Anda di halaman 1dari 10

Geotextiles and Geomembranes 42 (2014) 329e338

Contents lists available at ScienceDirect

Geotextiles and Geomembranes


journal homepage: www.elsevier.com/locate/geotexmem

Modeling the pullout behavior of short fiber in reinforced soil


Hong-Hu Zhu 1, Cheng-Cheng Zhang 1, Chao-Sheng Tang*, Bin Shi 2, Bao-Jun Wang 3
School of Earth Sciences and Engineering, Nanjing University, Nanjing 210023, China

a r t i c l e i n f o a b s t r a c t

Article history: Fiber reinforcement is an effective method for improving engineering properties of soil. However, the
Received 31 May 2013 interaction mechanism of the fiber and the surrounding soil is not well understood. Based on mechanical
Received in revised form analysis of fiber-soil interface under pullout condition, a tri-linear model is proposed to describe the
11 May 2014
shear stress-displacement relationship. The progressive pullout process of a short fiber in soil is divided
Accepted 20 May 2014
into five consecutive phases: (1) the initial pure elastic phase (Phase I); (2) the elastic-softening phase
Available online 22 June 2014
(Phase II); (3) the pure softening phase (Phase III); (4) the softening-residual phase (Phase IV); and (5)
the final pure residual phase (Phase V). For each phase, the analytical solutions of the distributions of
Keywords:
Fiber reinforced soil
tensile force, interfacial shear stress and displacement are derived. Through a comparison between the
Pullout pullout test results of polypropylene fiber (PP-fiber) and the predicted results, the effectiveness of the
Stress softening proposed model in capturing the progressive load-deformation behavior of a short fiber in soil is verified.
Interaction mechanism Moreover, the effects of water content and dry density of soil on the model parameters are analyzed in
Progressive failure detail. It is found that the interfacial peak/residual shear resistance and shear stiffness of fiber reinforced
soil significantly depend on soil compaction conditions. In general, two transition phases (Phase II and
Phase IV) are not evident during the whole pullout process of PP-fiber.
© 2014 Elsevier Ltd. All rights reserved.

1. Introduction Michalowski and Zhao (1996) first proposed a close-form failure


criterion for fiber-reinforced sand using the energy-based homog-
The concept of reinforcing soils by introducing tension-resisting enization scheme. Both the fiber and the granular fill are assumed
elements, such as fibers, is becoming increasingly popular in to be perfectly plastic without considering the influence of
ground improvement engineering. Previous experimental studies confining stress on the fiber tensile strength. Zornberg (2002)
on fiber reinforced soils have shown significant increase of cohe- established a discrete framework to predict the equivalent shear
sion and friction angle (e.g., Maher and Gray, 1990; Consoli et al., strength parameters based on the independent properties of fiber
1998, 2009; Michalowski and Cermark, 2003; Yetimoglu and and soil. A theoretical model was proposed by Rifai and Miller
Salbas, 2003; Yetimoglu et al., 2005; Cai et al., 2006; Tang et al., (2009) to quantitatively describe the contribution of randomly
2007), reduction of desiccation cracking (Consoli et al., 2003; distributed fiber to cracking reduction in soil undergoing desicca-
Miller and Rifai, 2004; Tang et al., 2012), improvement of hydrau- tion. In this model, the Mohr-Coulomb failure criterion is used to
lic conductivity (Miller and Rifai, 2004), mitigation of liquefaction model the fiber-soil interface behavior. Taking the fiber orientation
risk (Ibraim et al., 2010; Liu et al., 2011), and improvement of piping and strain rate/stress direction at failure into consideration, Gao
resistance of hydraulic structures (Estabragh et al., 2014). In and Zhao (2013) presented a three-dimensional anisotropic fail-
modeling the performance of fiber reinforced soils, most traditional ure criterion, and verified their findings using laboratory test
approaches assume that the fiber reinforced soil is a composite results.
material with improved properties from a macroscopic scale. Generally, the pullout failure of discrete fiber from soil matrix as
subjected to external load is recognized as one of the dominant
failure mode of fiber reinforced soils. The interaction mechanism
* Corresponding author. Tel.: þ86 25 83597888; fax: þ86 25 89680137.
between a fiber under pullout condition and the surrounding soil is
E-mail addresses: zhh@nju.edu.cn (H.-H. Zhu), zhangchengcheng@gmail.com therefore of great importance. In order to quantify the interfacial
(C.-C. Zhang), tangchaosheng@nju.edu.cn (C.-S. Tang), shibin@nju.edu.cn (B. Shi), shear strength, Tang et al. (2010) conducted a series of single fiber
baojun@nju.edu.cn (B.-J. Wang). pullout tests using a modified testing apparatus. Soil water content,
1
Tel.: þ86 25 83597888.
2 dry density and cement content were varied in these tests and the
Tel.: þ86 25 89680137.
3
Tel.: þ86 25 89686194. results show that both the peak and the residual strengths are

