Geosynthetics International, 2017, 24, No.

Liquefaction resistance of sand reinforced with randomly

distributed polypropylene fibres
B. Ye1,2, Z. R. Cheng3, C. Liu4, Y. D. Zhang5 and P. Lu6
1
Associate Professor, Department of Geotechnical Engineering, Tongji University, Shanghai 200092,
China, E-mail: [email protected] (corresponding author)
2
Research Scholar, State Key Laboratory of Geohazard Prevention and Geoenvironment Protection,
Chengdu 610059, China, E-mail: [email protected]
3
Graduate Student, Department of Geotechnical Engineering, Tongji University, Shanghai, 200092,
China, E-mail: [email protected]
4
Graduate Student, Department of Geotechnical Engineering, Tongji University, Shanghai, 200092,
China, E-mail: [email protected]
5
Graduate Student, Department of Geotechnical Engineering, Tongji University, Shanghai, 200092,
China, E-mail: [email protected]
6
Assistant Professor, College of Surveying and Geo-informatics, Tongji University, Shanghai 200092,
China, E-mail: [email protected]

Received 23 December 2016, revised 05 April 2017, accepted 17 September 2017, published
15 November 2017

ABSTRACT: This paper presents an experimental study on the effect of randomly distributed fibres
on enhancing the liquefaction resistance of sand. A series of undrained cyclic triaxial compression tests
and hollow cylinder torsional shear tests were conducted on saturated sand samples with and without
fibres. The influencing factors (including fibre length, fibre content and relative density), which are
closely related to the liquefaction resistance of sand, were investigated. The test results indicated that
samples reinforced with randomly distributed fibres can sustain many more loading cycles than can
samples without fibres before liquefaction occurs; that is, the inclusion of fibres can effectively improve
the liquefaction resistance of sand. The liquefaction resistance increased with increasing fibre length
and fibre content. The reinforcement effect is found to be significant in both medium dense samples
and loose samples. The results from triaxial tests and hollow cylinder shear tests were similar, indicating
that the different loading conditions in the two types of tests will not influence the reinforcement effect
of randomly distributed fibres.

KEYWORDS: Geosynthetics, Liquefaction, Polypropylene fibres, Cyclic triaxial compression test,

Hollow cylinder shear test

REFERENCE: Ye, B., Cheng, Z. R., Liu, C., Zhang, Y. D. and Lu, P. (2017). Liquefaction resistance of
sand reinforced with randomly distributed polypropylene fibres. Geosynthetics International, 24, No. 6,
625–636. [http://dx.doi.org/10.1680/jgein.17.00029]

soil-related structures. Compared to traditional soil rein-

1. INTRODUCTION
forcing techniques with geosynthetics (e.g. geotextiles and
When earthquakes occur, liquefaction is a major threat geogrids), randomly distributed fibres have the following
to the safety of civil engineering structures constructed advantages: (i) prevention of potential weak planes that
on sandy ground. Great efforts have been devoted by commonly develop parallel to traditional planar reinfor-
researchers and engineers to investigating the mechanisms cement materials (Noorzad and Amini 2014); and
of soil liquefaction, because it can cause loss of life and (ii) maintenance of the isotropic strength characteristics
property (Huang and Yu 2013). Many techniques and (Maher and Gray 1990).
methods, such as compaction, enhanced drainage, soil Fibre reinforcement technology can generally be
replacement, and soil reinforcement, have been applied divided into two categories: oriented reinforcement
in geotechnical engineering to reduce the risk of soil and randomly distributed reinforcement, according to
liquefaction. A soil reinforcement technique using ran- the method by which fibres are distributed within the
domly distributed fibres can be an effective solution soil matrix. The traditional soil reinforcing techniques
to decrease the liquefaction potential of embankments, (e.g. geogrids and geotextiles) have similarities with
retaining structures, slopes, subgrade soils, and other oriented fibre reinforcement, because they are planar.
1072-6349 © 2017 Thomas Telford Ltd 625

Downloaded by [ University of Auckland] on [24/12/17]. Copyright © ICE Publishing, all rights reserved.
 626 Ye, Cheng, Liu, Zhang and Lu

