Journal of the Chinese Institute of Engineers, Vol. 24, No. 6, pp.

659-671 (2001) 659

CAVITY EXPANSION AND CONE PENETRATION RESISTANCE

IN ANISOTROPIC CLAY

Shi-Fon Su
Department of Civil Engineering
Tung-Nan Institute of Technology
Taipei, Taiwan 222, R.O.C.

Hung-Jiun Liao*
Department of Construction Engineering
National Taiwan University of Science and Technology
Taipei, Taiwan 106, R.O.C.

Key Words: spherical cavity expansion, cone factor, strength

anisotropy, and rigidity index.

ABSTRACT
Due to strength anisotropy, the undrained shear strength of natu-
rally deposited clay can vary with its position in the coordinate. When
subjected to a spherical expansion, the undrained shear strength of clay
yields the lowest value in the horizontal direction and the largest value
in the vertical direction. To represent the nature of clay, a model based
on the anisotropic undrained strength criterion has been developed in
this paper to determine the anisotropic undrained shear strength under
a spherical expansion condition. To simplify the model, the effects of
in-situ initial stresses, anisotropy, and the anisotropy of soil
deformability are not taken into account here. Combining the spheri-
cal cavity expansion theory and the anisotropic undrained shear strength
under spherical expansion conditions, a simplified limit pressure of
spherical cavity expansion in anisotropic clay has been established.
The magnitude of limit pressure is related to its position in the spheri-
cal cavity. Based on the limit pressure in spherical cavity expansion
derived from anisotropic strength, the cone factor of an advancing cone
has been derived to correlate the cone resistance and undrained shear
strength of clay. The calculated cone factors uniquely correspond to
the undrained shear strengths determined from different tests and are
interrelated to each other in terms of strength anisotropy ratio A r of
clay. The accuracy of the calculated cone factors has been satisfacto-
rily verified with laboratory and field CPTU results. However, it has
also been found that the effect of strength anisotropy only becomes
obvious when clay has a low rigidity and high strength anisotropy. If
the strength anisotropy of clay is not considered, a less than 15% error
on the value of the cone factor will result.

I. INTRODUCTION tests, determining the mechanical properties of natu-

rally deposited soft clay from cone resistance has been
Among many applications of cone penetration widely used in geotechnical practice. For example,

*Correspondence addressee
660 Journal of the Chinese Institute of Engineers, Vol. 24, No. 6 (2001)

the undrained shear strength can be correlated to the

cone resistance either empirically as proposed by Aas
et al. (1986) or theoretically by the following
methods: (1) bearing capacity theory (Durgunoglu and
Mitchell, 1975), (2) cavity expansion theory [Vesic
(1977); Yu and Houlsby (1991); and Salgado (1993)],
(3) steady state deformation method [Baligh (1985);
Teh and Houlsby (1991)], and (4) incremental finite-
element analysis (Van den Berg, 1994). Sometimes,
the results from the calibration chamber test of the
cone can also be used to obtain the undrained shear
strength. Among the above mentioned methods, the
cavity expansion theory was considered to be sim-
pler and more appropriate than others for analyzing
the resistance to an advancing cone (Yu and Mitchell,
1998).
However, natural clays usually show some in-
herent anisotropy due to the shape of clay particles
Fig. 1 Predominant failure modes around an advancing cone
and the process of deposition. In addition, soil ele- (Keaveny & Mitchell, 1986)
ments around an advancing cone are subjected to prin-
cipal stress rotation (Baligh, 1984). So the effect of
stress-induced anisotropy on cone resistance should
be taken into account also. But not much attention horizontal tangential direction. The soil located in
has been paid to the influence of anisotropy on the between these two zones is the transition zone. The
cone resistance of an advancing cone in clay so far direction of major principal stress rotates from the
(Meigh, 1987). vertical direction (beneath the cone) to the horizon-
This paper will propose a model for investigat- tal direction (around the shaft).
ing the effect of strength anisotropy on cone resis- To obtain the undrained shear strength of a
tance based on the anisotropic strength criteria pro- clayey soil from the cone penetration test, Eq. (1) is
posed by Su et al. (1998) and Prevost (1979). In or- commonly used in practice due to its simplicity.
der to do so, the undrained shear strength under
spherical expansion conditions will be established q t – σ vo
su = (1)
first and the limit pressure in spherical cavity expan- N kt
sion derived from anisotropic strength criteria will
be developed also. This limit pressure will be used where su is the undrained shear strength of clayey soil,
to evaluate the effect of strength anisotropy on the q t is the cone tip resistance corrected for pore pres-
cone resistance of an advancing cone. To verify the sure effect (Campanella et al., 1983), Nkt is the cone
suitability of the proposed model, the field piezocone factor, and σ vo is the vertical overburden pressure.
test results (Nash et al., 1992) and CK oU triaxial tests Since the stress paths to failure vary with the loca-
results (Hight et al., 1992) of the Bothkennar clay in tions around the cone tip, the degree of mobilized
Scotland as well as the field piezocone test results undrained shear strength is different also. So, the
reported by Aas et al. (1986) are compared. undrained shear strength obtained from Eq. (1) can
only be treated as an apparent undrained shear
II. MOBILIZED UNDRAINED SHEAR strength.
STRENGTH AROUND AN ADVANCING CONE In addition, it is not clearly defined what kind of
undrained shear strength can be obtained from
As indicated by Keaveny and Mitchell (1986), Eq. (1). Sometimes it stands for the shear strength
the stress states of soil around an advancing cone can obtained from the field vane (FV) shear test; some-
be simplified into three zones as shown in Fig. 1. Soil times it means the shear strengths determined from
in the zone beneath the cone is subjected to axial the triaxial UU, CIUC, or CKoUC test. But the range
compression. Its major principal stress (σ 1) is verti- of shear strength variation obtained from different test
cal and its minor principal stress ( σ 3) is horizontal. methods can be rather significant for naturally de-
For soil in the zone around the shaft, the stress state posited clays (Fig. 2). For example, the cone factor
is close to that of a laterally expanding cavity. Its used to calculate the shear strength of the field vane
major principal stress ( σ 1 ) is in the horizontal radial shear test is higher than that needed to calculate the
direction and its minor principal stress ( σ 3) is in the shear strength of triaxial CK o UC test for clay with
 S.F. Su and H.J. Liao: Cavity Expansion and Cone Penetration Resistance in Anisotropic Clay 661

