively. Adhesive contact is again
frequency can be approximated with reasonable assumed between footing and soil, while 5 = 0.05.
accuracy by the fundamental natural frequency of Several differences in the response of a founda-
the deposit in shear vibration. The latter, obtained tion to vertical and horizontal loads are evident
as for a cantilever shear beam (see, for example, from a comparison of Figs 68 with Figs 24.
Newmark et al., 1974), is given by Vulnerability. For a given soil deposit, founda-
tions are more vulnerable to horizontal than
vertical statically applied loads, since they
experience horizontal displacements larger, by a
and leads to a resonant frequency factor factor of at least 2, than the settlements due to a
7CB vertical load of equal magnitude.
A V,r = 2H(4-n)m’i2 (21) Layer thickness. Static settlements exhibit a
stronger dependence on layer thickness but are less
if account is taken of the constraint relation in sensitive to the degree of soil anisotropy than
undrained conditions. As an example, for H/B = 1 horizontal displacements are. This is better
and n = 4, 1 and 2, equation (21) yields A”., = 082, illustrated in Fig. 9, as compared with Fig. 5. It can
0907 and 1.111, respectively, which compare be seen that foundations on very shallow layers (e.g.
favourably with the values 0.78086 and 1.08 read H/B = 1) settle by an amount which is essentially
from Fig. 3. independent of n. The phenomenon can be attri-
Figures 2-5 and equations (19) and (21) can be buted to the one-dimensional nature of deforma-
utilized in practice to predict the performance of a tions that take place in shallow deposits under the
variety of foundations and structures subjected to central part of relatively large loading areas (i.e. in
static and dynamic horizontal loads. The latter may cases of small H/B ratio). It seems reasonable to
either be applied directly on the structure and then argue that such deformations depend primarily on
transmitted into the ground through the founda- E, and vvu; E, affects only the deformations under
tion (as with machine, wind or sea-wave loads) or the foundation perimeter and its importance dimi-
transmitted from the ground to the structure (as nishes with H/B. Thus n is unimportant in such
with earthquake and traffic-induced loads). Pro- cases.
cedures that employ the compliance functions to The phenomenon is reminiscent of the develop-
STRIP FOUNDATIONS ON CROSS-ANISOTROPIC SOIL LAYER 173

Frequency factor, A,

(b)

Fig. 11. Undrained dynamic angle of rotatiowmoment ratio (H/B = 1): (a) in-phase component; (b) WA-out-of-phase
component

ment of excess pore pressures in clayey strata


6v,s,atic = 4$(4-r1)(~~~)t~‘~’ 1 + 3.5 g (22a)
subjected to axisymmetric loading, as discussed by ” ( >
Skempton & Bjerrum (1957). Figure 10 shows the
dependence on overconsolidation (OCR) and H/B with
ratios of the correction factor which should multi-
ply the one-dimensionally determined consolida- 1<HJB<4
tion settlement in order to compensate for three- 0.5 < n < 2.5
dimensional (3-D) effects on the initial excess pore-
water pressure distribution. It is clear that in fits the numerical data with reasonable accuracy
shallow strata (H/B < 1) the 3-D correction is minor (error within 10%); for deeper deposits the approxi-
and, moreover, almost independent of OCR. mation takes the form
Certainly, under plain-strain loading conditions, as
is the case here, 3-D effects would be even smaller 6V,static =2(4-n) log,,: (22b)
and thus settlement would practically be indepen- ”
dent of the n ratio which, in general, increases with with
increasing OCR (e.g. Gibson, 1974).
HJB>8
For relatively shallow deposits (1 < H/B< 4), the
simple expression all n values
174 G. GAZETAS

(4

_o.5~_~_
0 1 2 3 4
Frequency factor, A,

(b)
Fig. 12. Undrained dynamic angle of rotatiorr-momentratio (H/B = 3): (a) in-phase component; (b) !N’Z-out-of-phase
component