Many researchers have studied the liquefaction behaviour strength of fly ash reinforced with randomly distributed
of soils reinforced with geotextiles. For example, fibres, and found that their addition can significantly
Krishnaswamy and Isaac (1994, 1995) studied the effect increase the liquefaction strength of fly ash. Maher and
of reinforcement using woven or nonwoven geotextiles in Woods (1990), Omarov (2010), Noorzad and Amini (2014)
enhancing the liquefaction resistance of soil samples and Eskisar et al. (2016) investigated the different factors
through a series of cyclic undrained triaxial compression that influence the liquefaction behaviour of soils reinforced
tests. The results showed that liquefaction resistance by randomly distributed fibres, such as fibre species, fibre
increased when the geotextile was included in the soil content, fibre length, and soil density. In addition to cyclic
matrix, and the increasing effects were more significant loading-induced liquefaction, Ibraim et al. (2010) and Liu
for less dense samples. Altun et al. (2008) obtained et al. (2011) investigated the static liquefaction behaviour of
and compared the liquefaction strength curves of soils soils reinforced with randomly distributed fibres. Their
reinforced by woven and nonwoven geotextiles, and they results showed that the randomly distributed fibres have a
found that nonwoven geotextiles performed better than significant effect on increasing the static liquefaction
woven geotextiles as reinforcement materials to improve resistance of soils. Although there have been some studies
the liquefaction resistance of soils. Vercueil et al. (1997) reporting the liquefaction characteristics of soils reinforced
investigated the influence of compressibility and surface with randomly distributed fibres, the results remain very
roughness of nonwoven geotextiles on the liquefaction limited compared to the number of research studies focused
resistance of reinforced soils, and reported that both of on planar reinforcement materials.
those characteristics of the geotextile have significant This paper presents an experimental study on the
effects in increasing the liquefaction resistance of soils. liquefaction characteristics of sands reinforced with
In addition to laboratory cyclic triaxial tests, some randomly distributed polypropylene fibres through a
researchers have studied the liquefaction characteristics series of undrained cyclic triaxial compression tests and
of geosynthetic-reinforced ground using shaking table hollow cylinder shear tests. Several factors that will
model tests. Maheshwari et al. (2012) and Che (2012) influence the liquefaction characteristics of reinforced
conducted shaking table tests on model ground reinforced soils, including fibre content, fibre length and relative
by geogrids. It was found that the inclusion of geogrids density, are investigated in detail. It is shown that the
in the ground could decrease the generated excessive inclusion of randomly distributed fibres in the soil matrix
pore water pressure within the ground and the sub- is an effective measure to reduce liquefaction suscepti-
sidence of the ground surface during and after shaking; bility, and that liquefaction resistance increases with the
that is, the liquefaction resistance of the model ground was increment of fibre length and fibre content. In addition,
effectively improved by geogrid reinforcement. the reinforced fibres can work effectively both in medium-
In reinforcement technology with randomly distributed dense sand and loose sand.
fibres, the fibres are directly mixed with soils. The use
of randomly distributed fibres as reinforcement has
attracted the attention of researchers and engineers 2. TESTING MATERIALS
because of its more satisfactory performance compared
with that of traditional reinforcements (e.g. geogrids and 2.1. Tested sand
geotextiles). Numerous experimental studies have proved The soil tested in this study is Fujian standard sand, which
that the inclusion of randomly distributed fibres can is widely used in geotechnical experiments in China. To
effectively improve the strength and deformation behaviour make the sand more prone to liquefaction, grains larger
of soils under static loading and cyclic loading conditions than 1 mm in size were sieved out from the tested sand.
(Li and Ding 2002; Yetimoglu and Salbas 2003; Lovisa The final grain size distribution curve of the tested sand is
et al. 2010; Falorca and Pinto 2011; Viswanadham et al. shown in Figure 1. The physical properties of the tested
2011; Ple and Le 2012; Fatahi et al. 2013; Hamidi and sand are shown in Table 1.
Hooresfand 2013; Olgun 2013; Tafreshi et al. 2013;
Estabragh et al. 2014; Cristelo et al. 2015). As for the 2.2. Reinforced fibres
liquefaction characteristics of soils reinforced by randomly The reinforcement material used in this test was poly-
distributed fibres, the test results of Noorany and propylene fibre, which is commonly used in geotechnical
Uzdavines (1989) indicated that the improvement in the engineering as a geosynthetic reinforcement material. The
liquefaction resistance of soil from the use of randomly diameter of the fibres was 31 μm, and they were cut into
distributed fibres was better than that achieved using planar small pieces of 6, 12 and 19 mm in length to investigate
reinforcement materials. The same conclusion was the influence of fibre length. The properties of the
obtained by Maheshwari et al. (2012) through shaking polypropylene fibres are shown in Table 2.
table tests. They compared the vibration subsidence and
excess pore water pressure of sandy ground reinforced by
geogrid and randomly distributed synthetic fibres. It was 3. TESTING METHODS
found that randomly distributed fibres can restrain the
generation of excess pore water pressure and reduce the 3.1. Tested samples
vibration subsidence more effectively than geogrids. To evaluate the influence of different factors, a series
Boominathan and Hari (2002) investigated the liquefaction of undrained cyclic triaxial compression tests was first
Geosynthetics International, 2017, 24, No. 6

