SCIENCE CHINA

Technological Sciences
• Article • doi: 10.1007/s11431-016-9024-9

Model predictions for behaviors of sand-nonplastic-fines

mixturesusing equivalent-skeleton void-ratio state index
XIAO Yang1,2,3*, SUN YiFei4, LIU HanLong1,3, XIANG Jia3, MA QiFeng3 & LONG LeiHang3
1
Key Laboratory of New Technology for Construction of Cities in Mountain Area, Chongqing University, Chongqing 400045, China;
2
State Key Laboratory of Coal Mine Disaster Dynamics and Control, Chongqing University, Chongqing 400030, China;
3
School of Civil Engineering, Chongqing University, Chongqing 400045, China;
4
Faculty of Engineering and Information Sciences, University of Wollongong, Wollongong 2522, Australia

Received August 14, 2016; accepted March 6, 2017; published online May 16, 2017

It is a challenge to suggest a constitutive model for describing the stress-strain behavior of sand-fines mixtures due to that
these granular mixtures exhibited very complex behaviors at different densities, pressures and fines contents. In this study,
an elastoplastic constitutive model within the framework of the bounding surface plasticity and critical state theories was
proposed for sand-nonplastic-fines mixtures by using the concept of the equivalent-skeleton void ratio and equivalent-skeleton
void-ratio state index. The proposed model with a set of material constants calibrated from a few tests could be used to model
the fines-dependent and state-dependent behaviors of the sand-nonplastic-fines mixture including the strain-softening and
volumetric-expansion behaviors in the drained triaxial compression tests, and also the effects of fines content on the critical state
lines in both the deviatoric stress versus mean effective stress and the void ratio versus mean effective stress planes.

sand, fines, equivalent-skeleton void ratio, critical-state strength, dilatancy, constitutive model

Citation: Xiao Y, Sun Y F, Liu H L, et al. Model predictions for behaviors of sand-nonplastic-fines mixturesusing equivalent-skeleton void-ratio state index. Sci
China Tech Sci, 2017, 60, doi: 10.1007/s11431-016-9024-9

1 Introduction shifted downwards with increasing the fines content until the
so-called threshold fines [10] was reached. This behavior
Natural sands usually contain amounts of fines which could of sand-fines mixtures was found by many researchers, e.g.,
significantly influence their behaviors [1–8]. Seed et al. Huang et al. [11], Yang et al. [6], Murthy et al. [12], Chiu
[9] and Pitman et al. [1] found that the addition of fines and Fu [13], Papadopoulou and Tika [14], Bobei et al. [15],
could increase the liquefaction resistance of the sand deposit. Stamatopoulos [16], Rahman et al. [7], Yang and Wei [17],
However, Thevanayagam and Mohan [2] recognized that Rahman and Lo [18], Yin et al. [19], Yang et al. [20], and
these findings of the effect of fines on the behavior the Yang and Luo [21]. Then, the CSL moved upwards with
sand-fines mixture were limited. The complexity due to continuously increasing the fines content [10]. This parallel
the addition of fines could be reflected by the shift of the CSL is similar to that in crushable soils [22–24] or to that in
critical state line (CSL) or steady state line (SSL) with the transitional soils [25]. Rahman et al. [26] found that the crit-
fines content in the global void ratio versus mean effective ical-state friction angle was independent of the fines content
stress (i.e., e-p′) plane. The CSL in the e-p′ plane initially while other researchers, e.g., Salgado et al. [3], Yang et al.
[6], Chiu and Fu [13], Murthy et al. [12], and Papadopoulou
and Tika [14] found that the critical-state friction angle was
* Corresponding author (email: [email protected]) dependent on the fines content.

© Science China Press and Springer-Verlag Berlin Heidelberg 2017 tech.scichina.com link.springer.com
2 Xiao Y, et al. Sci China Tech Sci

Modeling the stress-strain relations of sand-fines mixtures skeleton or they partially separated the sand grains. Con-
is a challenge because not only the density and pressure but sequently, the skeleton void ratio was then extended as an
also the interaction of sand grains and fines should be con- equivalent-skeleton void ratio by incorporating a factor b:
sidered in the description of constitutive equation. Chang
and Yin [27] and Lashkari [28,29] proposed valuable models e = [e + (1 b)f c ] / [1 (1 b)f c ], (2)
to describe the stress-strain behaviors of sand-fines mixtures where e is the equivalent-skeleton void ratio; b is defined as
and the evolution of the CSL in the e-p′ plane. Rahman et al. the portion of the fines grains that contribute to the force chain
[26] suggested a modified unified clean-sand model proposed in the microstructure of the sand-fines mixture.
by Li and Dafalias [30] for the sand-fines mixture by using The factor b ranges from 0 to 1. Evidently, e reduces
the equivalent-skeleton state parameter for the dilatancy and to es with b=0, indicating that the fines would act as voids;
peak-failure stress ratios. The determinations of these model e becomes e with b=1, meaning that the fines would act
parameters were similar to that by Li and Dafalias [30] based like the base-sand grains. In recent years, the equiva-
on a few drained tests with a single fines content. This model lent-skeleton void ratio was used by many investigators
could predict the flow, limited flow and non-flow behaviors [5–7,13,18,20,21,26–29,35–37] to characterize the behavior
of sand-fines mixtures. However, this model could not con- of the sand-fines mixtures.
sider the effect of fines content on the critical-state friction
angle of sand-fines mixtures as observed by Salgado et al.
[3], Yang et al. [6], Chiu and Fu [13], Murthy et al. [12], and 3 Drained triaxial tests
Papadopoulou and Tika [14].
The host sand used in this study is Fujian sand, which is
In the current study, an elastoplastic model within the
quartzitic. The minerology of the nonplastic fines is the same
framework of the bounding surface plasticity and critical
as that of the host sand. Fines contents in this study are
state theories was proposed for the sand-fines mixture by
10% and 20%. The particle size distributions of the host
using the concept of the equivalent-skeleton void ratio and
sand ( f c =0) and other two mixtures are shown in Figure 1.
state index and also by considering the effect of the fines on
The grain size of the nonplastic fines ranges from 0.005 to
the critical-state strength.
0.075 mm while the grain size of host sand ranges from 0.075
to 1 mm, as shown in Figure 1.
2 Effective void ratio Specimen was prepared with higher relative density rang-
ing from 0.7 to 0.95. The specimen size was 39.1 mm in
Sand’s behavior could be changed with addition of fines into a diameter and 80 mm in a height. The specimen was sat-
its voids and surfaces, which could be attributed to the mi- urated with the B-values larger than 0.94. The confining
crostructure of the sand-fines mixture with different types of pressure ranged from 0.1 to 0.8 MPa. All specimens were
inter-grain contacts. In the current study, the microstructure sheared under the conventional drained triaxial compres-
of the sand matrix where the sand is primarily in contact is sion condition with a constant vertical displacement rate
investigated. It is necessary to find an effective density index of 0.17 mm min‒1. The test details of sand-nonplastic-fines
for characterizing the behavior of this sand-fines mixture. mixtures are listed in Table 1. The test data in the Group
A are used for calibration of model constants, while the
2.1 Skeleton void ratio
test data in the Group B are used for validation of model
Based on the hypothesis that fines may be confined within capabilities (Figure 2).
the void spaces between sand grains with little contribution
to support the sand grain skeleton, the skeleton void ratio
was used as an alternative to describe the sand-fines mixture
[1–4,14,31–34].