http://dx.doi.org/10.1016/j.geotexmem.2014.05.005
0266-1144/© 2014 Elsevier Ltd. All rights reserved.
330 H.-H. Zhu et al. / Geotextiles and Geomembranes 42 (2014) 329e338

influenced by these factors. More recently, Li and Zornberg (2013)


and Hejazi et al. (2013) performed fiber pullout tests to study the
mobilization of reinforcement forces due to the inclusion of straight
fiber and loop-formed fiber, respectively. The slippage theory,
together with an artificial neural network technique, was utilized
by Hejazi et al. (2013) to interpret the experimental data. However,
the literature review shows that few attempts have been made to
develop a comprehensive theoretical model to describe the fiber-
soil interaction.
In this study, the interaction between a fiber and the sur-
rounding soil is systematically investigated and the pullout process
of a short fiber in soil is divided into five consecutive phases. A
theoretical model based on the tri-linear interfacial shear stress-
displacement relationship was derived and the distributions of
tensile force, interfacial shear stress and displacement were ob- Fig. 2. Proposed model defining the relationship between shear stress and shear
displacement at the fiber-soil interface.
tained, as well as the pullout force-displacement relationship. This
model was used to interpret the pullout test results of poly- Based on the above assumptions, the progressive pullout behavior
propylene fiber (PP-fiber). Some conclusions have been made on of a short fiber in soil can be divided into five consecutive phases, as
the influence of water content and dry density of soil on the model shown in Fig. 3. These five phases are described as following.
parameters.
(1) Initial pure elastic phase (Phase I): When a relatively small
pullout force is applied on the fiber head, the mobilized
2. Mechanical analysis of fiber-soil interface interfacial shear stress follows a linear relationship with
respect to the shear displacement. Neither stress softening
An illustrative model of the pullout of a fiber in soil is shown in nor debonding occurs in this phase.
Fig. 1. The fiber is idealized as a cylindrical rod with a diameter D (2) Elastic-softening phase (Phase II): The fiber-soil interface
along the fiber length L. During pullout, the fiber is assumed to be remains elastic until the shear stress reaches the peak shear
an axially loaded tension member while radial deformation is resistance at the fiber head, from which stress softening
neglected. When a pullout force F0 is applied, shear stresses are initiates and propagates to the fiber tail. As a result, a tran-
mobilized at the fiber-soil interface to resist the pullout force. sition point P1 (x ¼ Ls), as shown in Fig. 3(b), is introduced
A tri-linear model is adopted to quantify the shear stress- here to divide the elastic and softening zones (Misra et al.,
displacement relationship of the fiber-soil interface, as depicted 2004). Note that the interfacial shear stress at this point is
in Fig. 2. In this model, the fiber-soil interface first behaves elasti- equivalent to the peak shear resistance.
cally, which is characterized by an ascending branch up to the peak (3) Pure softening phase (Phase III): As the softening zone ex-
shear resistance. Afterward, stress softening emerges at the fiber- tends towards the fiber tail, the interfacial shear stress suc-
soil interface. Once the interfacial shear stress decreases to the cessively reaches the peak shear resistance, and gradually
residual shear resistance, the interface is entirely debonded and the decreases thereafter. Once the shear stress at the fiber tail
shear stress remains constant. The relationship between shear increases to the peak shear resistance, the softening zone
stress tðxÞ and shear displacement uðxÞ can be expressed by occupies the entire fiber, as shown in Fig. 3(c). Both the
8 pullout force and the interfacial shear stress decrease in this
< GuðxÞ ð0  u < u1 Þ ð1aÞ stage, while the shear displacement continues to increase.
tðxÞ ¼ 2tmax  GuðxÞ ðu1  u < u2 Þ ð1bÞ (1) (4) Softening-residual phase (Phase IV): Similar to Phase II, once
:
tres ðu  u2 Þ ð1cÞ the interfacial shear stress at the fiber head decreases to the
residual shear resistance, the fiber turns into the softening-
where G ¼ shear stiffness at the fiber-soil interface that should be
residual transition state. Again, a transition point P2 (x ¼ Lr)
determined experimentally (Unit: Pa/m); u1 and u2 ¼ shear dis-
is introduced here to divide the softening and residual zones,
placements corresponding to the peak shear resistance tmax and
as shown in Fig. 3(d). In this phase, both the pullout force and
the residual shear resistance tres , respectively. Considering that
the interfacial shear stress decrease slightly.
u1 ¼ tmax =G and u2 ¼ ð2tmax  tres Þ=G, there are totally three in-
(5) Final pure residual phase (Phase V): The final stage starts
dependent model parameters, namely, G, tmax and tres . It should be
when the interfacial shear stress at the fiber tail decreases to
noted that these parameters are influenced not only by the physical
the residual shear resistance. As shown in Fig. 3(e), the re-
and mechanical properties of soil, but also by the fiber properties,
sidual zone now occupies the entire fiber. In this stage, the
such as surface roughness.
pullout force remains constant, whereas the pullout
displacement increases continuously.