Downloaded by [ University of Auckland] on [24/12/17]. Copyright © ICE Publishing, all rights reserved.
 Liquefaction resistance of sand reinforced with randomly distributed polypropylene fibres 627

100 Table 3. The main characteristics of the tested samples in

triaxial tests

80 Specimen CSR FC (%) FL (mm) Dr (%) Skempton’s

Percent finer by weight (%)

no. B value

60 [1] 0.1 0 Unreinforced 30 0.972

sample
[2] 0.1 0.2 12 30 0.965
40 [3] 0.1 0.4 12 30 0.962
[4] 0.1 0.6 12 30 0.974
[5] 0.1 0.8 12 30 0.963
20 [6] 0.15 0 Unreinforced 30 0.981
sample
[7] 0.15 0.2 12 30 0.963
0 [8] 0.15 0.4 12 30 0.974
0.01 0.1 1.0 10 [9] 0.15 0.6 12 30 0.981
Grain size (mm) [10] 0.15 0.8 12 30 0.968
[11] 0.2 0 Unreinforced 30 0.976
Figure 1. Grain size distribution curve of the tested sand sample
[12] 0.2 0.2 12 30 0.969
[13] 0.2 0.4 12 30 0.970
Table 1. Physical properties of the tested sand [14] 0.2 0.6 12 30 0.981
[15] 0.2 0.8 12 30 0.962
[16] 0.1 0.6 6 30 0.964
Density of soil particles ρs (g/cm3) 2.64
[17] 0.1 0.6 19 30 0.981
Maximum void ratio emax 0.78
[18] 0.15 0.6 6 30 0.985
Minimum void ratio emin 0.50
[19] 0.15 0.6 19 30 0.979
[20] 0.2 0.6 6 30 0.987
[21] 0.2 0.6 19 30 0.969
[22] 0.15 0 Unreinforced 20 0.971
Table 2. Properties of the polypropylene fibres
sample
[23] 0.15 0.2 12 20 0.979
Properties Values
[24] 0.15 0.4 12 20 0.968
[25] 0.15 0.6 12 20 0.965
Diameter (μm) 31
[26] 0.15 0.8 12 20 0.969
Density (kg/m3) 0.91 × 103
Length (mm) 6, 12, 19