Contrast equation (22b) with the expression for more, introducing the constraint relationship in
the undrained settlement of a uniformly loaded equation (2) (derived by Gibson & Sills, 1975),
circular area given by Hooper (1975) (on the basis of yields for a linearly heterogeneous cross-
results by Gerrard & Harrison, 1970a). Use of the anisotropic halfspace
constraint relationship reduces that expression to
6V,static =%(4-n)
6 = PY(4-n) (23)2
Vsstat’c
2xR-b in which Pv, avis the average vertical pressure on the
footing and r the rate of soil heterogeneity (equa-
with
tion (1)). A similar dependence of undrained settle-
H/B = co ment on the degree of anisotropy n for both
homogeneous and heterogeneous deep soil
all n values
deposits can be inferred from equation (24).
which reveals a similaily strong dependence on Resonance. Resonance phenomena are again
settlement on n for both circular and strip footings observed at one or two frequencies of vibration but
on very deep homogeneous soil deposits. Further- the corresponding peaks are not as sharp as those
of the horizontal dispiacements. In fact, on very
z P, in equation (23) is total applied force on a circular shallow deposits (H/B = 1) a single flat resonance
area, whereas P, in the previous equations is total force per takes place, which is characteristic of a highly
unit length on a strip footing. R = the radius of the damped system.
foundation. A possible explanation of such behaviour stems
STRIP FOUNDATIONS ON CROSS-ANISOTROPIC SOIL LAYER 175

1.5
r

I I I

0 2 3 4
Frequency factor, A,

04

Fig. 13. Undrained dynamic angle of rotation-moment ratio (H/B = co): (a) in-phase component; (b) !Xt%-out-of-phase

from the fact that both compressional and shear moment ratio E, B2 4,/M,.
waves participate in the motion at first resonance The static value of the ratio converges to a finite
which occurs at a frequency factor value (e.g. 0.92, for n = 1) as the depth of the deposit
grows beyond any limit, in contrast with the static
horizontal and vertical displacement-load ratios
that tend to infinity as the stratum tends to become
that lies in between the fundamental natural a halfspace (Figs 4, 8). Increasing layer depth
frequency factors of the deposit in pure shear beyond a value corresponding to H/B = 3 has
(equation (21)) and in pure dilatation (A,,, = CO, practically no effect on rotation. This implies that
due to incompressibility of the material). Surface the stress and strain fields caused by moment loads
waves are also present during resonance, as a result are of limited extent, thus influencing only the near-
of the interference of the two types of waves; the surface soil (Gazetas & Roesset, 1976; Gazetas,
ensuing radiation damping contributes to further 1980a).
limiting the peak amplitudes of motion (Gazetas & One relatively flat resonance takes place at a
Roesset, 1979). frequency factor A,!, which can be approximated
by equation (25). This indicates that an interference
Moment loading between dilatational and shear waves is responsible
The response of a rigid foundation to a harmonic for the observed peak, much in a similar way as with
moment M, eiot is described in Figs 11-14 through vertical vibrations. However, the effect of soil
the normalized dynamic angle-of-rotation- anisotropy is evident in rocking vibrations even at
176 G. GAZETAS

very shallow deposits (e.g. H/E = 1) and small


frequencies of oscillation (e.g. A,< 1.5); this is
hardly the case with vertical motion, as discussed
previously (Fig. 6).

EFFECT OF FRICTIONLESS CONTACT


Under static loading conditions no secondary
stresses develop in the soil-foundation interface if
the material is incompressible. That is, no shear
tractions are generated during vertical and moment
1 2 3 4 loading and no normal tractions during horizontal
n = E,IE, loading. Under dynamic excitation these
conclusions appear to be true for horizontal load-
Fig. 14. Undrained static angle of rotation-moment ratio ing throughout the frequency range examined; the
as a function of the degree of anisotropy n, for various response curves corresponding to the two types of
values of the H/B ratio contact behaviour (i.e. allowing or not allowing

2
Frequency factor, A,

(b)
Fig. 15. Effect of frictionless contact on dynamic angle of rotation-moment ratio (H/B = 1, n = l/3); solid line = adhesive
contact, dotted line = frictionless contact; (a) in-phase component; (b) Wk+,ut-of-phase component
STRIP FOUNDATIONS ON CROSS-ANISOTROPIC SOIL LAYER 177