e s = (e + f c ) / (1 f c ), (1)

where e is the global void ratio; es is the skeleton void ratio;

f c is the fines content. Fines acted like voids based on the
definition of eq. (1). For a clean sand ( f c =0), es is exactly
identical to e .

2.2 Equivalent-skeleton void ratio

Thevanayagam and Shenthan [10] recognized that not all Figure 1 Grain size distribution of sand-fines mixtures with fines content
fines acted like voids but they partially supported the sand ranging from 0 to 20%.
 Xiao Y, et al. Sci China Tech Sci 3

Table 1 Test details for sand-nonplastic-fines mixturesa)

Series Group No. Fines content (%) Confining pressure (MPa) Relative density (%)
T-0-0.1 0 0.1 79.7
T-0-0.6 0 0.6 86.7
Group A
T-10-0.4 10 0.4 79.8
T-20-0.4 20 0.4 89.2
T-0-0.2 0 0.2 79.5
T-0-0.4 0 0.4 81.3
T-0-0.8 0 0.8 87.9
T-10-0.1 10 0.1 78.5
T-10-0.2 10 0.2 76.2
T-10-0.6 Group B 10 0.6 84.5
T-10-0.8 10 0.8 84.6
T-20-0.1 20 0.1 87.4
T-20-0.2 20 0.2 86.1
T-20-0.6 20 0.6 92.9
T-20-0.8 20 0.8 95.0
a)
In “T-a-b”, “T”, “a” and “b” denote test, fines content and confining pressure, respectively.

5 Critical state line and equivalent-skeleton

state index
Figure 3(a) shows that the critical state line (CSL) observed in
the drained triaxial compression tests for the host sand ( f c =0)
could be described by the formulation proposed by Li and
Wang [39]:

ecs = ecsfc00 + cs (p cs / pa ) , (4)

where ecs is the critical-state void ratio; p cs is the critical-state

mean effective stress; ecsfc00, cs and are material constants.
Figure 2 Critical-state stress ratios of sand-nonplastic-fines mixtures in re-
lation to the fines content. is usually set as 0.7 for sands. The values of ecsfc00 and cs can
be determined by fitting the test data in Figure 3(a).
Figure 3(a) illustrates that the CSL for f c =20% is parallel
4 Critical-state stress ratio
to that for f c =0. The material constant ecsfc020 is the intercept
The observation in the current study is that an increase in the for the mixture with f c =20%. The difference between the
nonplastic fines could increase the critical-state friction angle initial state and the corresponding CSL in Figure 3(a) could be
of the sand-nonplastic-fines mixture, which is also verified described by a state parameter proposed by Been and Jefferies
by the tests conducted by Salgado et al. [3], Yang et al. [6], [40]:
Chiu and Fu [13], Murthy et al. [12], Papadopoulou and Tika 0 = e0 ecs, (5)
[14] and Carraro et al. [38]. The critical-state stress ratio by
where 0 is an initial state parameter; e 0 is an initial global
considering the effect of fines content can be given as
void ratio.
M cs = Mcs0 + f ,
m cs c (3) The application of the equivalent-skeleton void ratio is to
where M cs is the critical-state stress ratio; M cs0 and m cs are obtain a unique CSL and a unique state parameter. The CSL
material constants. of the host sand (i.e., f c =0) is a basis line. All test data with
The material constants M cs0 and m cs could be determined different fines contents are close to this basis line by using
by the values of the critical-state stress ratio M cs at f c =0 and e where the value of the factor b needs to be firstly evaluated.
another specific f c . These M cs values could be obtained from It implies that the intercept ecsfc020 of the CSL for f c =20% needs
two drained tests in the Group A. It is noted that the test data to be transformed to the intercept ecsfc00 of the CSL for f c =0.
in the Group A are used for calibrating model constants. Based on eq. (2), the intercepts ecsfc00 and ecsfc020 in Figure 3(a)
4 Xiao Y, et al. Sci China Tech Sci

I e = e / ecs, (9a)

I e0 = e 0 / ecs, (9b)
where I e is the equivalent-skeleton void-ratio state index;
I e0 is the initial equivalent-skeleton void-ratio state index.