is electro-mechanical and has a maximum operating fre-

conducted on 26 saturated samples with different fibre quency of 5 Hz and a minimum operating frequency of
contents (FC), fibre lengths (FL) and relative densities 0.0001 Hz. The apparatus can provide sinusoidal cyclic
(Dr) under different cyclic stress ratios (CSR). Next, control of either the axial displacement or the axial force.
another series of hollow cylinder shear tests was con- In this study, a 0.01 Hz sinusoidal cyclic control of the
ducted on five samples with different fibre contents axial force was used. The available axial load of the
to investigate whether the different loading conditions apparatus was in the range of 5 kN, and the available cell
in triaxial tests and hollow cylinder shear tests would pressure and back pressure were in the range of 2 MPa.
influence the reinforcement effect of fibres. Tables 3 and 4 The available diameter and height of specimens were 39.1
summarise the main characteristics of all the tested and 80.0 mm, respectively. The cell pressure, shear strain,
samples. In the tables, the fibre content was defined shear stress, back pressure, and vertical and lateral defor-
as the ratio of the fibre mass to the soil mass. Here, the mation of the specimen were measured using a 16-bit
cyclic stress ratio in the triaxial test is defined as dynamic data logging system.
To prepare high-quality soil-fibre mixture samples,
σ′1  σ′3
CSR ¼ ð1Þ sand of the required weight was first mixed with a small
2σ′3 amount of water; next, the required amount of fibres was
where σ′1 and σ′3 are the vertical and the horizontal added to the moist sand, and they were then carefully
effective stresses, respectively. mixed together. In this process, a small amount of water is
In this study, when calculating the relative density necessary because the fibres cannot be evenly mixed with
for the fibre-sand mixture, the void ratio occupied by the the dry sand. Conversely, an excessive amount of water
fibres was not considered, and the maximum and mini- should not be added because the fibres might float in the
mum void ratios for pure sand were employed. water and cause uneven mixing. By trial and error, the
appropriate water content was found to be approximately
10%. Next, the soil-fibre mixture was divided into
3.2. Undrained cyclic triaxial compression tests four equivalent parts (each layer 20 mm high) and care-
A dynamic triaxial test apparatus was used in this fully filled into a split mould using the moist tamping
experimental study. The actuator system of the apparatus method (Ishihara 1993). Subsequently, the mould with the
Geosynthetics International, 2017, 24, No. 6

Downloaded by [ University of Auckland] on [24/12/17]. Copyright © ICE Publishing, all rights reserved.
 628 Ye, Cheng, Liu, Zhang and Lu

Table 4. The main characteristics of the tested samples in hollow cylinder shear tests

Specimen no. Shear stress (kPa) FC (%) FL (mm) Dr (%) Skempton’s B value

[27] 35 0 Unreinforced sample 30 0.977

[28] 35 0.2 12 30 0.974
[29] 35 0.4 12 30 0.968
[30] 35 0.6 12 30 0.981
[31] 35 0.8 12 30 0.972