3. Formulation of the fiber pullout model

The following derivation uses the coordinate system shown in


Fig. 1. As the fiber is assumed to be elastic throughout the pullout
process, we have
Fig. 1. Schematic illustration of pullout mechanism of a fiber in soil. The fiber is p 2
assumed to be a cylindrical rod. During pullout, the fiber is assumed to be an axially FðxÞ ¼ D EεðxÞ (2)
4
loaded tension member while radial deformation is neglected.
H.-H. Zhu et al. / Geotextiles and Geomembranes 42 (2014) 329e338 331

Fig. 3. Progressive pullout process of a short fiber in soil in the analytical model proposed in this study.

in which FðxÞ ¼ tensile force in the fiber (tension is considered to be and the closed-form solutions can be derived for all the pullout
positive); E ¼ Young's modulus of the fiber; and εðxÞ ¼ uniaxial phases.
strain of the fiber. Because εðxÞ can be written in terms of shear
displacement uðxÞ (displacement in the pullout direction is taken as 3.1. Pure elastic phase
positive), Eq. (2) can be expressed as
In this phase, the relationship between the interfacial shear
p duðxÞ stress and the shear displacement can be expressed by Eq. (1a). By
FðxÞ ¼  D2 E (3) combining Eqs. (1a), (3) and (4), we get
4 dx
According to the mechanical model mentioned above, the d2 F
resistance of the fiber against the pullout force is provided by the  a2 F ¼ 0 (5)
dx2
shear stress acting on the fiber-soil interface. Hence, the following
differential equation can be derived from the equilibrium of a where
uniaxial fiber element (Sawicki, 1998, 2000) rffiffiffiffiffiffi
4G
a¼ (6)
dFðxÞ DE
¼ pDtðxÞ (4)
dx Note that a depends on the diameter and the Young’s modulus
Given specified boundary conditions, Eqs. (1), (3) and (4) are of the fiber and the shear stiffness of fiber-soil interface (Unit: m1).
sufficient to determine the three unknowns (FðxÞ, tðxÞ, and uðxÞ), A general solution of Eq. (5) is
332 H.-H. Zhu et al. / Geotextiles and Geomembranes 42 (2014) 329e338

Fe ðxÞ ¼ C1 eax þ C2 eax (7) pDtmax tanh aðL  Ls Þ


F00 ¼ (17)
a
where Fe ðxÞ ¼ tensile force in fiber in the pure elastic phase; C1 and
In the softening zone (0  x  Ls ), the relationship between the
C2 ¼ integration constants, which can be determined with specific
interfacial shear stress and the shear displacement is defined by Eq.
boundary conditions.
(1b). Combining Eqs. (1b), (3) and (4), we get
In a fiber pullout test, the pullout force F0 applied on the fiber
head can be measured by a load cell, and the tensile force at the
fiber tail can be neglected. Thus we have d2 F
þ a2 F ¼ 0 (18)
dx2

Fe ð0Þ ¼ F0 A general solution of Eq. (18) is
(8)
Fe ðLÞ ¼ 0
Fs ðxÞ ¼ C3 cos bx þ C4 sin bx (19)
The solution of the boundary-value problem defined by Eqs. (7)
and (8) is
where C3 and C4 ¼ integration constants. Similarly, the boundary
conditions are
sinh aðL  xÞ
Fe ðxÞ ¼ F0 (9) 
sinh aL Fs ð0Þ ¼ F0
(20)
The corresponding interfacial shear stress te ðxÞ and shear Fs ðLs Þ ¼ Fe ðLs Þ
displacement ue ðxÞ are then calculated using Eqs. (4) and (1a),
So we get
respectively, which can be expressed by
8
aF cosh aðL  xÞ < C3 ¼ F0
>
te ðxÞ ¼ 0 (10) (21)
pD sinh aL > pDtmax tanh aðL  Ls Þ
: C4 ¼  F0 cot aLs
a sin aLs
aF0 cosh aðL  xÞ
ue ðxÞ ¼ (11) Again, considering the continuity for interfacial shear stress at
pDG sinh aL the transition point P1 , the relationship between the pullout force
Let x ¼ 0 in Eq. (11), we get and the length of the softening zone is derived as