6 Elastoplastic constitutive model

Many constitutive models were proposed for geomaterials,
e.g., the bounding surface model [30,41–50], generalized
plasticity model [51–54], disturbed state concept model
[55–61] and unified hardening model [62–65]. The frame-
work of the unified state-dependent bounding surface
proposed by Li and Dafalias [30] for sand is simple, which
has been extended for overconsolidated clay [66], transparent
granular soils [67] and rockfill materials [68].
In this study, the bounding-surface plasticity framework
proposed by Manzari and Dafalias [43] and Li and Dafalias
[30] was used for the sand-nonplastic-fines mixtures.

6.1 Elastic-strain increments

Figure 3 Critical state lines of sand-nonplastic-fines mixtures in terms of
global void ratio (a) and equivalent-skeleton void ratio (b). The total-strain increment d i can be expressed as the sum of
the elastic-strain increment d ie and plastic-strain increment
should satisfy the following equations: d i p:

e = ecsfc00 = ecsfc020 + (1 b)f c / [1 (1 b)f c ] . (6) d i=d i

e
+d i
p
. (10)

Rearrangement of eq. (6) gives The elastic-strain increment is defined as

d e
= dp / Be, (11a)
b = ecsfc020 ecsfc00 + (1 + ecsfc00)f c / (1 + ecsfc00)f c . (7) v

The value of the factor b for f c =20% could be obtained from d s

e
= dq / (3Ge), (11b)
eq. (7) as 0.381. In the same way, b=0.422 for f c =10%. All e
where d is an elastic volume-strain increment; Be is an elas-
v
test data can be plotted in the equivalent-skeleton void ratio tic bulk modulus; dp is a mean effective-stress increment;
versus the mean effective stress plane by using the equivalent- d se is an elastic shear-strain increment; Ge is an elastic shear
skeleton void ratio. As can be seen from Figure 3(b), the basis modulus; dq is a deviatoric stress increment.
CSL could well describe the test data from both Group A and The elastic shear modulus Ge and elastic bulk modulus
Group B, implying that the determination of the factor b is Be (proposed by Hardin and Richart [69], Richart et al.
reasonable. [70] and Li and Dafalias [30]) can be rewritten for the
Figure 3(b) shows that an equivalent-skeleton state param- sand-nonplastic-fines mixture by using e :
eter is defined in the equivalent-skeleton void ratio versus the
mean effective stress plane, which was originally proposed Ge = G0(2.97 e ) 2 p pa / (1 + e ), (12a)
by Rahman and Lo [36],
=e ecs, (2.97 e ) 2 2(1 + )
(8a) Be = G0 p pa , (12b)
1+e 3(1 2 )
0 = e0 ecs, (8b) where G0 is an elastic material constant; is the Poison’s
where is the equivalent-skeleton state parameter; 0 is the ratio. =0.25 in the current study for the sand-nonplastic-
initial equivalent-skeleton state parameter; e 0 is the initial fines mixture.
equivalent-skeleton void ratio; ecs is the critical-state equiv-
alent-skeleton void ratio. The mean effective stress for e 0 is 6.2 Plastic-strain increments
the same as that for ecs . According to the assumption of non-associated flow rule, the
In the current study, an equivalent-skeleton void-ratio state plastic volume-strain increment d vp and plastic shear-strain
index was used: increment d sp are defined as
 Xiao Y, et al. Sci China Tech Sci 5

d v
p
= (n vf dp + n s f dq )n vg / H p, (13a) where d is the dilatancy state index; kd is a material constant.
Consequently, a state-dependent stress-dilatancy equation
d s
p
= (n v dp + n s dq )n / H p,
f f
s
g
(13b)
for the sand-nonplastic-fines mixture can be given as
f f
where n and n s are compression-related and shearing-re-
v
lated components of a loading unit vector, respectively; sin (1 + kd lnI e )sin cs
= Rd
m
sin . (19)
n vg and n sg are compression-related and shearing-related
m
1 (1 + kd lnI e )sin m sin cs
components of a plastic-flow unit vector, respectively; Hp is The above general state-dependent stress-dilatancy equa-
a plastic modulus. tion is expressed as a function of the equivalent-skeleton
6.3 Dilatancy-state index and stress-dilatancy equation void-ratio state index I e for sand-nonplastic-fines mixtures.

The stress-dilatancy equation proposed by Rowe [71] can be 6.4 Peak-state index and peak-state stress ratio
given as A peak-state index in relation to the equivalent-skeleton void-
sin m = (sin m sin cs) / (1 sin m sin cs), (14) ratio state index is defined as
where m is a dilatancy angle; m is a mobilized friction angle; =1 k p lnI e ,
p (20)
cs is a critical-state friction angle.
It is noted that eq. (14) is based on the analogy of the ir- where p is the peak-state index; k p is a material constant.
regular packing of uniform rods using the assumption of the The peak-state stress ratio M p is given as
minimum energy ratio at failure, which cannot describe the
dilatancy in three-dimensional stress space [72,73]. There M p = M cs p = M cs(1 k plnI e ) . (21)
would be great differences in the particle shape, particle size, 6.5 Yield surface and loading unit vector
particle grading, particle texture and particle strength for the
actual soils (sand and rockfills) and the uniform rods used in A simple yield surface [30] was adopted in the current study:
the literature [71]. Therefore, Xu and Song [74] revised the f=q p = 0, (22)
Rowe’s stress-dilatancy equation by adding a parameter for
where is a stress ratio.
considering these differences:
The loading unit vector which is normal to the yield surface
sin m = R d (sin m sin cs ) / (1 sin m sin cs ), (15) under the triaxial compression condition can be given as
where R d is a dilatancy material constant. 1 f
In addition, the stress-dilatancy equation of eq. (14) cannot n vf = = , (23a)
Lf p Lf
capture the state-dependent evolution of the dilatancy (i.e.,
1 f 1
the effect of the void ratio and pressure on the dilatancy), n sf = = , (23b)
Lf q Lf
as pointed by Yang and Li [75]. As a result, Wan and Guo
2 2
[76] modified the Rowe’s stress-dilatancy equation by incor- f f
porating the void-ratio state index: Lf = + = 1+ 2
, (23c)
p q
sin m = [sin m (Ie) sin cs] / [1 (Ie) sin m sin ],
cs (16) where L f is a gradient amplitude of the loading vector.
where Ie is the void-ratio state index (Ie = e / ecs ); is a dila-
6.6 Dilatancy and plastic-flow unit vector
tancy material constant.
It should be noted that the state-dependent characteristics is A dilatancy d by neglecting the elastic strains is defined as:
not incorporated into the modification of Rowe’s stress-dila- d p
3d
d= =
v v
tancy equation [74], while the differences in particle charac- p
, (24)
d s 3d a d v
teristics between the real soils and idealized uniform rods are
not taken into account in the modification of Rowe′s stress-di- where d a and d v are axial-strain and volume-strain incre-
latancy equation [76]. Thus, Xiao et al. [54] proposed a ments, respectively.
state-dependent stress-dilatancy equation by combining the According to the Rowe’s stress-dilatancy theory, the fol-
modifications from Wan and Guo [76] and Xu and Song [74]: lowing dilatancy angle was given as