soil-fibre mixture was set up in the pressure chamber. A chambers to individually control the outer cell pressure,
loading cap was placed on the top of the sample, and inner cell pressure, and back pressure. The available outer
a small vacuum pressure was applied through the back cell pressure, inner cell pressure, and back pressure are in
pressure outlet to allow the sample to maintain its cy- the range of 0.8 MPa.
lindrical shape after the moulds were removed. The sample preparation process was the same as the
Afterwards, a small cell pressure of 30 kPa was applied process for the triaxial tests described in the previous
through the hydraulic devices, and the vacuum pressure section. After preparation, the sample was consolidated at
was simultaneously released. The air bubbles within the a confining pressure (inner and outer cell pressures) of
samples were removed by flushing approximately 1000 ml 200 kPa and back pressure of 100 kPa; that is, the same
of de-aired water slowly through the samples from the conditions as those of the triaxial tests. After consolidation,
bottom. When Skempton’s B value (pore pressure coeffi- the samples were subjected to a cyclic torsional shear load
cient) was greater than 0.96, the sample was regarded as under undrained conditions until liquefaction occurred;
fully saturated. that is, the mean effective stress of the sample reached the
After the sample was prepared, cell pressure and back zero-stress state. During this loading process, the loading
pressure were applied, and the sample was isotropically frequency was set as 0.01 Hz, and the amplitude of the
consolidated. In all the tests, the confining pressure and torque was set as 7.2 N·m, which corresponded to a
the back pressure were set at 200 kPa and 100 kPa, res- shearing stress of 35 kPa applied to the samples.
pectively, and the corresponding effective confining
pressure was 100 kPa. After the consolidation process
was completed, stress-controlled testing was performed
4. RESULTS AND DISCUSSION
under an undrained condition. A sinusoidal waveform
loading with the frequency of 0.01 Hz was applied on the 4.1. Results of the triaxial tests
samples until liquefaction occurred. In the present study, 4.1.1. Influence of fibre content
liquefaction onset was considered as the mean effective A series of samples with the same fibre length (12 mm)
stress when first reaching the zero-stress state, that is, when and initial relative density of 30%, with a fibre content
the generated excess pore water pressure was equal to the varying from 0 to 0.8% in increments of 0.2%, was tested
effective confining pressure (100 kPa). to investigate how the fibre content influences the lique-
faction resistance of the sand. The cyclic stress ratio (CSR)
3.3. Hollow cylinder shear tests was set as 0.1, 0.15 and 0.2, which corresponded to
A hollow cylinder testing apparatus was employed in deviator stresses (the difference between the vertical and
this study. This apparatus can test larger samples than horizontal stresses) of 20, 30 and 40 kPa, respectively.
the triaxial apparatus described in the previous section. Figure 2 displays the time histories of the excess pore
The available outer and inner diameters of the hollow water pressure (EPWP) for the specimens under the
cylindrical samples were 100 mm and 60 mm, respect- loading condition of CSR = 0.15. In the figure, the value
ively. The available heights of the samples were 200 mm. of the effective confining stress (100 kPa) is displayed by
This apparatus can simultaneously impose an axial load, a dashed line. Clearly, with the increment of the fibre
torque, outer cell pressure, inner cell pressure and back contents, the rate of increase of the EPWP became lower.
pressure on the sample and control all five components Here, we employ a parameter Nre, which is defined
individually. The axial load and torque are provided by as the cycle number at which the EPWP first reached
the machinery system at the top of the apparatus; this the effective confining stress, to indicate the loading
system has a maximum operating frequency of 2 Hz and cycle number for triggering liquefaction. Obviously,
a minimum operating frequency of 0.001 Hz. The larger values of Nre corresponded to greater resistance
available axial load of the apparatus is in the range of to liquefaction in the sand. In Figure 2, the Nre values
3 kN, and the available torque is in the range of 30 N×m. increases from 7 (FC = 0) to 27 (FC = 0.8%) with the
The pressure cell is separated into three hermetic cham- increment of the fibre contents; that is, the inclusion of
bers by latex films: the outer chamber, the specimen fibre in specimens can greatly enhance the liquefaction
chamber, and the inner chamber. Each chamber has two resistance of sand, and the enhanced effect increases with
outlets. The outlet at the bottom is used for a water supply, the fibre content.
and the outlet at the top is used for water or gas exhaust. The inclusion of fibre can not only influence the develop-
Three hydraulic devices are connected to the three ment of the EPWP, but also reduce the deformation of the
Geosynthetics International, 2017, 24, No. 6

Downloaded by [ University of Auckland] on [24/12/17]. Copyright © ICE Publishing, all rights reserved.
 Liquefaction resistance of sand reinforced with randomly distributed polypropylene fibres 629

120 10
Nre = 7

Axial strain (%)

5
EPWP (kPa)

80
0
40 FC = 0
FC = 0 –5

0 –10
0 10 20 30 0 10 20 30
Number of cycles (N) Number of cycles (N)

120 10
Nre = 13

Axial strain (%)

5
EPWP (kPa)

80
0
40 FC = 0.2%
FC = 0.2% –5

0 –10
0 10 20 30 0 10 20 30
Number of cycles (N) Number of cycles (N)

120 10
Nre = 16

Axial strain (%)

5
EPWP (kPa)

80
0
40
–5
FC = 0.4%
FC = 0.4%
0 –10
0 10 20 30 0 10 20 30
Number of cycles (N) Number of cycles (N)

120 10
Nre = 19
Axial strain (%)

5
EPWP (kPa)

80
0
40
–5
FC = 0.6% FC = 0.6%
0 –10
0 10 20 30 0 10 20 30
Number of cycles (N) Number of cycles (N)

120 10
Nre = 27
Axial strain (%)

5
EPWP (kPa)

80
0
40
–5
FC = 0.8% FC = 0.8%
0 –10
0 10 20 30 0 10 20 30
Number of cycles (N) Number of cycles (N)