Fig. 3 Schematic diagram of spherical expansion and soil par-

Fig. 2 Normalized undrained shear strengths from laboratory tests ticles orientation
and field vane test (Mesri, 1989)

low plasticity (PI<60%). But, the cone factor for the

FV test becomes smaller than that of the CKoUC test
for clay with high plasticity (PI>60%). In other
words, if Eq. (1) is to be used to determine the shear
strength, the cone factor Nkt should vary with the types
of undrained shear strength to be determined.

III. ANISOTROPIC UNDRAINED SHEAR

STRENGTH UNDER SPHERICAL
EXPANSION CONDITIONS

For the cavity expansion theories used in incom-

pressible material (such as saturated clay under
undrained conditions), Tresca and von Mises
undrained strength criteria are most commonly used. Fig. 4 Comparison of different undrained strength criteria on
The undrained shear strengths under spherical expan- deviatoric plane
sion conditions obtained from the above criteria are
the same in value and are irrelevant to their positions
on the spherical coordinates. In other words, if an
isotropic strength criterion (i.e., the undrained represent the center and the size of the failure func-
strength used in the analysis is constant in value) is tion in the stress space respectively; Ar is the strength
adopted, the undrained shear strength to resist the anisotropy ratio (=s ue/s uc); s uc is the undrained shear
spherical expansion in the vertical direction will be strength obtained from CK oUC triaxial tests, and s ue
the same as the one to resist the spherical expansion is the undrained shear strength obtained from CKoUE
in the horizontal direction. However, due to the ori- triaxial tests. When plotted on the deviatoric plane,
entation of naturally deposited clay particles, the iso- the failure function of Eq. (2) is elliptical and is not
tropic strength criterion is generally not suitable for symmetrical with respect to the origin (i.e., the hy-
spherical expansion in natural clay (Fig. 3). To deal drostatic axis) (Fig. 4). In comparison, the Prevost’s
with the problem associated with spherical expansion failure function is circular in shape but not symmetri-
in anisotropic clay, the undrained anisotropic strength cal to the origin; the failure function of von Mises is
criterion proposed by Su et al. (1998) is adopted in circular in shape and is symmetrical to the origin.
this study (Eq. (2)): When the major principal stress at failure is not in
the vertical direction, the difference between the von
F = 1 {A r(σ x – σ y)2 + [(σ y – σ z) + α]2 + [(σ z – σ x) – α]2} Mises criterion and the anisotropic strength criteria
6
will become obvious.
2A r + 1 2 For spherical cavity expansion in an isotropic
+ τ xy + A –r 1/4(τ yz
2 + τ2 )– K2 = 0
zx (2) medium, the responded radial stress σ r is the major
3
principal stress σ 1 ; while the tangential stress σ θ and
where F is the undrained failure function; α and K the hoop stress σ φ are the minor principal stress σ 3.
662 Journal of the Chinese Institute of Engineers, Vol. 24, No. 6 (2001)

The stress state at failure should meet the require-

ment of the strength criterion shown in Eq. (2). So,
by substituting the stress components of Eq. (4) into
Eq. (2), the following equation is obtained:

[(1+A r)sin 4θ +2cos 4θ +(6 A r −2)sin 2θ cos 2θ ]( σ r− σ θ) 2

– 1/4

+[2α (sin 2θ −2cos 2θ )](σ r− σ θ )+{2 α2−6K 2}=0 (5)

Let the undrained shear strength obtained from the

spherical expansion (s u(sc)) be equal to the maximum
shear stress at failure (=0.5(σ r− σ θ )f). Then the ratio
between s u(sc) and s uc can be expressed as

– α b(θ ) + [ α b(θ )]2 – a(θ )[2 α – 6 K ]

2 2
Fig. 5 Spherical cavity expansion in an infinite medium and its s u(sc)/s uc = J(θ ) =
coordinate system 2a(θ )
(6)

In other words, σ 1= σ r and σ 3 = σ θ = σ φ . To simplify a(θ )=(1+A r)sin 4θ +2cos 4θ +(6 A r