sin = R d [sin (Ie) sin ] / [1 (Ie) sin m sin ]. (17) d 1 + sin

=
v m
m m cs cs 1 . (25)
d 1 sin
Eq. (17) cannot consider the effect of fines content and a m

therefore can be further revised with a dilatancy state index Combination of eqs. (24) and (25) gives
in relation to the equivalent-skeleton void-ratio state index
d = 6sin m / (3 sin m ). (26)
= 1 + kd lnI e = 1 + kd ln(e / ecs ), (18)
d Substitution of eq. (19) into eq. (26) gives
6 Xiao Y, et al. Sci China Tech Sci

6R d sin m (1 + kd lnI e )sin cs

d= . (27)
{3 1 (1 + kd lnI e )sin m sin cs R d sin m (1 + kd lnI e )sin cs }

Eq. (27) indicates that the dilatancy is correlated to the Gi = q / (3 s), (32)
equivalent-skeleton void-ratio state index I e . where s is a shear strain.
The plastic-flow unit vector under the triaxial compression The elastic material G0 can be obtained from combinations
condition can be given as of eqs. (12a) and (32) as

n vg = d/ 1 + d2, (28a) q (1 + e )
G0 = . (33)
3 s (2.97 e ) 2 p p
a

n sg = 1 / 1 + d2 . (28b) 7.2 Critical state

Substitution of eq. (27) into eq. (28) gives As mentioned in the third section, the material constants
M cs0 and m cs could be determined by the critical-state stress
n vg = 6sin m / 36sin 2 m + (3 sin m )2 , (29a)
ratio from two drained triaxial compression test data at f c =0
and another specific f c . These test data are from the Group
n sg = 3 sin m / 36sin 2 m + (3 sin m )2 . (29b) A.
The material constants ecs0 and cs can be obtained from
Eq. (29) illustrates that the plastic flow direction is in rela-
eq. (4) by fitting the test data, as can be seen from Figure 3(a).
tion to the equivalent-skeleton void-ratio state index.
7.3 Dilatancy
6.7 Plastic modulus
According to eqs. (24) and (27), the dilatancy material con-
Following the work [30], a plastic modulus Hp in relation to
stant R d can be determined as
the peak-state stress ratio M p can be given as

Rd =
3d {sin
v m 1 + kd ln(e / ecs ) sin }
cs
.
e0 e 0 )(M p ) /[
Hp = Ge(h e0 (1 + ) 2], (30) (34)
(d v 6d ){1 s 1 + kd ln(e / ecs) sin m sin cs }
where h e0 and e0 are material constants for the plastic modu-
Therefore, R d can be obtained from the v– s curve. The di-
lus.
latancy is zero when a phase-transformation state is reached.
Substitution of eqs. (12) and (21) into eq. (30) gives
Therefore, the material constant kd can be determined from
G0 (h e0 e ) M cs(1
e0 0 k plnI e ) eq. (27) as
Hp = . (31)
(1 + e )( (1 + ) 2) / (2.97 e ) 2 p pa kd = (sin opt sin cs ) / sin ln(eopt / ecs) ,
cs (35)

Eq. (10) shows that the plastic modulus is pertaining to where opt and eopt is the observed mobilized friction angle and
the equivalent-skeleton void ratio and equivalent-skeleton observed equivalent-skeleton void ratio at the phase-transfor-
void-ratio state index. Consequently, the fines-dependent and mation state.
state-dependent constitutive model is proposed for sand-non- The critical-state friction angle cs is given as
plastic-fines mixtures with a set of material constants, i.e., sin cs = (3Mcs) / (6 + M cs) . (36)
G0, , M cs0 , m cs, ecs0 , cs, R d , kd , k p, h e0 and e0 . As mentioned
cs can be determined by substituting eq. (3) into eq. (36):
before, the values of and for granular soils are usually
set as 0.25 and 0.7, respectively. The calibrations of other sin cs = 3(Mcs0 + mcs cf ) / (6 + M cs0 + mcs c f ). (37)
material constants are introduced in the following section. 7.4 Peak state
The plastic modulus at the peak state is zero. Thus, the peak-
7 Calibrations of material constants state material constant k p can be derived as
7.1 Elasticity
k p = (M cs M ops) / M csln(eops / ecs) , (38)
The elastic material constant G0 can be determined from the
initial shear modulus Gi , which was obtained from an initial where M ops and eops is the observed stress ratio and observed
slope in the q – s curve of this mixture. equivalent-skeleton void ratio at the peak state.
 Xiao Y, et al. Sci China Tech Sci 7

7.5 Plastic modulus

The stress path under the conventional drained triaxial com-
press can be expressed as

dp = dq / 3, (39)

Based on eqs. (10)–(13), (23), (28) and (30), the plastic-

modulus parameter h 0 can be determined as

h 0 = (h e0 e
e0 0 )
(1 / 3)(1 + )2 / 1 + 2
= .
1 1
Ge 1 + d (M p
2
) dq / d s 3Ge
(40)
The sand-nonplastic-fines mixture with two independent
e 0 from the Group A yields two h 0 due to eq. (40). Thus,
the material constants h e0 and e0 can be evaluated from two
sets of e 0 and h 0. The values of all material constants are
listed in Tables 1 and 2.