Figure 2. Time histories of the EPWP of specimens with different Figure 3. Time histories of the axial strain of specimens with
fibre contents (at FL = 12 mm, CSR = 0.15, and Dr = 30%) different fibre contents (at FL = 12 mm, CSR = 0.15, and
Dr = 30%)

specimens during cyclic loading. Figure 3 displays the

time history of the axial strain of the specimens during loading cycles (for the specimen with FC = 0, only the first
cyclic loading. It can be seen that the axial strain prior to eight cycles are displayed because of the failure of the
liquefaction was very small, but afterwards it developed specimen). With increasing fibre contents, the maximum
rapidly. With the increase in fibre content, the develop- axial strain decreases from 5.55% (FC = 0.2%) to 0.42%
ment of axial strain was obviously delayed; that is, larger (FC = 0.8%). Thus, from the perspective of deformation,
fibre contents will lead to smaller axial strain at the same the effect of fibre content on the enhancement of the
loading cycles. To observe this phenomenon more clearly, liquefaction resistance is also significant.
Figure 4 displays the stress-strain relationship of speci- Table 5 summarises all the experimental results of the
mens with different fibre contents during the first 15 Nre values and the maximum axial strain of the specimens
Geosynthetics International, 2017, 24, No. 6

Downloaded by [ University of Auckland] on [24/12/17]. Copyright © ICE Publishing, all rights reserved.
 630 Ye, Cheng, Liu, Zhang and Lu

40 40

20 20
Deviator stress (kPa)

Deviator stress (kPa)

0 0

FC = 0.2%
–20 FC = 0 –20

–40 –40
–10 –5 0 5 10 –10 –5 0 5 10
Axial strain (%) Axial strain (%)

40 40

20 20
Deviator stress (kPa)

Deviator stress (kPa)

0 0

–20 FC = 0.4% –20

FC = 0.6%

–40 –40
–10 –5 0 5 10 –10 –5 0 5 10
Axial strain (%) Axial strain (%)

40

20
Deviator stress (kPa)

–20 FC = 0.8%

–40
–10 –5 0 5 10
Axial strain (%)

Figure 4. Stress-strain relationship of specimens with different fibre contents (at N = 15, FL = 12 mm, CSR = 0.15, and Dr = 30%)

with different fibre contents. In Table 5, it is noticeable that Furthermore, from Figure 5, the Nre values under
the maximum axial strains for different CSRs are checked CSR = 0.1 rose more rapidly than did the values under
at different N numbers because the development speed of CSR = 0.15 and 0.2; that is, the influence of fibre content is
the axial strain is greatly different for different CSRs. more remarkable under smaller loading conditions.
Therefore, an appropriate N number should be selected at Finally, Figure 7 shows the liquefaction resistance
which the axial strain has been fully developed but the curves for different fibre contents in terms of the rela-
sample has not failed. Figures 5 and 6 illustrate the vari- tionship between the cyclic stress ratio and corresponding
ations of the Nre values and the maximum axial strain with number of cycles required to cause liquefaction. The
the fibre contents, respectively. Under the loading con- figure clearly shows that the liquefaction resistance
ditions of CSR = 0.1 and CSR = 0.2, the Nre values and is increased with an increase in the fibre content.
maximum axial strain are found to have variation trends Consequently, the reinforcement has a significant effect
with fibre contents similar to the trend of CSR = 0.15. in reducing the liquefaction susceptibility.
Geosynthetics International, 2017, 24, No. 6

Downloaded by [ University of Auckland] on [24/12/17]. Copyright © ICE Publishing, all rights reserved.
 Liquefaction resistance of sand reinforced with randomly distributed polypropylene fibres 631

Table 5. Summary of the test results of the specimens with 0.25

different fibre contents (at FL = 12 mm and Dr = 30%)

CSR FC (%) Nre Maximum axial strain (%) at (i) N = 100

0.20
for CSR = 0.1; (ii) N = 15 for CSR = 0.15;