– 1/4
−2)sin 2θ cos 2 θ
the problem, it is assumed that σ r > σ θ = σ φ and
τrθ=τrφ=τθφ=0 in this paper. However, it is understood (7)
that the stresses acting on an expanding sphere in
anisotropic material can be much more complicated b(θ )=sin 2θ −2cos 2 θ (8)
because of the changes in material yielding properties.
The stress symbols used in the strength crite- α = 1 – Ar (9)
rion are defined in the Cartesian coordinate system
and the z-direction is taken as the gravity direction. 2 (1 + A r)2
K = (10)
So the stress components of the spherical expanding 3
pressure are transferred to Cartesian coordinates as
Similarly, if Prevost’s (1979) undrained strength cri-
follows (refer to Fig. 5):
terion is adopted, then the ratio of s u(sc) /s uc can be
written as
sinθ cosφ cosθ cosφ – sinφ σr 0 0
– α b(θ ) + [ α b(θ )]2 – 2[2 α 2 – 6 K ]
2
sinθ sinφ cosθ sinφ cosφ 0 σθ 0
s u(sc)/s uc =
cosθ – sinθ 0 0 0 σφ 4
(11)

sinθ cosφ sinθ sinφ cosθ In fact, when a( θ )=2, the undrained shear strength of
• cosθ cosφ cosθ sinφ – sinθ the expanding sphere shown in Eq. (6) is equal to the
– sinφ cosφ 0 undrained shear strength (Eq. (11)) determined from
the Prevost’s strength criterion. The undrained shear
strengths of spherical expansion obtained from both
σ x τ xy τ xz methods are compared in Fig. 6 and the following
statements can be made:
= τ yx σ y τ yz (3)
(1) The undrained shear strength of sphere expansion
τ zx τ zy σ z (s u(sc)) decreases with increasing position angle θ
and decreasing strength anisotropy ratio A r. s u(sc)
Eq. (3) can be rearranged as follows: in the horizontal direction ( θ =90° and the major

sin2θ cos2φ(σ r – σ θ) + σ θ sin2θ sinφcosφ(σ r – σ θ) sinθ cosθ cosφ(σ r – σ θ) σ x τ xy τ xz

sin θ sinφcosφ(σ r – σ θ)
2
sin θ sin φ(σ r – σ θ) + σ θ
2 2
sinθ cosθ sinφ(σ r – σ θ) = τ yx σ y τ yz (4)
sinθ cosθ cosφ(σ r – σ θ) sinθ cosθ sinφ(σ r – σ θ) cos2θ (σ r – σ θ) + σ θ τ zx τ zy σ z
 S.F. Su and H.J. Liao: Cavity Expansion and Cone Penetration Resistance in Anisotropic Clay 663

Table 1 Comparison of calculated and laboratory

measured undrained shear strengths for
true triaxial test (data from Liao and Su,
1999)
Normalized undrained
shear strength (s u/ σ ' p)
Shearing mode
Calc.
Lab.
Eq. 6 Eq. 11
Axial compression 0.320
− −
Axial extension 0.210
Horizontal expansion
0.264 0.254 0.233
Fig. 6 Comparison of undrained shear strengths under spherical ( σ y> σ z= σ x)
expansion condition obtained from different strength cri-
teria

von Mises criterion. Since the undrained shear

principal stress is in the horizontal direction) has strengths determined from these theories are irrel-
the lowest value. But it is still larger than s ue, the evant to positions on the sphere, the limit pressure in
undrained shear strength obtained from the triaxial spherical cavity expansion derived from isotropic
CK oUE test. strength is irrelevant to the position also.
(2) When θ is smaller than 55°, s u(sc) determined from To study the existing cavity expansion theories,
Prevost’s criterion is larger than that obtained from Houlsby and Withers’ theory is adopted here as an
Su’s criterion; when θ is larger than 55°, s u(sc) de- example. The initial in-situ stresses are assumed to
termined from the Prevost’s criterion is smaller be isotropic, i.e., the vertical stress σ vo and horizon-
than that obtained from Su’s criterion. tal stress σ ho resulting from the soil overburden are
(3) Only when the strength anisotropy of clay is not assumed to be equal. In addition, the undrained yield-
significant (Ar>0.8), is s u(sc) close to the undrained ing behavior of clay is assumed to satisfy the Tresca
shear strength obtained from the Tresca or von criterion. Combining the spherical symmetrical equi-
Mises criterion. librium equation for the plastic zone (Eq. (12)) and
To verify the undrained shear strength s u(sc) deter- the boundary condition between the elastic zone and
mined from Prevost’s criterion or Su’s criterion, a true the plastic zone (Eq. (13)) (refer to Fig. 5), the limit
triaxial test was carried out on remolded Taipei silty pressure in cavity expansion can be derived (Eq. (14))
clay. The soil specimens were consolidated under as follows:
K o condition. So, the initial stress condition acting
on the specimen was σz> σy=σx. During the undrained d σ r 4s u
+ r =0 (12)
loading stage, one horizontal stress, σ x, was increased dr
to an amount of σ z first; and then the other horizontal
4s u
stress, σ y , was increased to make the specimen fail. σr = σo + , when r=r p (13)
3
In other words, the specimen failed under a stress con-
dition of σ y> σ z=σ x. It was equivalent to the horizon-
ψ s, I = σ o + 4 s u [1 + ln(I r)] (14)
tal limit pressure in spherical cavity expansion (i.e., 3
θ = 90°). The calculated shear strengths are com-
pared with the laboratory results and shown in Table where r p is the radius of the plastic zone, ψs, I = limit
1. In general, the shear strength calculated from Su’s pressure in spherical cavity expansion derived from
criterion (Eq. (6)) can better predict the undrained isotropic strength, σ o = initial in-situ stress (= verti-
shear strength determined from the laboratory test. cal overburden pressure σ vo or horizontal pressure
σ ho ), I r = rigidity index (=G/s u , G is the undrained
IV. SIMPLIFIED SPHERICAL EXPANSION shear modulus).
THEORY IN ANISOTROPIC CLAY However, it is understood that σ vo is not equal
to σ ho for in-situ clay. So, Eq. (14) proposed by
Many cavity expansion theories for studying the Houlsby and Withers (1988) is actually not suitable
undrained behavior of saturated clay have been pro- for natural clay. If the effect of strength anisotropy
posed (Vesic, 1972; Houlsby and Withers, 1988). is taken into account, it will further complicate the
Most of them are based on either the Tresca or the problem. To deal with the strength anisotropy of
664 Journal of the Chinese Institute of Engineers, Vol. 24, No. 6 (2001)