8 Model predictions
Figures 4 and 5 show the comparisons between model predic-
tions and test data of the sand-nonplastic-fines mixture from
the Group A. These test data in the Group A are used for
calibrations of material constants. It is found that the evo-
lutions of the deviatoric stress, mobilized friction angle, vol-
umetric strain, global void ratio and equivalent-skeleton state
index with the axial strain for the four test data in the Group
A could be well described by the fines-dependent state-de-
pendent constitutive model. In addition, the strain-softening
and volume-expansion behaviors of the sand-nonplastic-fines
mixture could be well captured by the proposed model.

Table 2 Material constants for the sand-nonplastic-fines mixture

Material constants Symbols Values
G0 133.04
Elastic constant
0.25
M cs0 1.23

m cs 0.86
Critical-state constants ecs0 0.677

cs 0.022

0.7
Rd 0.6
Dilatancy constants
kd 4.77

Peak-state constant kp 2.52

Plastic-modulus h e0 13.14 Figure 4 (Color online) Comparisons between model simulations and test
constants data in Group A for sand-fines mixtures with fc=0%: (a) q versus εa; (b) ϕ
e0 22.95
versus εa; (c) εv versus εa; (d) e versus εa; (e) Ie versus εa.
8 Xiao Y, et al. Sci China Tech Sci

The validations of the current model are presented in

Figures 6–11 in terms of evolutions of the deviatoric stress,
mobilized friction angle, volumetric strain, global void ratio
and equivalent-skeleton state index with the axial strain.
Comparisons between model predictions and test data in
the Group B show that the proposed model can well de-
scribe the state-dependent and fines-dependent behaviors
of the sand-nonplastic-fines mixture. It was observed from
Figures 5, 9, 10 and 11 that the critical-state friction an-
gle increases with increasing the fines content, while the
critical-state friction angle is independent of the confining
pressure, as can be seen in Figures 4, 6, 7 and 8.
The evolutions of the stress-dilatancy relation of the sand-
nonplastic fines are presented in Figure 12. The proposed
model could well describe the stress-dilatancy behaviors at
various confining pressures and fines contents. The mixture
at a low confining pressure exhibits higher dilation and strain-
softening behaviors. A high confining pressure could oppress
the dilation and strain-softening behaviors.

9 Discussion
9.1 Limitations of the equivalent-skeleton void ratio
The concept of equivalent skeleton void ratio was incorpo-
rated into the bounding surface plasticity model to capture the
stress-strain behaviors of the sand-nonplastic-fines mixture.
The central hypothesis of this concept is that the mixed soil
and its host sand with the same equivalent skeleton void ra-
tio should exhibit the same stress-strain behavior. However,
as pointed by Yang et al. [20], this concept cannot provide
a universal way for characterizing the stress-strain behaviors
of the mixture, due to its lack of mechanisms to interpret the
highly complex inter-granular contacts.
As shown in Figures 13 and 14, specimens of the host sands
(Toyoura sand and Leighton Buzzard sand) and mixtures
were sheared at the same confining pressure of 0.5 MPa [20].
The stresses in the two figures were normalized by the
atmospheric pressure pa of 0.1 MPa. It was found from the
two figures that the clean Toyoura sand and clean Leighton
Buzzard sand exhibited more dilatant than their mixtures
with nonplastic fines. For Toyoura sand and its mixture with
nonplastic fines, the post-peak responses are significantly
different although the peak deviatoric stress is similar. The
clean Toyoura sand exhibited a slight drop in deviatoric
stress and regained the deviatoric stress after a quasi-steady
state, while the mixture with 10% nonplastic fines showed
a marked reduction in the deviatoric stress. Figures 13 and
14 also show that the clean Leighton Buzzard sand exhibited
the higher peak strength than its mixture with 5% nonplastic
fines. In addition, the mixture was more susceptible to the
Figure 5 (Color online) Comparisons between model simulations and test
data in Group A for sand-fines mixtures at σ3=0.4 MPa: (a) q versus εa; (b) onset of flow liquefaction than the clean Leighton Buzzard
ϕ versus εa; (c) εv versus εa; (d) e versus εa; (e) Ie versus εa. sand.
 Xiao Y, et al. Sci China Tech Sci 9

Figure 6 (Color online) Comparisons between model simulations and test Figure 7 (Color online) Comparisons between model simulations and test
data in Group B for sand-fines mixtures with fc=0%: (a) q versus εa; (b) ϕ data in Group B for sand-fines mixtures with fc=10%: (a) q versus εa; (b) ϕ
versus εa; (c) εv versus εa; (d) e versus εa; (e) Ie versus εa. versus εa; (c) εv versus εa; (d) e versus εa; (e) Ie versus εa.
10 Xiao Y, et al. Sci China Tech Sci

Figure 8 (Color online) Comparisons between model simulations and test Figure 9 (Color online) Comparisons between model simulations and test
data in Group B for sand-fines mixtures with fc=20%: (a) q versus εa; (b) ϕ data in Group B for sand-fines mixtures at σ3=0.1 MPa: (a) q versus εa; (b)
versus εa; (c) εv versus εa; (d) e versus εa; (e) Ie versus εa. ϕ versus εa; (c) εv versus εa; (d) e versus εa; (e) Ie versus εa.
 Xiao Y, et al. Sci China Tech Sci 11

Figure 10 (Color online) Comparisons between model simulations and test Figure 11 (Color online) Comparisons between model simulations and test
data in Group B for sand-fines mixtures at σ3=0.2 MPa: (a) q versus εa; (b) data in Group B for sand-fines mixtures at σ3=0.6 MPa: (a) q versus εa; (b)
ϕ versus εa; (c) εv versus εa; (d) e versus εa; (e) Ie versus εa. ϕ versus εa; (c) εv versus εa; (d) e versus εa; (e) Ie versus εa.
12 Xiao Y, et al. Sci China Tech Sci

Figure 13 (Color online) Stress-strain characteristics of mixtures at a simi-

lar equivalent-skeleton void ratio under undrained shearing tests: (a) Toyoura
sand and Toyoura sand with nonplastic fines; (b) Leighton Buzzard sand and
Leighton Buzzard sand with nonplastic fines (data from Yang et al. [20]).