pDG tanh aL pDtmax


F0 ¼ ue0 (12) F0 ¼ ½sin aLs þ tanh aðL  Ls Þcos aLs  (22)
a a
Accordingly, the solutions for the softening zone are
in which ue0 ¼ shear displacement at the fiber head, i.e. the pullout
displacement. It is observed that there is a linear relationship be-
pDtmax
tween the pullout force and the shear displacement in this phase. Fs ðxÞ ¼ ½sin aðLs  xÞ þ tanh aðL  Ls Þcos aðLs  xÞ
a
As the shear displacement reaches u1 , Eq. (12) becomes
(23)
c pDtmax tanh aL
Fes0 ¼ (13) ts ðxÞ ¼ tmax ½cos aðLs  xÞ  tanh aðL  Ls Þsin aðLs  xÞ (24)
a
c ¼ the critical pullout force between the pure elastic and
where Fes0
2tmax tmax
the elastic-softening phases, which is also the maximum pullout us ðxÞ ¼  ½cos aðLs  xÞ  tanh aðL  Ls Þsin aðLs  xÞ
G G
force in the pure elastic phase.
(25)
For the pullout of a reinforcement element in soil, the equation
3.2. Elastic-softening phase Fu ¼ pDLtmax is widely adopted to calculate the pullout capacity
during the whole pullout process (e.g. Yin et al., 2012; Zhu et al.,
In this phase, the distributions of tensile force, interfacial shear 2011). However, this equation is not valid here, as it over-
stress and shear displacement in the elastic zone (Ls  x  L) are estimates the real pullout capacity of a fiber in soil. The peak
similar to those in the pure elastic phase. Therefore, the following pullout resistance Fu can be calculated when dF0
dLs
¼ 0. Combining it
equations are easily obtained, with Eq. (23), we get

sinh aðL  xÞ tanh aðL  Ls Þ ¼ tan aLs (26)


Fe ðxÞ ¼ F 0 0 (14)
sinh aðL  Ls Þ Let b ¼ aL and t ¼ Ls
L, Eq. (26) turns into

aF 0 0 cosh aðL  xÞ tanh bð1  tÞ ¼ tan bt (27)


te ðxÞ ¼ (15)
pD sinh aðL  Ls Þ To the best of the authors’ knowledge, no exact solution is
available for Eq. (27). In this study, this equation is approximated
aF 0 0 cosh aðL  xÞ through evaluating its Taylor series expansion. It is found that when
ue ðxÞ ¼ (16)
pDG sinh aðL  Ls Þ b ¼ aL  1, t can be approximated as 0.5 with the relative error less
than 10%. That means, given a relatively low a value, the pullout
where Ls ¼ length of the softening zone; and F 0 0 ¼ tensile force at force is expected to peak when the softening zone occupies about
the transition point P1 . Considering that the interfacial shear stress half length of the fiber. However, when b is relatively large (b > 1), t
at P1 equals the peak shear resistance, we get will be much less than 0.5.
H.-H. Zhu et al. / Geotextiles and Geomembranes 42 (2014) 329e338 333

Let Ls ¼ L in Eq. (22), the critical pullout force between the Table 2
elastic-softening and the pure softening phases is given by Physical and mechanical parameters of silty clay.

Specific Liquid Plasticity Optimum moisture Maximum dry


c pDtmax sin aL gravity Gs limit wL (%) index Ip content wopt (%) density rd (g/cm3)
Fs0 ¼ (28)
a 2.71 36.4 17.8 16.5 1.7

3.3. Pure softening phase Table 3


Physical and mechanical parameters of polypropylene fiber.
The interfacial shear stress of the fiber tail increases to the peak
c . As long as the Unit weight Diameter Length Breaking tensile Elongation Modulus of
shear resistance when the pullout force reaches Fs0 r (g/cm3) (mm) (mm) strength (MPa) at break (%) elasticity (GPa)
interfacial shear stress at the fiber head is higher than the residual
0.91 0.048 150 350 ± 20 16e20 3.5
shear resistance, the entire fiber is occupied by the softening zone. The
pure softening phase exists only when ts ðxÞLs ¼L  tres is satisfied, i.e.,
3.4. Softening-residual phase
arccos k
L  Lc ¼ (29)
a c , the residual state initiates
As the pullout force decreases to Fsr0
at the fiber head and extends towards the fiber tail. The residual
where k ¼ tres =tmax ; and Lc ¼ a critical length for a fiber under and softening zones are divided by the transition point P2 . In the
pullout condition. Eq. (29) can be used to mechanically determine softening zone, the distributions of tensile force, interfacial shear
whether a fiber is “short enough” to achieve the pure softening stress, and shear displacement are given by Eqs. (31)e(33),
phase for fiber reinforced soil. With respect to the new definition, respectively, with x replaced by ðx  Lr Þ, L by ðL  Lr Þ, and F0 by F00
all the theoretical and experimental works in the present study are
for this type of “short” fiber. Theoretically, if the fiber is long enough Fs ðxÞ ¼ F 0 0 ½cos aðx  Lr Þ  cot aðL  Lr Þsin aðx  Lr Þ (35)
(L  Lc ), there will be three types of stress state within the entire
fiber in Phase III, namely the residual state, the softening state, and
aF 0 0
the elastic state. However, as longer fibers are much easier to bend, ts ðxÞ ¼ ½sin aðx  Lr Þ þ cot aðL  Lr Þcos aðx  Lr Þ (36)
distort and twist, the interaction between a long fiber and soil will
pD
be intractably complicated. On the other hand, the dominant failure
pattern of long fiber reinforced soil will be the breaking of fiber. 2tmax aF 0 0
us ðxÞ ¼  ½sin aðx  Lr Þ þ cot aðL  Lr Þcos aðx  Lr Þ
Therefore, the present study only considers the interaction be- G pDG
tween a short fiber and soil under pullout condition. (37)
Similar to the analysis of the softening zone in the elastic-
softening phase, Eq. (18) is still applicable in the pure softening
phase. The boundary conditions are