secondary tractions to develop) are therefore anisotropy greatly influences both static and
essentially identical for all the profiles examined. dynamic foundation displacements under
However, dynamic shear tractions do arise undrained conditions. Soils characterized by large
during resonance in vertical and rocking vibrations n values are likely to experience static settlements as
of footings adhesively connected to the soil surface. much as 50% smaller than what computations
Such tractions are not present under an ideally based on classical isotropic theories indicate; this
smooth (frictionless) footing. As a consequence, may well be the reason for the usual overprediction
the peaks of the frictionless response curves occur of settlements on heavily overconsolidated clays.
at slightly higher frequencies, i.e. they move away On the other hand, soils exhibiting n values smaller
from the fundamental natural frequency of the than unity, like sands and sensitive clays, settle
deposit in shear vibrations. Figure 15 illustrates this more than isotropic soils with the same vertical
observation as it applies to rocking vibrations stiffness.
(Vertical vibration curves for frictionless and Finally, when the dynamic performance of
adhesive contact show similar differences.) Figure foundations is studied, neglecting soil anisotropy
15 also shows the somewhat higher peaks of the may lead to unsafe conclusions regarding the
frictionless curves. Overall the observed discrep- possibility of resonance and the response at high
ancies are rather insignificant and, furthermore, frequencies of oscillation.
decrease with increasing degree of anisotropy n.
ACKNOWLEDGEMENT
SUMMARY AND CONCLUSIONS
The Author wishes to acknowledge partial
The Paper has presented a rigorous solution to
financial support of this work by the US National
the problem of determining the static and dynamic
Science Foundation.
response of rigid strip footings resting on the
surface of a soil deposit consisting of any number of
cross-anisotropic soil layers. Vertical, horizontal REFERENCES
and moment loading has been considered and the American Society of Civil Engineers (1979). Soil-structure
interaction effectsfor nuclear power plants. New York:
results of comprehensive parametric studies have
American Society of Civil Engineers.
been presented in the form of normalized displace- Arthur, J. R. F. & Menzies, B. K. (1972). Inherent
ment-load ratios as functions of dimensionless anisotropy in sand. Gdotechnique 22, No. 1, 115-129.
geometric and material parameters. These results Atkinson, J. H. (1975). Anisotropic elastic deformations in
can be readily used in practice to study the laboratory tests on undisturbed London clay.
performance of foundations and structures in a Geotechniaue 25. No. 2. 357-374.
variety of static and dynamic loading Barden, L. (1963). Stresses and displacements in a cross-
environments. Simple characteristic soil profiles anisotropic soil. Geotechnique 13, No. 3, 198-210.
and undrained soil behaviour have been Barden, L. (1971). Examples of clay structure and its
influence on engineering behaviour. Proceedings of
considered, although the method can treat as easily
Roscoe memorial symposium, Cambridge, 1955205.
any horizontally layered deposit and drained Cambridge: Fouhs.
conditions. Barden, L. (1972). Influence of structure on deformation
Three main factors influence the normalized and failure in clay soil. Geotechnique 22, No. 1,
displacement-load ratios of rigid foundations on a 159-163.
cross-anisotropic soil stratum Bianchini, G. (1980). Effects ofanisotropy and strain on the
dynamic properties ofclays. PhD thesis, Case Western
(a) the stratum depth to foundation halfwidth ratio Reserve University, Cleveland.
HIB Burland, J. B., Broms, B. B. & de Mello, V. F. B. (1977).
(b) the degree ofsoil anisotropy, measured with the Behaviour of foundations and structures. Proc. 9th
ratio n = .E,/E, Inc. Conf Soil Mech., Tokyo, 4955546.
(c) the dimensionless frequency factor A, Carrier, G. F. (1946). Propagation ofwaves in orthotropic
media. Q. Appl. Math. 4, l&165.
The importance of these factors has been demon- Franklin, A. G. & Mattson, P. A. (1972). Directional
strated through a number of parametric plots for all variation of elastic wave velocities in oriented clay.
three types of loading and through extensive Clays Clay Miner. 20, 285-293.
comparisons with pertinent results of other Gassman, F. (1953). Elastic waves through a packing of
researchers for isotropic and anisotropic soils. On spheres. Geophysics 18, 268-273.
Gazetas, G. (1980a). Static and dynamic displacements of
the basis of the presented numerical data, simple
foundations on heterogeneous multilayered soils.
and sufficiently accurate formulas have been Geotechnique 30, No. 2, 1599177.
developed and offered in the Paper for static Gazetas, G. (1980b). Deformational anisotropy of soik
displacements and resonant frequencies in terms of experimental investigation and mathematical modeling.
n and HJB. Research report, Dept. Civil Engineering, Case
The results of these studies suggest that soil Western Reserve University, Cleveland.
178 ‘3. GAZETAS