Figure 12 (Color online) Comparisons between model simulations and test

data on the stress-dilatancy relations: (a) fc=0%; (b) fc=10%; (c) fc=20%.

Figure 15 shows comparisons of the stress-strain relations

of Ottawa sand with nonplastic fines at the similar skeleton
void ratio under the same confining pressure of 400 kPa [38].
It was found that the mixture with 5% nonplastic fines showed
higher peak-state strength and critical-state strength but lower
dilatancy in comparisons with the mixture with 15% nonplas-
tic fines.
Consequently, the host sand and its mixtures at the similar
skeleton or equivalent-skeleton void ratio cannot behave in
the same way in terms of the strength, stress path and stress-
strain relations. In this paper, the critical-state stress ratio or
critical-state friction angle was pertaining to the fines content
(i.e., eq. (3)), which is mainly due to that the application
of the equivalent-skeleton void ratio as a universal density
parameter could not fully capture the complex stress-strain
behaviors of sand-fines mixtures
Figure 14 (Color online) Stress path of mixtures at a similar equivalent-
9.2 Limitations of the proposed constitutive model skeleton void ratio under undrained shearing tests: (a) Toyoura sand and
Toyoura sand with nonplastic fines; (b) Leighton Buzzard sand and Leighton
Limitations of the proposed constitutive model for the sand- Buzzard sand with nonplastic fines (data from Yang et al. [20]).
 Xiao Y, et al. Sci China Tech Sci 13

arithmic functions of this equivalent-skeleton void-ratio state

index.
(2) A modified Rowe’s stress-dilatancy equation incor-
porating the equivalent-skeleton void ratio and equiva-
lent-skeleton void-ratio state index was proposed. The plastic
modulus was also expressed as a function of the equiva-
lent-skeleton void ratio and equivalent-skeleton void-ratio
state index. A fines-dependent and state-dependent con-
stitutive model in the framework of the critical state and
bounding surface plasticity theories was proposed for the
sand-nonplastic-fines mixture.
(3) The test data in the Group A were used to calibrate the
material constants for the proposed model. While the test data
in the Group B were used to validate the prediction capability
of the proposed model. The comparisons between model pre-
dictions and test data showed that the proposed model could
well capture the strain-softening and volume-expansion be-
haviors of the sand-nonplastic-fines mixture.
(4) It was found in the drained triaxial compression tests
that the critical-state friction angle and critical state line were
dependent on the fines content. These observations could also
be captured in the proposed model.
Figure 15 (Color online) Drained shear responses of Ottawa sand with non-
plastic fines at a similar equivalent-skeleton void ratio: (a) normalized de-
viatoric stress versus axial strain relation; (b) volumetric strain versus axial This work was supported by the National Natural Science Foundation of
strain relation (data from Carraro et al. [38]). China (Grant Nos. 51509024 and 51678094), the Fundamental Research
Funds for the Central Universities (Grant No. 106112015CDJXY200008)
fines mixture are given as follow. and the Project funded by China Postdoctoral Science Foundation (Grant
No. 2016M590864).
(1) The mixture with plastic fines cannot be predicted by
the proposed model.
1 Pitman T D, Robertson P K, Sego D C. Influence of fines on the col-
(2) The proposed model could only used for mixtures,
lapse of loose sands. Can Geotech J, 1994, 31: 728–739
whose mechanical behavior should be controlled by the 2 Thevanayagam S, Mohan S. Intergranular state variables and stress-
sand matrix, implying that the fines content should not be strain behaviour of silty sands. Géotechnique, 2000, 50: 1–23
over a threshold value as proposed by Thevanayagam and 3 Salgado R, Bandini P, Karim A. Shear strength and stiffness of silty
sand. J Geotech Geoenviron Eng, 2000, 126: 451–462
Shenthan [10]. Sand grains are primarily in contact within
4 Chu J, Leong W K. Effect of fines on instability behaviour of loose
this threshold fines content. sand. Géotechnique, 2002, 52: 751–755
(3) The effect of particle shape of the host sand and fines 5 Ni Q, Dasari G R, Tan T S. Equivalent granular void ratio for char-
on the strength and stress-strain relation of mixtures is not acterization of Singapore’s old alluvium. Can Geotech J, 2006, 43:
563–573
considered in the proposed model.
6 Yang S L, Sandven R, Grande L. Steady-state lines of sand-silt mix-
It is necessary in the future work to propose a more inno- tures. Can Geotech J, 2006, 43: 1213–1219
vative constitutive model to overcome the limitations in the 7 Rahman M M, Lo S R, Baki M A L. Equivalent granular state param-
current model. eter and undrained behaviour of sand-fines mixtures. Acta Geotech,
2011, 6: 183–194
8 Xiao Y, Liu H, Chen Y, et al. Strength and dilatancy of silty sand. J
10 Conclusions Geotech Geoenviron Eng, 2014, 140: 06014007
9 Seed H B, Idriss I M, Arango I. Evaluation of liquefaction potential
A fines-dependent and state-dependent constitutive model in using field performance data. J Geotech Eng, 1983, 109: 458–482
the framework of the bounding surface plasticity and critical 10 Thevanayagam S, Shenthan T, Mohan S, et al. Undrained fragility of
clean sands, silty sands, and sandy silts. J Geotech Geoenviron Eng,
state theories was proposed for the sand-nonplastic-fines mix-
2002, 128: 849–859
ture by incorporating the equivalent-skeleton void ratio and 11 Huang Y T, Huang A B, Kuo Y C, et al. A laboratory study on the
equivalent-skeleton void-ratio state index. The main conclu- undrained strength of a silty sand from Central Western Taiwan. Soil
sions in this study can be summarized as follow. Dyn Earthq Eng, 2004, 24: 733–743
12 Murthy T G, Loukidis D, Carraro J A H, et al. Undrained monotonic
(1) An equivalent-skeleton void-ratio state index was pro-
response of clean and silty sands. Géotechnique, 2007, 57: 273–288
posed for the sand-nonplastic-fines mixture. Then, the dila- 13 Chiu C F, Fu X J. Interpreting undrained instability of mixed soils
tancy-state index and peak-state index were suggested as log- by equivalent intergranular state parameter. Géotechnique, 2008, 58:
14 Xiao Y, et al. Sci China Tech Sci