Fig. 7 Comparison of limit pressure in spherical cavity expan-

sion in isotropic and anisotropic medium

Fig. 8 Pile tip failure mechanism (Yasufuku & Hyde, 1995)

natural clay and to simplify the problem, the follow-
ing assumptions are made in this study:
(1) The anisotropy of initial in-situ stresses is
may seem reasonable, the octahedral strain around the
neglected.
tip of the cone can be as high as 100% (Kiousis et
(2) The stresses σθ and σ φ within plastic zone are as-
al., 1988). It indicates that soil around the tip of an
sumed to be equal.
advancing cone is fractured and pushed away by the
(3) Shear stress τrθ within the plastic zone is relatively
cone. As a result, the stress-stain relationship of soil
small and is neglected.
is difficult to quantify. Thus, it is not easy to obtain
The first two assumptions follow those made by
the cone resistance by integrating the stresses on the
Houlsby and Withers (1988). In other words, the
surface of cone. To avoid this problem, the cone re-
equilibrium equation shown in Eq. (12) is still valid
sistance will be calculated alternatively in this study
when deriving the limit pressure for the expansion of
by the following steps:
a spherical cavity in anisotropic clay. So, by substi-
(1) establishing the relationship between cone resis-
tuting the undrained shear strength of an expanding
tance and limit pressure in spherical cavity
sphere su(sc) (Eq. (6)) to replace the s u of Eq. (14), the
expansion,
limit pressure for the expansion of a spherical cavity
(2) determining the cone factor for the isotropic
in anisotropic clay (ψ s, A) can be approximated as:
undrained shear strength,
(3) determining the cone factor for the anisotropic
ψ s, A = σ o + 4 s u(sc)[1 + ln(I r)] (15) undrained shear strength and verifying it with the
3
field test results,
The limit pressures for the expansion of a spherical (4) comparing the difference between the cone fac-
cavity in isotropic and anisotropic clays are compared tors determined from isotropic and anisotropic
in Fig. 7. The larger the θ angle, the smaller the undrained shear strengths.
ψ s, A. When θ =90° (i.e., the horizontal direction), When studying the pile tip bearing capacity,
the ψ s, A reaches its lowest value. It indicates that Yasufuku and Hyde (1995) assumed that the relation-
clay is weaker in the horizontal direction. Thus, it is ship among the pile tip bearing pressure, active earth
easier for the sphere to expand horizontally. pressure, and cavity expansion pressure can be rep-
resented by the failure mechanism shown in Fig. 8.
V. CONE RESISTANCE IN ANISOTROPIC Since the failure mechanism beneath the pile tip is
CLAY very similar to that beneath a cone tip, the failure
mechanism proposed by Yasufuku and Hyde (1995)
At present, only limited efforts have been made is adopted in this study for convenience. However,
to study the effect of strength anisotropy on cone some modification is still needed because soil around
resistance. The cone resistance of an advancing cone the cone shaft is subjected to axial symmetrical cav-
is usually calculated by integrating the stresses act- ity expansion and the spherical expansion around the
ing on the surface of the cone. Although this approach cone tip will not extend to the shaft. So, the Yasufuku
 S.F. Su and H.J. Liao: Cavity Expansion and Cone Penetration Resistance in Anisotropic Clay 665

Table 2 Comparison of cone factors derived using different methods

Large Strain
Cavity Expansion Theory Strain Path Method
Finite-Element Method
Ladanyi
Teh and Teh and Van den Van den
and Vesic Baligh Yu Yu
Houlsby Houlsby Berg Berg
Proposed Johnston (1977): (1975): (1993): (1993):
Ir (1991): (1991): (1994): (1994):
method (1974): rough rogh smooth rough
smooth rough smooth rough
rough cone cone cone cone
cone cone cone cone
cone
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11)
50 12.5 8.3 9.1 15.9 8.5 13.8 8.3 10.6 9.4 13.8
100 13.5 9.2 10.0 16.6 9.3 14.6 9.7 11.9 11.5 15.3
150 14.0 9.7 10.6 17.0 9.8 15.0 10.6 12.8 12.0 15.8
200 14.4 10.1 10.9 17.3 10.1 15.4 11.3 13.5 12.4 16.3
300 14.9 10.6 11.5 17.7 10.6 15.8 12.5 14.7 12.6 16.9
400 15.3 11.0 11.9 18.0 10.9 16.2 13.5 15.7 12.8 17.2
500 15.6 11.3 12.2 18.2 11.2 16.4 14.4 16.6 13.0 17.5

and Hyde (1995) failure mechanism is modified to

the one shown in Fig. 9.
To justify the suitability of the mechanism
shown in Fig. 9, the relationship between cone resis-
tance and limit pressure in spherical cavity expan-
sion derived from isotropic strength ( ψ s,I ) is estab-
lished based on the Tresca strength criterion. As
shown in Fig. 9, the following equation can be ob-
tained from the force equilibrium in the vertical
direction:

( π d 2)q t +(3 π d2)( ψ s,I −2s u )−(4 πd 2) ψ s,I =0 (16)

where d = radius of cone, (3 π d 2)( ψ s,I −2s u ) is the re-

sultant force acting on surface B. The stress acting
on surface B is the tangential stress of the spherical
expansion and is assumed to be uniformly distributed
(note: The tangential stress should become larger Fig. 9 Assumed relationship between cone tip resistance and limit
when it gets closer to the cone. The assumption made pressure in spherical cavity expansion
here is for simplicity). (4πd2)ψs,I is the resultant force
acting on surface C. The stress acting on surface C
factors listed in Table 2 are derived from the cavity
is the radial stress generated by a spherical cavity
expansion theory, the strain path method, and the
expansion. In addition, the initial stress state for soil
large strain finite element method respectively. The
underneath the cone tip (Fig. 9) is close to the verti-
proposed cone factor N kt,I is slightly lower than the
cal overburden pressure σ vo. So, the σ o of Eq. (14)
cone factor for a cone with rough surface, i.e., the
(= ψ s,I ) can be replaced with σ vo and the following
shear stress acting on the cone surface is equal to
equation can be obtained:
undrained shear strength of soil. When compared with
q t – σ vo Teh and Houlsby’s (1991) cone factor (for rough
N kt, I = su = 6 + 4 [1 + ln(I r)] (17) surface), N kt,I is slightly greater when I r<300 (= me-
3
dium to low rigidity index) but slightly lower when
where N kt,I is the cone factor for the isotropic Ir>300 (= high rigidity index). Furthermore, the pro-
undrained shear strength. The cone factors calculated posed cone factor is lower than Van den Berg’s (1994)
from Eq. (17) are compared with those obtained cone factor (for rough surface) but larger than Van
from other methods and shown in Table 2. The cone den Berg’s cone factor (for smooth surface). In
666 Journal of the Chinese Institute of Engineers, Vol. 24, No. 6 (2001)

summary, the proposed cone factors for the isotropic

undrained shear strength (N kt,I) calculated from Eq.
(17) are close to the average of the cone factors (for
rough surface) proposed by Yu (1993), Teh and
Houlsby (1991), and Van den Berg (1994). It veri-
fies the suitability of the proposed cone factor to de-
termine the cone penetration resistance of the CPTU
test.
Similarly, the equation used to determine the
limit pressure in spherical cavity expansion derived
from anisotropic strength (ψs, A) can be obtained from
the force equilibrium in the vertical direction (Fig.
9):
Fig. 10 Relationship between Jθ and A r of clay
(π d 2)q t + (3π d 2)(ψ s, A – 2s u(sc))
θ = 60°

60°
– 10.667π d 2 ψ s, Acosθ sinθ d θ = 0 (18) 1992 and Hight et al., 1992). The relationship be-
0
tween corrected cone resistance (qt) and depth can be
established from the piezocone test results and ex-
where (3 π d2 )(ψ s, A−2s u(sc) )|θ =60° is the resultant force
60° pressed as
acting on surface B and 10.667 π d 2 ψ s, A cos θ sin θ
0
. d θ is the resultant force acting on surface C. Re- qt (kPa)=129+50.9 depth(m) (21)
place the σ o of Eq. (15) with vertical overburden pres-
The rigidity index Ir determined from the self-boring
sure σ vo, then Eq. (18) can be written as:
pressuremeter test (Nash et al., 1992) ranges between
q t – σ vo 100 and 200. Since the rigidity index is related to
N kt, A = = 6J(60°) + 4 [1 + ln(I r)] the shear modulus of soil, it tends to have a wide range
S uc 3
of values. However, the cone factor is not sensitive
⋅ [10.667J θ – 3J(60°)] (19) to I r as shown in Table 2. When I r increases from
100 to 200, it merely results a 10% increase in cone
60° factor. So, for the Bothkennar clay with only a 30%
Jθ = J(θ )cosθ sinθ d θ (20) variation in I r value, an I r value of 150 is chosen to
0
calculate the cone factor Nkt, A and not much error will
Nkt, A is the cone factor for the anisotropic undrained be expected. The normalized undrained shear
shear strength. J(60°) of Eq. (19) stands for the value strengths of CKoUC and CKoUE tests vary from 0.45
of J( θ ) with θ =60° (refer to Eq. (6)). The J θ value to 0.58 and from 0.20 to 0.25 respectively (Hight et
that corresponds to different strength anisotropy ra- al., 1992). So, the strength anisotropy ratio A r of
tios A r is shown in Fig. 10. When A r =1 (i.e., no Bothkennar clay is between 0.38 and 0.49 with an
strength anisotropy), J θ =0.375 and J(60°)=1. In this average of 0.44. Substitute the above soil parameters
case, Eq. (19) is equal to Eq. (17). into Eq. (19), a cone factor Nkt, A of 13.12 for aniso-
tropic undrained shear strength of Bothkennar clay
VI. CASE STUDIES can be calculated. The calculated undrained shear
strengths of the CKoUC test based on Nkt, A=13.12 are
Soil parameters needed for determining the cone plotted versus depth in Fig. 11. They are compared
factor N kt, A include the rigidity index I r and the with the undrained shear strengths of specimens
undrained shear strengths obtained from CK o U tests sampled with two different kinds of sampler (Laval
(s ue and s uc). To verify the suitability of Eq. (19) on sampler and piston sampler). In general, the calcu-
the in-situ clay, the piezocone test results reported lated undrained shear strengths are in good agreement
by Nash et al. (1992) and the CK oU triaxial test re- with the laboratory test results. The calculated
sults reported by Hight et al . (1992) on the strength is only slightly lower than the test results
Bothkennar clay in Scotland are compared. In for soil located above 2.5 m and is slightly higher
addition, the cone factors for nine different clays sum- than the test results for soil below 10 m deep.
marized by Aas et al. (1986) are compared also. The cone factor Nkt, A for anisotropic undrained
Bothkennar clay is slightly over-consolidated shear strength derived in the study is to calculate the
with an average plasiticity index of 40% (Nash et al., undrained shear strength obtained from the CK o UC
 S.F. Su and H.J. Liao: Cavity Expansion and Cone Penetration Resistance in Anisotropic Clay 667