Fs ð0Þ ¼ F0
(30)
Fs ðLÞ ¼ 0

The solutions for the pure softening phase are

Fs ðxÞ ¼ F0 ðcos ax  cot aL sin axÞ (31)

aF0
ts ðxÞ ¼ ðsin ax þ cot aL cos axÞ (32)
pD

2tmax aF0
us ðxÞ ¼  ðsin ax þ cot aL cos axÞ (33)
G pDG
In the pure softening phase, both the pullout force and the
interfacial shear stress decrease. As the interfacial shear stress at
the fiber head decreases to the residual shear resistance, the pullout
force applied on the fiber head is given as

c pDtres tan aL
Fsr0 ¼ (34)
a
c ¼ minimum pullout force in the pure softening phase,
where Fsr0
i.e. the critical pullout force between the pure softening and
softening-residual phases.

Table 1
Critical pullout forces of five progressive phases in the analytical fiber pullout model
proposed in this study.

Pullout phase IeII IIeIII IIIeIV IVeV Fig. 4. Setup of the single fiber pullout tests performed by Tang et al. (2010). An
c c c c electronic balance and a displacement transducer were used to measure the applied
Critical pullout force Fes0 Fs0 Fsr0 Fr0
pullout force and the pullout displacement, respectively. The pullout displacement rate
Expression pDtmax tanh aL pDtmax sin aL pDtres tan aL pDtres L
a a a was controlled at 1.0 mm/min.
334 H.-H. Zhu et al. / Geotextiles and Geomembranes 42 (2014) 329e338

Because the interfacial shear stress at P2 equals the residual  


shear resistance, F 0 0 in the above three equations can be re-written 2tres 2 4tres 4tres
ur ðxÞ ¼ x  tan aðL  Lr Þ þ Lr x þ C5 (42)
in terms of Ls , i.e. DE aDE DE

pDtres tan aLs where


F0' ¼ (38)
a
2tres L2r 4tres Lr 2tmax  tres
As for the residual zone, debonding and slippage occur at the C5 ¼ þ tan aðL  Lr Þ þ (43)
DE aDE G
fiber-soil interface, and the interfacial shear stress equals to the
residual shear resistance. Eqs. (1c) and (4) lead to Let x ¼ 0 and Lr ¼ L in Eq. (41), the critical pullout force between
the softening-residual and the pure residual phases is given by
dF
¼ pDtres (39) c
Fr0 ¼ pDtres L (44)
dx
Combining it with the boundary conditions

Fr ð0Þ ¼ F0
(40) 3.5. Pure residual phase
Fr ðLr Þ ¼ Fs ðLr Þ
In the last pullout phase, the fiber-soil interface is completely
the distribution of tensile force is obtained by integrating Eq. (39),
debonded, and the pullout capacity is solely due to the sliding
i.e.
friction. Let x ¼ 0 and Lr ¼ L in Eq. (42), the critical pullout
pDtres displacement at the fiber head is
Fr ðxÞ ¼ pDtres x þ tan aðL  Lr Þ þ pDtres Lr (41)
a
2tres L2 2tmax  tres
Substituting Eq. (41) into Eq. (3) and considering ur ðLr Þ ¼ us ðLr Þ at uc0 ¼ þ (45)
DE G
P2 , the shear displacement in the residual zone can be expressed by

Fig. 5. Comparisons between predicted results and test results with different water content at the soil dry density of 1.7 g/cm3 (data from Tang et al., 2010).
H.-H. Zhu et al. / Geotextiles and Geomembranes 42 (2014) 329e338 335