751–755 sands containing plastic or nonplastic fines. J Geotech Geoenviron

14 Papadopoulou A, Tika T. The effect of fines on critical state and liq- Eng, 2009, 135: 1167–1178
uefaction resistance characteristics of non-plastic silty sands. Soils 39 Li X S, Wang Y. Linear representation of steady-state line for sand. J
Found, 2008, 48: 713–725 Geotech Geoenviron Eng, 1998, 124: 1215–1217
15 Bobei D C, Lo S R, Wanatowski D, et al. Modified state parameter 40 Been K, Jefferies M G. A state parameter for sands. Géotechnique,
for characterizing static liquefaction of sand with fines. Can Geotech 1985, 35: 99–112
J, 2009, 46: 281–295 41 Bardet J P. Bounding surface plasticity model for sands. J Eng Mech,
16 Stamatopoulos C A. An experimental study of the liquefaction 1986, 112: 1198–1217
strength of silty sands in terms of the state parameter. Soil Dyn 42 Dafalias Y F, Herrmann L R. Bounding surface plasticity. II: Appli-
Earthq Eng, 2010, 30: 662–678 cation to isotropic cohesive soils. J Eng Mech, 1986, 112: 1263–1291
17 Yang J, Wei L M. Collapse of loose sand with the addition of fines: 43 Manzari M T, Dafalias Y F. A critical state two-surface plasticity
The role of particle shape. Géotechnique, 2012, 62: 1111–1125 model for sands. Géotechnique, 1997, 47: 255–272
18 Rahman M M, Lo S R. Undrained behavior of sand-fines mixtures 44 Ling H I, Liu H, Mohri Y, et al. Bounding surface model for geosyn-
and their state parameter. J Geotech Geoenviron Eng, 2014, 140: thetic reinforcements. J Eng Mech, 2001, 127: 963–967
04014036 45 Hu C, Liu H, Huang W. Anisotropic bounding-surface plasticity
19 Yin Z Y, Zhao J, Hicher P Y. A micromechanics-based model for model for the cyclic shakedown and degradation of saturated clay.
sand-silt mixtures. Int J Solids Struct, 2014, 51: 1350–1363 Comp Geotechnics, 2012, 44: 34–47
20 Yang J, Wei L M, Dai B B. State variables for silty sands: Global void 46 Xiao Y, Liu H, Chen Y, et al. Bounding surface model for rockfill
ratio or skeleton void ratio? Soils Found, 2015, 55: 99–111 materials dependent on density and pressure under triaxial stress con-
21 Yang J, Luo X D. Exploring the relationship between critical state and ditions. J Eng Mech, 2014, 140: 04014002
particle shape for granular materials. J Mech Phys Solids, 2015, 84: 47 Xiao Y, Liu H, Chen Y, et al. State-dependent constitutive model for
196–213 rockfill materials. Int J Geomech, 2015, 15: 04014075
22 Muir Wood D, Maeda K. Changing grading of soil: Effect on critical 48 Xiao Y, Liu H, Chen Y, et al. Bounding surface plasticity model incor-
states. Acta Geotech, 2008, 3: 3–14 porating the state pressure index for rockfill materials. J Eng Mech,
23 Xiao Y, Liu H, Desai C S, et al. Effect of intermediate principal-stress 2014, 140: 04014087
ratio on particle breakage of rockfill material. J Geotech Geoenviron 49 Xiao Y, Liu H L, Yang G, et al. A constitutive model for the state-de-
Eng, 2016, 142: 06015017 pendent behaviors of rockfill material considering particle breakage.
24 Xiao Y, Liu H, Ding X, et al. Influence of particle breakage on critical Sci China Tech Sci, 2014, 57: 1636–1646
state line of rockfill material. Int J Geomech, 2016, 16: 04015031 50 Xiao Y, Liu H L, Liu H, et al. Unified plastic modulus in the bounding
25 Xiao Y, Coop M R, Liu H, et al. Transitional behaviors in well- surface plasticity model. Sci China Tech Sci, 2016, 59: 932–940
graded coarse granular soils. J Geotech Geoenviron Eng, 2016, 142: 51 Pastor M, Zienkiewicz O C, Chan A H C. Generalized plasticity and
06016018 the modelling of soil behaviour. Int J Numer Anal Methods Geomech,
26 Rahman M M, Lo S C R, Dafalias Y F. Modelling the static lique- 1990, 14: 151–190
faction of sand with low-plasticity fines. Géotechnique, 2014, 64: 52 Ling H I, Yang S. Unified sand model based on the critical state and
881–894 generalized plasticity. J Eng Mech, 2006, 132: 1380–1391
27 Chang C S, Yin Z Y. Micromechanical modeling for behavior of silty 53 Liu H, Zou D. Associated generalized plasticity framework for mod-
sand with influence of fine content. Int J Solids Struct, 2011, 48: eling gravelly soils considering particle breakage. J Eng Mech, 2013,
2655–2667 139: 606–615
28 Lashkari A. Recommendations for extension and re-calibration of an 54 Xiao Y, Liu H, Chen Y, et al. Testing and modeling of the state-
existing sand constitutive model taking into account varying non-plas- dependent behaviors of rockfill material. Comput Geotech, 2014, 61:
tic fines content. Soil Dyn Earthquake Eng, 2014, 61-62: 212–238 153–165
29 Lashkari A. Prediction of flow liquefaction instability of clean and 55 Desai C S. Mechanics of Materials And Interfaces: The Disturbed
silty sands. Acta Geotech, 2016, 11: 987–1014 State Concept. Boca Raton: CRC Press, 2001
30 Li X S, Dafalias Y F. Dilatancy for cohesionless soils. Géotechnique, 56 Desai C S. Constitutive modeling for geologic materials: Significance
2000, 50: 449–460 and directions. Int J Geomech, 2005, 5: 81–84
31 Thevanayagam S. Effect of fines and confining stress on undrained 57 Desai C S, Chen J Y. Parameter optimization and sensitivity analysis
shear strength of silty sands. J Geotech Geoenviron Eng, 1998, 124: for disturbed state constitutive model. Int J Geomech, 2006, 6: 75–88
479–491 58 Desai C S. Unified DSC constitutive model for pavement materials
32 Chien L K, Oh Y N, Chang C H. Effects of fines content on liquefac- with numerical implementation. Int J Geomech, 2007, 7: 83–101
tion strength and dynamic settlement of reclaimed soil. Can Geotech 59 Desai C S, Sane S, Jenson J. Constitutive modeling including creep-
J, 2002, 39: 254–265 and rate-dependent behavior and testing of glacial tills for prediction
33 Belkhatir M, Arab A, Della N, et al. Influence of inter-granular void of motion of glaciers. Int J Geomech, 2011, 11: 465–476
ratio on monotonic and cyclic undrained shear response of sandy soils. 60 Desai C S. Constitutive modeling of materials and contacts using the
Comptes Rendus Mécanique, 2010, 338: 290–303 disturbed state concept: Part 2-Validations at specimen and boundary
34 Cabalar A F. The effects of fines on the behaviour of a sand mixture. value problem levels. Comput Struct, 2015, 146: 234–251
Geotech Geol Eng, 2011, 29: 91–100 61 Desai C S. Constitutive modeling of materials and contacts using the
35 Ni Q, Tan T S, Dasari G R, et al. Contribution of fines to the com- disturbed state concept: Part 1-Background and analysis. Comput
pressive strength of mixed soils. Géotechnique, 2004, 54: 561–569 Struct, 2015, 146: 214–233
36 Rahman M M, Lo S R. The prediction of equivalent granular steady 62 Yao Y P, Hou W, Zhou A N. UH model: Three-dimensional unified
state line of loose sand with fines. Geomech Geoeng, 2008, 3: hardening model for overconsolidated clays. Géotechnique, 2009, 59:
179–190 451–469
37 Rahman M M, Baki M A L, Lo S R. Prediction of undrained mono- 63 Yao Y P, Yang Y F, Niu L. UH model considering temperature effects.
tonic and cyclic liquefaction behavior of sand with fines based on Sci China Tech Sci, 2011, 54: 190–202
the equivalent granular state parameter. Int J Geomech, 2014, 14: 64 Yao Y P, Kong Y X. Extended UH model: Three-dimensional uni-
254–266 fied hardening model for anisotropic clays. J Eng Mech, 2012, 138:
38 Carraro J A H, Prezzi M, Salgado R. Shear strength and stiffness of 853–866
 Xiao Y, et al. Sci China Tech Sci 15