Gazetas, G. (1981). Indentation of anisotropic halfspace of earthquake engineering. Englewood Cliffs, New
by yielding circular foundation. J. Engng Mech. Div., Jersey: Prentice-Hall.
Am. Sot. Civ. Engrs 107, EM2. Pagano, N. J. & Halpin, J. C. (1968). Influence of end
Gazetas, G. & Roesset, J. M. (1976). Forced vibrations of constraint in the testing of anisotropic bodies. J.
strip footings on layered soils. Meth. Struct. Anal., Am. Compos. Mat. 2, No. 1, 18-31.
Sot. Ciu. Engrs I, 115-131. Parkin, A. K., Gerrard, C. M.&Willoughby, D. R. (1968).
Gazetas, G. & Roesset, J. M. (1978). Plane-strain soil- Deformation of sand in hydrostatic compression
structure interaction for seismic response. Proc. 6th (discussion). J. Geotech. Engng Div., Am. Sot. Ciu.
Symp. Earthq. Engng, Roorkee, India 1, 259-264. Engrs, 94, 336340.
Gazetas, G. & Roesset, J. M. (1979). Vertical vibration of Payton, R. G. (1975). Green’s tensor for a constrained
machine foundations. J. Geotech. Engng Dia., Am. transversely isotropic elastic solid. Q. JI Mech. Appt.
Sot. Ciu. Enars 105. GT12. 1435-1454. Math. 28, 473481.
Gerrard, C. My (1977). Background to mathematical Pickering, D. J. (1970). Anisotropic elastic parameters for
modelling in geomechanics: the roles of fabric and soil. Gkotechnique 20, No. 3, 271-276.
stress history. Finite elements in geomechanics Poulos, H. G. & Davis, E. H. (1974). Elastic solutions for
(Gudehus, G., Ed.), 333120. New York: J. Wiley. soil and rock mechanics. New York: J. Wiley.
Gerrard, C. M. & Harrison, W. J. (1970a). Circular loads Ratay, R. T. (1971). Sliding-rocking vibration of body on
applied to a cross-anisotropic halfspace. Paper 8. elastic medium. J. Geotech. Engng Div., Am. Sot. Ciu.
Commonwealth Scientific and Industrial Research Engrs, 97, SMl, 1777192.
Organization: Div. Appl. Mech. Richart, F. E., Woods, R. D. & Hall, J. R. (1970). Vibrations
Gerrard, C. M. & Harrison, W. J. (1970b) Stresses and ofsoils and foundations. Enalewood Cliffs, New Jersev:
displacements in a loaded orthorombic ha&pace. Prentice-Hall.
Canberra: Commonwealth Scientific and Industrial Rowe, P. W. (1962). The stress-dilatancy relation for static
Research Organization: Div. Appl. Mech. equilibrium of an assembly of particles in contact.
Gibson, R. E. (1974). The analytical method in soil Proc. R. Sot. A269, 50&527.
mechanics. Geotechnique 24, No. 2, 115-140. Saada, A. S. & Bianchini, G. F. (1977). Strength of l-
Gibson, R. E. & Kalsi, G. S. (1974). The surface settlement dimensionally consolidated clays. J. Geotech. Engng
of a linearly inhomogeneous cross-anisotropic elastic Div., Am. Sot. Ciu. Engrs 103, GT6, 660.
halfspace. Z. Angew. Math. Phys. 25, 843-847. Saada, A. S., Bianchini, G. F. & Shook, L. (1978). The
Gibson, R. E. & Sills, G. E. (1975). Settlement of a strip dynamic response of normally consolidated aniso-
load on a non-homogeneous orthotropic incom- tropic clay. Proceedings of conference on earthquake
pressible elastic halfspace. Q. JI. Mech. Appl. Math. 28, engineering and soil dynamics, Pasadena 2, 7777801.
2333243. New York: American Society of Civil Engineers.
Harrison, W. J. & Gerrard, C. M. (1972). Elastic theory Saada, A. S. & Ou, C. D. (1973). Strainstress relations
applied to reinforced earth. J. Geotech. Engng Div., Am and failure ofanisotropicclays. J. Geotech. Engng Div.,
Sot. Engrs 98, SM12, 132551345. Am. Sot. Ciu. Engrs 99, SM12, 1091-1111.
Hearmon, R. F. S. (1961). An introduction to applied Skempton, A. W. & Bjerrum, L. (1957). A contribution to
anisotropic elasticity. Oxford: Oxford University the settlement analysis of foundations on clay. Gko-
Press. technique 7, No. 3, 168-178.
Highway Research Board. (1973). Estimating consolida- Sveklo, V. A. (1970). The action of a stamp on an elastic
tion settlements of shallow foundations on over- anisotropic halfspace. Prikl. Mat. Mekh. 34, No. 1,
consolidated clay. Bull. Highw. Res. Bd. 172-178 (in translation).
Hooper, J. A. (1975). Elastic settlement of circular raft in Uriel, A. 0. & Kaiiizo, L. (1971). On the elastic anisotropv__
adhesive contact with a transversely isotropic of soil. Gtotechnique 21, No. 3, 262-267.
medium. Geotechnique 25, No. 4, 691-711. _.
Valliappan, S., White. W. & Lee. I. K. (1976). ~ I
Enerev
-2