Table 3 Comparison of calculated and laboratory measured undrained shear strengths for direct simple
test
Triaxial test Direct simple shear test

s uc s ue τ h/ σ ′p References
Types of clay
σ ′p σ ′p Lab. Calc.
Bangkok 0.720 0.370 0.550 0.522 Prevost (1979)
Drammen 0.390 0.200 0.320 0.282 Prevost (1979)
Boston 0.330 0.155 0.210 0.229 Ladd and Edgers (1972)
AGS 0.315 0.195 0.255 0.250 Koutsoftas and Ladd (1985)
Bothkennar 0.380 0.190 0.310 0.272 Mesri (1993)
Taipei 0.302 0.197 0.246 0.250
Note: Normalized undrained shear strength determined from the laboratory test taken at peak shear stress

following study.
Let the failure function (F) on the stress space
(Eq. (2)) be equal to the yielding function (f) under
critical state conditions; and the associated flow rule
is adopted. The total strain increment (dεij) at failure
is about equal to the plastic strain increment (d ε pij ) if
the elastic contribution to the total strain increment
is neglected (Prevost, 1979). Thus, the following
equation can be obtained for the critical state
condition:

d ε ij = d λ ∂f (22)
∂σ ij

where d λ is a proportionally scalar function of the

stress-strain history of the soil. For a direct simple
shear test (d ε x=d ε y=d ε z = 0), the following relation-
ships can be obtained by combining Eqs. (2) and (22):
Fig. 11 Comparison of calculated and measured CK o UC
undrained shear strength (s uc) of Bothkennar clay τ zx = 1 A 1/8
r (1 + A r)s uc (23)
3

test. Its value ranges between 10.5 and 15.6 for most σ z− σ x= α (24)
naturally deposited clays of which the upper and lower
bound values of Ir are equal to 500 and 30; the strength σ y= σ x (25)
anisotropy ratios A r are equal to 0.4 and 0.9.
However, the variation range of cone factor tends to The direct simple shear test tends to underesti-
vary with the type of test conducted for the undrained mate the shear stress τzx by about 10 % (Roscoe, 1953,
shear strength. For example, Aas et al. (1986) re- and Prevost, 1979). So, the direct simple shear test
ported a range of 11~20.5 for cone factors correspond- undrained shear strength (s ud) can be obtained from
ing to the average undrained shear strengths (i.e., the Eq. (23):
average of undrained shear strengths obtained from
CKoUC, CKoUE, and direct simple shear tests) of nine s ud 1 1/8 1/8
different clays with the OCR values varying from 1 s uc = 0.9 3 A r (1 + A r) = 0.52A r (1 + A r) (26)
to 8. But such a wide value range can not be covered
by current analytical methods. To further verify the The suitability of Eq. (26) is verified in Table 3. The
suitability of the proposed cone factor, the test data maximum difference between measured and calcu-
presented by Aas et al. (1986) will be adopted for the lated s ud is only 12%.
668 Journal of the Chinese Institute of Engineers, Vol. 24, No. 6 (2001)

*
Fig. 12 Range of N kt, A obtained from Su et al.’s anisotropic Fig. 14 Comparison of Nkt,I and N kt,A at different strength anisot-
undrained strength criterion ropy ratios

*
limit pressure in spherical cavity expansion, N kt, A
varies from 12.3 to 23.3 (Fig. 13).
In summary, anisotropic strength criteria such
as Su’s (1998) and Prevost’s (1979) have demon-
strated their versatility in generating cone factors for
undrained shear strengths that are determined from
different tests. These cone factors uniquely corre-
spond to different undrained shear strengths and are
interrelated to each other in terms of strength anisot-
ropy ratio A r of clay (Eq. (28)). It is not possible to
do this if an isotropic strength criterion is used to gen-
erate the cone factor.
*
Fig. 13 Range of N kt, A obtained from Prevost’s anisotropic VII. EFFECT OF STRENGTH ANISOTROPY
undrained strength criterion
ON CONE FACTOR

By replacing the s u of Eq. (17) with the s uc of

So, the average undrained shear strength for the CK oUC test, the isotropic cone factor Nkt, I which
three different tests (CK o UC test, CK o UE test, and corresponds to the s uc of the CK o UC test can be
direct simple shear test) can be expressed as obtained. The effect of strength anisotropy on cone
factor can be evaluated by comparing this N kt, I with
(1 + A r)(1 + 0.52A 1/8
r ) Nkt, A (Fig. 14). The smaller the A r and I r, the larger
s u, lab = s uc (27)
3 the difference between Nkt, I and N kt, A. However, for
m o s t n a t u r a l l y d e p o s i t e d c l a y s (A r = 0 . 4 ~ 0 . 9 ,
Combining Eqs. (19) and (27), the cone factor N *kt, A Ir=30~500), the difference is less than 15%. Thus, it
corresponding to the average laboratory undrained can be concluded that even if the effect of strength
shear strength is expressed as anisotropy is neglected, the error resulting, on the
calculated cone factor, is rather limited.
q t – σ vo 3
N *kt, A = = N kt, A (28)
s u, lab (1 + A r)(1 + 0.52A 1/8 VIII. CONCLUSIONS
r )