Assuming that the displacement of fiber head is known as u0 0 , were employed here to do simulation. The silty clay soil used in
the distribution of displacement is derived as the tests was obtained from a site in Nanjing, China. The
commercially available PP-fiber was selected as the test fiber. The
2tres  2  4t L tres  2tmax physical and mechanical properties of the soil and the fiber are
þ ðu0 0  u0 Þ
res
uðxÞ ¼ x  L2  ðx  LÞ 
DE DE G listed in Table 2 and Table 3, respectively.
2tres  2  As shown in Fig. 4, the test specimens were designed to be
¼ u0 0 þ x  2Lx
DE 5  5  5 mm in size to avoid breakage during the pullout tests.
(46) After compaction in the mould, the specimens were extracted and
laid on the self-designed pullout test machine. An electronic bal-
The pullout force and interfacial shear stress remain constants ance and a displacement transducer were used to measure the
and equal Fr0c and t , respectively. Therefore, the distribution of
res applied pullout force and the pullout displacement, respectively.
tensile force is given as The pullout displacement rate was controlled at 1.0 mm/min. More
details about the test setup can be found in Tang et al. (2010).
Fr ðxÞ ¼ pDtres ðL  xÞ (47)
The closed-form solutions are all derived now for all the five
4.2. Predicted results using the proposed model
pullout phases. The five critical pullout forces are summarized in
Table 1.
Based on the proposed model, the analytical solutions were
obtained to simulate the pullout of a short fiber in soil. The pa-
4. Analysis of fiber pullout test results rameters used in the simulation are as follows: L ¼ 5 mm,
D ¼ 0:048 mm, and E ¼ 3:5 GPa. The remaining three parameters,
4.1. Laboratory fiber pullout tests i.e. peak shear resistance tmax , residual shear resistance tres , and
shear stiffness G at the fiber-soil interface, which vary with water
To verify the above theoretical model and to get a better un- content and dry density of soil, were obtained using curving fitting
derstanding of the influential factors of model parameters, the technique. The critical fiber length Lc ranges from 21.15 mm to
published results of single fiber pullout tests of Tang et al. (2010) 55.52 mm, which is much longer than the fiber length L (5 mm) in

Fig. 6. Comparisons between predicted results and test results with different soil dry density at the water content of 16.5% (data from Tang et al., 2010).
336 H.-H. Zhu et al. / Geotextiles and Geomembranes 42 (2014) 329e338

Fig. 7. Evolution of interfacial shear stress along the fiber length during the pullout process. In Phase I and Phase III, the pattern of the interfacial shear stress distribution remains
unchanged. Phase II and Phase IV are two transition phases. The final pure residual phase where the interfacial shear stress equals the residual shear resistance is not presented. The
results given in the figures correspond to the test conducted at water content ¼ 20.5% and soil dry density ¼ 1.7 g/cm3.

the tests. The predicted results are shown in Figs. 5e8. It should be
noted that in the fiber pullout tests, there was a 50-mm free fiber
between the specimen and the electronic balance, as shown in
Fig. 4. The extension of the 50-mm free fiber is deducted and the
pullout displacement presented here is thus the relative fiber-soil
displacement, while the results shown in Tang et al. (2010) refers
to the displacement of the total fiber.
Figs. 5e6 show the fitted curves of pullout force versus
displacement. The stress softening phenomena and an approxi-
mately horizontal residual branch are observed from the test re-
sults. It is clear that the predicted results agree well with those
obtained from the pullout tests, indicating that the proposed
pullout model can effectively describe the progressive pullout
behavior of a short fiber in soil. Besides, the displacements in Phase
II and Phase IV are relatively small compared with those in Phase I
and Phase III. Therefore, the two transition phases can be neglected
Fig. 8. Contribution of three zones to the pullout force during the five pullout phases. when analyzing the pullout behavior of a short PP-fiber in soil.
The elastic zone, the softening zone and the residual zone contributes entirely to the In addition to the fitting results, the following study is also
pullout force in Phase I, III and V, respectively. Phase II and Phase IV are two transition presented to better understand the five pullout phases. The simu-
phases. In Phase II, the contribution of the elastic zone gradually decreases to 0% while
that of the softening zone gradually increases to 100%. The contributions of the soft-
lated results given in Figs. 7e8 correspond to the test conducted at
ening and residual zones in Phase IV are similar. The test condition is the same as that water content ¼ 20.5% and dry density ¼ 1.7 g/cm3. Evolution of the
in Fig. 7. interfacial shear stress is presented in Fig. 7. In Phase I and Phase III,
H.-H. Zhu et al. / Geotextiles and Geomembranes 42 (2014) 329e338 337

the interfacial shear stress varies with the pullout force but the the variations of the three model parameters with soil water con-
distribution pattern remains unchanged. In Phase II, the interfacial tent. It can be observed that both the peak and residual shear
shear stress starts to decrease once reaching the peak shear resis- resistance are approximately proportional to the water content. The
tance (147.0 kPa), indicating that stress softening gradually prop- decreasing rate of the residual shear resistance with respect to the
agates from the fiber head to the fiber tail. In Phase IV, the water content is higher than that of the peak shear resistance.
interfacial shear stress remains constant if the residual shear When the water content is 14.5%, the residual shear resistance is
resistance (48.0 kPa) is reached. Finally, the interfacial shear stress 67.2% of the peak shear resistance. When the water content in-
along the fiber equals the residual shear resistance and Phase V is creases to 20.5%, this value decreases to 32.7%. However, for shear
arrived. Fig. 8 illustrates the contributions of three components stiffness of the fiber-soil interface, it increases slightly with an in-
(due to elastic zone, softening zone and residual zone) to the crease in water content. One possible explanation for this phe-
pullout force and their variations in the five pullout phases. The nomenon is that the water acting as lubricating layer results in a
elastic zone, the softening zone and the residual zone contributes decrease in the interfacial strength. On the other hand, the test
entirely to the pullout force in Phase I, III and V, respectively. Phase results indicate that this effect may reduce the shear deformation of
II and Phase IV are two transition phases. In Phase II, the contri- the surrounding soil and thus increase the stiffness of the fiber-soil
bution of the elastic zone gradually decreases to 0% while that of interface. In addition, the optimum moisture content seems not
the softening zone gradually increases to 100%. The contributions of play an important role in governing the fiber-soil interaction.
the softening and residual zones in Phase IV are similar. Fig. 10 shows the effect of soil dry density on the model pa-
rameters. The peak shear resistance, residual shear resistance and
shear stiffness are observed to increase with the increase of soil dry
4.3. Influential factors of model parameters density. The peak shear resistance at soil dry density of 1.7 g/cm3 is
about three times higher than that at soil dry density of 1.4 g/cm3.
The effects of two factors, including water content and dry For the residual shear resistance, it is approximately five times
density of soil, on the model parameters are examined. Fig. 9 shows larger. As water content and dry density of soil are two key