Kausel, E., Roesset, J. M. & Christian, J. T. (1976). absorbing boundary for anisotropic material. Numeri-
Nonlinear behavior in soil-structure interaction, J. cal methods in geomechanics 2,1013-1024. New York:
Geotech. Engng Div., Am. Sot. Ciu. Engrs 102, GT12, American Society of Civil Engineers.
1159-1170. Ward, W. H., Marsland, A. & Samuels, S. G. (1965).
Kirkner, D. J. (1979). Steady-state response of a circtdar Properties of the London Clay at the Ashford Com-
foundation on a transversely isotropic half-space. PhD mon Shaft. Geotechnique 15, No. 4, 321-344.
thesis, Case Western Reserve University, Cleveland. Ward, W. H., Samuels, S. G. & Gutler, M. E. (1959).
Kirkpatric, W. M. & Rennie, I. A. (1972). Directional Further studies of the properties of London Clay.
properties of a consolidated caolin. Geotechnique 22, Geotechnique 9, No. 2, 33-58.
No. 1, 166169. Westergaard, H. M. (1938). A problem of elasticity
Lekhnitskii, S. G. (1963). Theory of elasticity of an suggested by a problem in soil mechanics. Contribu-
anisotropic elastic body (in translation). San Francisco: tions tomechanics ofsolids. Timoshenko 60th Anniver-
Holden-Day. sary Volume. New York: Macmillan.
Michell, J. H. (1900). The stress in an aeolotropic elastic Willoughby, D. R. (1967). The geometrical description of
solid with an infinite plane boundary. Proc. Land. linear and planar features in soils. Aust. J. Soil Res. 5,
Math. Sot. 32, 247-258. 21-36.
Milovic, D. M. (1972). Stresses and displacements in an Wolf, K. (1935). Ausbreitungder Kraft in der Halbene und
anisotropic layer due to a rigid circular foundation. im Halbraum bein anisotropen Material. Z. Angew.
Geotechnique 22, No. 1, 169-174. Math. Mech. 15, 249-254.
Newmark, N. M. & Rosenblueth, E. (1971). Fundamentals Wroth, C. P. (1971). Some aspects of the elastic behaviour
STRIP FOUNDATIONS ON CROSS-ANISOTROPIC SOIL LAYER 179

of overconsolidated clay. Proceedings of Roscoe mem- which, upon substitution of the drained values, yields
orial symposium, Cambridge. Cambridge: Foulis. n = 1.633. Assuming similar deformational behaviour of
Yong, R. N. & Silvestri, V. (1979). Anisotropic behaviour the soils at Barbican Arts Centre and at Ashford, the
of a sensitive clay. Can. Geotech. J. 16, 335-350. undrained value of G,,/E, can be estimated as equal to
Yoshimi, Y., Prakash, S., Barkan, D. D., Ilyichev, V. A. & 0.372 (Table 1). Furthermore, the condition that the shear
Richart, F. E. (1977). Soil dynamics and its applica- moduli GVH and Gnu remain constant during consolida-
tions to foundation engineering. Proc. 9th Int. Conf tion leads to a ratio of drained and undrained horizontal
Soil Mech., Tokyo, 613-650. moduli (see Hooper, 1975).

E” 4-n
APPENDIX 1 p= = 1.18
DRAINED GVHIEV OF LONDON CLAY E”’ 2(lf VW’)

Drained laboratory triaxial tests carried out by The vertical moduli ratio is then
Atkinson (1975) indicated that n’ = 2, vVH’ = 0.19 and E, E, n’ 2
vuu’ = OQO, where the primes denote drained parameters. -= - -e 1.18 x --= 1445
E,’ ER’ n 1.633
For a saturated poro-elastic soil, Uriel & Kafiizo (1971)
give the following relationship for undrained n and consequently

2n’( 1 - vu” - 2n’v,,‘*) G ’ = y-E’


GVH Ev
n- $ = 0.372~ 1445=0.538
1- 2n’v,,’ + n’ -n” vVH’*- 2n’vVu’vrm - vnn’* V v V

which is the value shown in Table 1.

S-ar putea să vă placă și