For most soils, the rigidity index I r ranges between By ignoring the anisotropic effects of initial in-
30 and 500; the strength anisotropy ratio A r ranges situ stresses and soil deformation modulus, a cone
between 0.4 and 0.9. The cone factors shown in Fig. factor for an advancing cone in anisotropic clay has
12 are the upper bound (=21.8) and lower bound been developed based on the spherical expansion
(=12.2) of N *kt, A . They are close to the variation range theory and anisotropic strength criteria. The suitabil-
of cone factor (=11~20.5) reported by Aas et al. ity of the proposed cone factor has been verified with
(1986). For comparison, if Prevost’s (1979) field test data. The following conclusions can be
undrained strength criterion is used to calculate the made from the findings of this study:
 S.F. Su and H.J. Liao: Cavity Expansion and Cone Penetration Resistance in Anisotropic Clay 669

1. An undrained shear strength of spherical cavity s u(sc) undrained shear strength under spheri-
expansion has been derived from Su’s (1998) and cal expansion conditions
Prevost’s (1979) anisotropic strength criteria. The dλ proportionally scalar function
undrained shear strength varies with the position σ x , σ y, σ z normal stresses acting on x-, y-, z-direc-
angle θ in a spherical cavity and reaches a mini- tions
mum value in the horizontal direction ( θ =90°). σ r , σ θ, σ φ radial, tangential, and hoop stresses of
Even so, it is still larger than the undrained shear an spherical cavity expansion
strength determined from the CK oUE triaxial test. σo initial in-situ stress
2. The cone factor N kt,I (Eq. (17)) for isotropic σ ho horizontal stress resulting from overbur-
undrained shear strength is close to the cone fac- den
tors for rough cone surfaces reported by other σ vo vertical overburden pressure
investigators. The cone factor Nkt,A (Eq. (19)) for τxy, τ yz, τ zx shear stresses acting at xy-, yz-, and zx-
anisotropic undrained shear strength is in good planes
agreement with the CPTU results of Bothkennar ψ s,A limit pressure in spherical cavity expan-
clay. sion derived from anisotropic strength
3. Based on the anisotropic undrained strength criterion
criterion, the calculated cone factors uniquely cor- ψ s,I limit pressure in spherical cavity expan-
respond to the undrained shear strengths deter- sion derived from isotropic strength cri-
mined from different tests. These cone factors are terion
interrelated to each other in terms of the strength θ position angle within the spherical cav-
anisotropy ratio, A r, of clay. ity (refer to Fig. 5)
4. The effect of strength anisotropy on the cone fac-
tor can become obvious only for soil with low ri- REFERENCES
gidity (I r<100) and high strength anisotropy (A r<
0.4). But the error resulting from not taking into 1. Aas, G., Lacasse, S., Lunne, T., and Hoeg, K.,
account the strength anisotropy on the cone factor 1986, “Use of In-Situ Tests for Foundation De-
is not larger than 15%. sign on Clay,” Proceedings of In-Situ ’86, A Spe-
cialty Conference sponsored by the Geotechnical
NOMENCLATURE Engineering Division of ASCE, Blacksburg,
Virginia, pp. 1-30.
Ar strength anisotropy ratio (=s ue/s uc) 2. Baligh, M. M., 1975, “Theory of Deep Static Cone
d radius of cone Penetration Resistance,” Research report R75-
Ir rigidity index = shear modulus to 56, Department of Civil Engineering, Massachu-
undrained shear strength ratio (G/s u) setts Institute of Technology.
Ko coefficient of lateral earth pressure at 3. Baligh, M. M., 1984, “The Simple-Pile Approach
rest to Pile Installation in Clay,” Proceedings of Sym-
N kt cone factor for correlating q t and s u posium on Analysis and Design of Pile
Nkt, I cone factor for the isotropic undrained Foundation, sponsored by the Geotechnical
shear strength Division, ASCE, National Convention, San
Nkt, A cone factor for the anisotropic undrained Francisco, pp. 310-330.
shear strength. 4. Baligh, M. M., 1985, “Strain Path Method,” Jour-
N *kt, A cone factor for correlating qt and s u,lab nal of the Geotechnical Engineering Division,
qt cone resistance corrected for pore pres- ASCE, Vol. 111, No. 9, pp. 1108-1136.
sure effects 5. Campanella, R. G., Robertson, P. K., and
r radius of the spherical cavity Gillespie, D., 1983, “Cone Penetration Testing in
rp radius of plastic zone around an expand- Deltaic Soils,” Canadian Geotechnical Journal,
ing sphere Vol. 20, No. 1, pp. 23-35.
su undrained shear strength 6. D u r g u n o g l u , H . T . , a n d M i t c h e l l , J . K . ,
s uc undrained shear strength obtained from 1975, “Static Penetration Resistance of Soils. I:
CK o UC triaxial test Analysis,” Proceedings, ASCE Specialty Confer-
s ue undrained shear strength obtained from ence on In Site Measurement of Soil Properties,
CK o UE triaxial test ASCE, New York, Vol. 1, pp. 151-171.
s ud undrained shear strength obtained from 7. Hight, D. W., Bond, A. J., and Legge, J. D.,
direct simple shear test 1992, “Characterization of the Bothkennar Clay:
s u,lab average laboratory undrained shear an overview,” Geotechnique, Vol. 42, No. 2, pp.
strength (= (s uc +s ud +s ue )/3) 303-347.
670 Journal of the Chinese Institute of Engineers, Vol. 24, No. 6 (2001)