Fig. 9. Correlation between model parameters and water content with the soil dry Fig. 10. Correlation between model parameters and soil dry density with the water
density of 1.7 g/cm3. content of 16.5%.
338 H.-H. Zhu et al. / Geotextiles and Geomembranes 42 (2014) 329e338

parameters for ground improvement works, it may be concluded compaction conditions (i.e. water content and dry density of
that, comparing with the peak shear resistance, the residual shear soil).
resistance is more sensitive to soil compaction conditions.

5. Discussion Acknowledgments

The proposed analytical model in this work is a simplified model The financial supports provided by the National Basic Research
derived based on a series of assumptions: (1) the fiber is straight Program of China (973 Program) (Grant No. 2011CB710605), the
and considered as a pure axially loaded tension member; (2) the National Natural Science Foundation of China (Grant Nos.
pullout load is always along the fiber longitude direction; (3) the 41302217; 41322019; 41230636; 41072211) are gratefully
fiber-soil interfacial mechanical behavior is described by the tri- acknowledged. The authors would like to express their sincere
linear model. Therefore, it can be used to interpret the experi- thanks to the two anonymous reviewers for their constructive
mental data of Tang et al. (2010), where the test conditions are in comments and suggestions.
ideal situations as the above assumptions. It is apparent that the
tests did not consider many factors, such as the effect of confine-
References
ment around the fiber.
Although the model predictions are consistent with the single Cai, Y., Shi, B., Ng, C.W.W., Tang, C.S., 2006. Effect of polypropylene fibre and lime
fiber pullout test results, the applicability of this model to the real admixture on engineering properties of clayey soil. Eng. Geol. 87 (3e4),
case of fiber reinforced soils, where a large number of fibers are 230e240.
Consoli, N.C., Prietto, D.M.P., Ulbrich, L.A., 1998. Influence of fiber and cement
randomly distributed in soil matrix, is to be further investigated. In addition on behavior of sandy soil. J. Geotechn. Geoenviron. Eng. 124 (12),
the discrete framework proposed by Zornberg (2002) for limit 1211e1214.
equilibrium analysis of fiber reinforced soil, the contributions of soil Consoli, N.C., Vendruscolo, M.A., Fonini, A., Rosa, F.D., 2009. Fiber reinforcement
effects on sand considering a wide cementation range. Geotext. Geomembr. 27
and fibers are treated separately. The equivalent shear strength of a (3), 196e203.
fiber reinforced soil specimen is defined as the fiber-induced ten- Consoli, N.C., Vendruscolo, M.A., Prietto, P.D.M., 2003. Behavior of plate load tests on
sion plus the shear strength of the unreinforced soil. Independent soil layers improved with cement and fiber. J. Geotech. Geoenviron. Eng. 129 (1),
96e101.
evaluation of the strength of soil and fiber, instead of the fiber
Estabragh, A.R., Soltannajad, K., Javadi, A.A., 2014. Improving piping resistance using
reinforced soil specimen, are therefore needed. The proposed randomly distributed fibers. Geotext. Geomembr. 42 (1), 15e24.
methodology somewhat facilitates the implementation of fiber Gao, Z.W., Zhao, J.D., 2013. Evaluation on failure of fiber-reinforced sand. J. Geotech.
reinforcement in engineering practice. Besides Tang et al. (2010), Li Geoenviron. Eng. 139 (1), 95e106.
Hejazi, S.M., Sheikhzadeh, M., Abtahi, S.M., Zadhoush, A., 2013. Shear modeling of
and Zornberg (2013) designed another pullout testing apparatus to fiber reinforced soil composite on the base of fiber pull-out test. Fibers Polym.
quantify the interfacial shear strength between fibers and soil, 14 (2), 277e284.
which is a required part of the proposed discrete framework of Ibraim, E., Diambra, A., Wood, D.M., Russell, A.R., 2010. Static liquefaction of fibre
reinforced sand under monotonic loading. Geotext. Geomembr. 28 (4),
Zornberg (2002). It should be noted that, the fiber-soil interaction 374e385.
are highly complicated and conditioned by many factors. This work Li, C.L., Zornberg, J.G., 2013. Mobilization of reinforcement forces in fiber-reinforced