65 Yao Y P, Zhou A N. Non-isothermal unified hardening model: A assembly of particles in contact. Proc R Soc A-Math Phys Eng Sci,
thermo-elasto-plastic model for clays. Géotechnique, 2013, 63: 1962, 269: 500–527
1328–1345 72 Xiao Y, Desai C S. General stress-dilatancy relation for granular soils.
66 Xiao Y, Liu H L, Sun Y F, et al. Modeling of strength and deformation J Geotech Geoenviron Eng, 2016, 142: 02816001
of overconsolidated clays based on bounding surface plasticity. Sci 73 Xiao Y, Liu H, Liu H, et al. Strength and dilatancy behaviors of dense
China Tech Sci, 2016, 59: 1452–1462 modeled rockfill material in general stress space. Int J Geomech,
67 Xiao Y, Sun Y, Yin F, et al. Constitutive modeling for transparent 2016, 16: 04016015
granular soils. Int J Geomech, 2017, 17: 04016150 74 Xu M, Song E. Numerical simulation of the shear behavior of rock-
68 Xiao Y, Liu H. Elastoplastic constitutive model for rockfill materials fills. Comput Geotech, 2009, 36: 1259–1264
considering particle breakage. Int J Geomech, 2017, 17: 04016041 75 Yang J, Li X S. State-dependent strength of sands from the perspective
69 Hardin B O, Richart F E. Elastic wave velocities in granular soils. J of unified modeling. J Geotech Geoenviron Eng, 2004, 130: 186–198
Soil Mech Found, 1963, 89: 33–66 76 Wan R G, Guo P J. A simple constitutive model for granular soils:
70 Richart F E, Jr Hall J R, Woods R D. Vibration of Soils and Founda- Modified stress-dilatancy approach. Comput Geotech, 1998, 22:
tions. Upper Saddle River: Prentice-Hall, 1970 109–133
71 Rowe P W. The stress-dilatancy relation for static equilibrium of an