8. Houlsby, G. T., and Withers, N. J., 1988, “Analy- Clay,” Journal of the Geotechnical Engineering
sis of the Cone Pressuremeter Test in Clay,” Division, ASCE, Vol. 105, No. 1, pp. 49-64.
Geotechnique, Vol. 38, No. 4, pp. 575-587. 20. Roscoe, K. H., 1953, “An Apparatus for the Ap-
9. K e a v e n y , J . M . , a n d M i t c h e l l , J . K . , plication of Simple Shear to Soil Samples,” Pro-
1986, “Strength of Fine Grained Soils Using the ceedings Of 3rd ICSMFE, Zurich, Session 2, pp.
Piezocone,” Proceedings of In-Situ ’86, a Spe- 186-191.
cialty Conference sponsored by the Geotechnical 21. Salgado, R., 1993, “Analysis of Penetration re-
Engineering Division of ASCE, Blacksburg, sistance in Sand,” Ph.D. thesis, University of
Virginia, pp. 668-685. California, Berkeley.
10. Kiousis, P. D., Voyiadjis, G. Z., and Tumay, 22. Su, S. F., Liao, H. J., and Lin, Y. H., 1998, “Base
M. T., 1988, “A Large Strain Theory and its Ap- Stability of Deep Excavation in Anisotropic Soft
plication in the Analysis of the Cone Penetration Clay,” Journal of Geotechnical and Geoenviron-
Mechanism,” International Journal for Numeri- mental Engineering, ASCE, Vol. 124, No. 9, pp.
cal and Analytical Methods in Geomechanics, 809-819.
Vol. 12, pp. 45-60. 23. Teh, C. I., and Houlsby, G. T., 1991, “An Ana-
11. Koutsoftas, D. C., and Ladd, C. C., 1985, “De- lytical Study of the Cone Penetration Test in
sign Strengths for an Offshore Clay,” Journal of Clay,” Geotechnique, Vol. 41, No. 1, pp. 17-34.
Geotechnical Engineering, ASCE, Vol. 111, No. 24. Van den Berg, V., 1994, “Analysis of Soil
3, pp. 337-355. Penetration,” Ph.D. thesis, Delft University,
12. Ladanyi, B., and Johnston, G. H., 1974, “Behav- Delft, The Netherlands.
ior of Circular Footings and Plate Anchors Em- 25. Vesic, A. S., 1972, “Expansion of Cavities in In-
bedded in Permafrost,” Canadian Geotechnical finite Soil Mass,” Journal of the Soil Mechanics
Journal, Vol. 11, No. 4, pp. 531-553. and Foundation Division, ASCE, Vol. 100, No.
13. Ladd, C. C., and Edgers, L., 1972, “Consolidated- 3, pp. 207-224.
undrained Direct Simple Shear Tests on Boston 26. Vesic, A. S., 1977, Design of Pile Foundations,
Blue Clay,” Research report R72-82, Department Synthesis of Highway Practice 42, Transportation
of Civil Engineering, Massachusetts Institute of Research Board, National Research Council,
Technology. Washington, D.C.
14. Liao, H. J., and Su, S. F., 1999, “Behavior Grout 27. Yasufuku, N., and Hyde, A. F. L., 1995, “Pile
Column Reinforced Clay under Lateral End-Bearing Capacity in Crushable Sands,”
Compression,” Proceedings of International Sym- Geotechnique, Vol. 45, No. 4, pp. 663-676.
posium on Geotechnical Aspects of Underground 28. Yu, H. S., and Houlsby, G. T., 1991, “Finite
Construction in Soft Ground, pp. 649-654. Cavity Expansion in Dilatant Soil: Loading
15. Meigh, A. C., 1987, Cone Penetration Testing, Analysis,” Geotechnique, Vol. 41, No. 2, pp. 173-
CIRIA Report, Butterworth’s. London U.K. 183.
16. Mesri, G., 1989, “A Reevaluation of S u(mob) = 29. Yu, H. S., 1993, “Discussion on: Singular Plastic
0.22 σ′p Using Laboratory Shear Tests,” Canadian Fields in Steady Penetration of a Rigid Cone,”
Geotechnical Journal, Vol. 26. No. 1, pp. 162- Journal of Applied Mechanics Division, ASME,
164. Vol. 60, pp. 1061-1062.
17. Mesri, G., 1993, “Discussion on: Initial Investi- 30. Yu, H. S., and Mitchell, J. K., 1998, “Analysis of
gations of the Soft Clay Test Site at Bothken- Cone Resistance: Review of Methods,” Journal
nar,” Geotechnique, Vol. 43, No. 3, pp. 503-504. of Geotechnical and Geoenvironmental Engineer-
18. Nash, D. F. T., Powell, J. J. M., and Lloyd, I. M., ing, ASCE, Vol. 124, No. 2, pp. 140-149.
1992, “Initial Investigations of the Soft Clay Test
Site at Bothkennar,” Geotechnique, Vol. 42, No. Manuscript Received: Mar. 02, 2000
2, pp. 163-181. Revision Received: Jun. 18, 2001
19. Prevost, J. H., 1979, “Undrained Shear Tests on and Accepted: July 04, 2001
S.F. Su and H.J. Liao: Cavity Expansion and Cone Penetration Resistance in Anisotropic Clay 671

 !"#$%&'()*+,-./



 !"#$%&'(



 !"#$%&'()*



 !"#$%&'()*+,-.%/0123 !456789:

 !"#$%&'( )*+%,-)./0123 45+26789

 !"#$%&'()*+,-.#/012345'6789:;<=

 !"#$
 %&'()*!+,-./012345678


 !"#$%&'(
)* +,-./0123456789:;<

 !"#$%&'(#$%&)*+,-./0123456789


 !"#$%&'()*+,-./01,234%5,2-.67(8

 !"#$%&'()*+,-./0123456789:;<=>:

 !"#$%&'()*+,-./0123456789*+:;<=
(A r)
 !"#$%&'()*+,-./0123456)789:(

 !"#
$%& '()*+,-./0,123 456789:

 !"#$%&'()*+,$%-./012 15%

 !"#$%&'()*+',-./01203+4