is the first step towards the solution of this issue. More refined soil. J. Geotech. Geoenviron. Eng. 139 (1), 107e115.
Liu, J., Wang, G.H., Kamai, T., Zhang, F.Y., Yang, J., Shi, B., 2011. Static liquefaction
experimental and theoretical investigations should be carried out behaviour of saturated fiber-reinforced sand in undrained ring-shear tests.
to provide a comprehensive picture of the macro/micro-mechanical Geotext. Geomembr. 29 (5), 98e107.
behavior of fiber reinforced soil. Maher, M.H., Gray, D.H., 1990. Static response of sands reinforced with randomly
distributed fibers. J. Geotech. Eng. 116 (11), 1661e1677.
Michalowski, R.L., Cerm 
ak, J., 2003. Triaxial compression of sand reinforced with
6. Conclusions fibers. J. Geotech. Geoenviron.Eng. 129 (2), 125e136.
Michalowski, R.L., Zhao, A.G., 1996. Failure of fiber-reinforced granular soils.
J. Geotech. Eng. 122 (3), 226e234.
In this paper, a tri-linear model is proposed to describe the shear
Miller, C.J., Rifai, S.M., 2004. Fiber reinforcement for waste containment soil liner.
stress-displacement relationship of fiber-soil interface. The full- J. Environ. Eng. 130 (8), 891e895.
range pullout process of a short fiber in soil is divided into five Misra, A., Chen, C.H., Oberoi, R., Kleiber, A., 2004. Simplified analysis method for
consecutive phases, and the distributions of tensile force, interfacial micropile pullout behavior. J. Geotech. Geoenviron. Eng. 130 (10), 1024e1033.
Rifai, S.M., Miller, C.J., 2009. Theoretical assessment of increased tensile strength of
shear stress and displacement, as well as the pullout force- fibrous soil undergoing desiccation. J. Geotech. Eng. 135 (12), 1857e1862.
displacement relationship, are derived. In addition, the effects of Sawicki, A., 1998. Modelling of geosynthetic reinforcement in soil retaining walls.
water content and dry density of soil on the model parameters are Geosynth. Int. 5 (3), 327e345.
Sawicki, A., 2000. Mechanics of Reinforced Soil. Balkema, Rotterdam, Netherlands.
investigated in detail. The main conclusions of the present work Tang, C.S., Shi, B., Cui, Y.J., Liu, C., Gu, K., 2012. Desiccation cracking behavior of
can be summarized as follows: polypropylene fiber reinforced clayey soil. Can. Geotech. J. 49 (9), 1088e1101.
Tang, C.S., Shi, B., Gao, W., Chen, F., Cai, Y., 2007. Strength and mechanical behavior
of short polypropylene fiber reinforced and cement stabilized clayey soil.
(1) The pullout process of a short fiber in soil can be divided into Geotext. Geomembr. 25 (2), 194e202.
five phases, i.e. pure elastic, elastic-softening, pure softening, Tang, C.S., Shi, B., Zhao, L.Z., 2010. Interfacial shear strength of fiber reinforced soil.
softening-residual, and pure residual phases. Geotext. Geomembr. 28 (1), 54e62.
Yetimoglu, T., Inanir, M., Inanir, O.E., 2005. A study on bearing capacity of randomly
(2) Through a comparison between the predicted results and the
distributed fiber-reinforced sand fills overlying soft clay. Geotext. Geomembr.
results obtained from PP-fiber pullout tests, the effectiveness 23 (2), 174e183.
of the proposed pullout model in capturing the progressive Yetimoglu, T., Salbas, O., 2003. A study on shear strength of sands reinforced with
randomly distributed discrete fibers. Geotext. Geomembr. 21 (2), 103e110.
loadedeformation behavior of a short fiber in soil is verified.
Yin, J.H., Hong, C.Y., Zhou, W.H., 2012. Simplified analytical method for calculating
(3) During the whole pullout process of a PP-fiber, the two the maximum shear stress of nail-soil interface. Int. J. Geomech. 12 (3),
transition phases (Phase II and Phase IV) are less evident 309e317.
compared with the other phases. Zhu, H.H., Yin, J.H., Yeung, A.T., Jin, W., 2011. Field pullout testing and performance
evaluation of GFRP soil nails. J. Geotech. Geoenviron. Eng. 137 (7), 633e641.
(4) The interfacial peak/residual shear resistance and shear Zornberg, J.G., 2002. Discrete framework for limit equilibrium analysis of fibre-
stiffness of fiber reinforced soil significantly depend on soil reinforced soil. Ge otechnique 52 (8), 593e604.

Anda mungkin juga menyukai