Cyclic stress ratio

(iii) N = 10 for CSR = 0.2

0.1 0 79 Failurea 0.15 FC = 0.4%

0.2 101 1.12 FC = 0.6%

FC = 0.8%
0.4 127 0.37
0.6 142 0.13
0.8 169 0.10 0.10 FC = 0
FC = 0.2%
0.15 0 7 Failure
0.2 13 5.55
0.4 16 3.32
0.05
0.6 19 2.29 0 50 100 150 200
0.8 27 0.42 Number of cycles (N)
0.2 0 2 Failure
0.2 4 7.73 Figure 7. Liquefaction strength curves with different fibre
0.4 6 6.75 contents
0.6 8 5.41
0.8 13 1.61
4.1.2. Influence of fibre length
a
In this study, a ‘failure’ is defined as the axial strain of a sample
exceeding 10%. To investigate how fibre length influences the liquefaction
resistance of reinforced sand, a series of samples with the
same fibre content (0.6%) for the fibre lengths of 6 mm,
12 mm and 19 mm were tested, with the loading
200
conditions of the cyclic stress ratio (CSR) set as 0.1,
0.15 and 0.2, respectively. The relative densities of the
Number of cyles to liquefaction, Nre

specimens were 30%. The test results of the Nre values and
150
the maximum axial strain during cyclic loading are
summarised in Table 6. Figures 8 and 9 illustrate the
changes of the Nre values and the maximum axial strain
100 CSR = 0.1
with the increase of fibre lengths. From Figure 8, the Nre
CSR = 0.15
CSR = 0.2
values increased almost linearly with fibre length; that is,
an increase in fibre length will increase the number of
50
cycles required to reach liquefaction. In other words,
under the same fibre contents, longer fibres lead to better
performance of the fibres in soils. In addition, Figure 8
0
0 0.2 0.4 0.6 0.8 1.0 shows that the Nre values for CSR = 0.1 rose more rapidly
FC: % than did the values for CSR = 0.15 and 0.2; that is, the
influence of fibre lengths is also more remarkable under
Figure 5. Relationship between Nre and fibre content smaller loading conditions. From Figure 9, the maxi-
mum axial strains during cyclic loading decreased with
increasing fibre lengths; that is, longer fibres help to
reduce the deformation of specimens due to liquefaction.
10
Note that the curves of CSR = 0.2 crossed the curves of
CSR = 0.1 CSR = 0.15; however, this did not indicate that the
8 CSR = 0.15 deformations under CSR = 0.2 were smaller than they
Maximum axial strain (%)

CSR = 0.2 were under CSR = 0.15 because the number of cycles at
which we checked the maximum axial strain were different
6
for different cyclic stress ratios. As shown in Table 6,
for CSR = 0.2, we checked the maximum axial strain
4 at N = 8, and for CSR = 0.15, we checked the maximum
axial strain at N = 30.
Finally, Figure 10 displays the liquefaction strength
2
curves of the sand samples reinforced with different fibre
lengths. The curve of pure sand (unreinforced specimens)
0 is also displayed for comparison. With the increase of fibre
0 0.2 0.4 0.6 0.8 1.0
FC (%)
lengths, the curves are found to be located in the upper
part of the graph, which indicates that an increase in fibre
Figure 6. Relationship between the maximum axial strain and length resulted in an improvement in the resistance to
fibre content liquefaction.
Geosynthetics International, 2017, 24, No. 6

Downloaded by [ University of Auckland] on [24/12/17]. Copyright © ICE Publishing, all rights reserved.
 632 Ye, Cheng, Liu, Zhang and Lu

Table 6. Summary of the test results of specimens with different 0.25

fibre lengths (at FC = 0.6% and Dr = 30%)

CSR FL (mm) Nre Maximum axial strain (%) at (i) N = 80 0.20

for CSR = 0.1; (ii) N = 30 for CSR = 0.15;

Cyclic stress ratio

(iii) N = 8 for CSR = 0.2

0.15 FL = 12 mm
0.1 6 103 0.16
FL = 19 mm
12 142 0.11
19 191 0.074
0.15 6 13 4.80
0.10 Pure sand
12 19 4.05 FL = 6 mm
19 29 2.63
0.2 6 5 6.55
12 8 2.28 0.05
0 50 100 150 200
19 14 0.92
Number of cycles (N)

Figure 10. Liquefaction strength curves for different fibre lengths

200
CSR = 0.1
Number of cycles to liquefaction, Nre

CSR = 0.15 The fibre contents varied from 0.2% to 0.8% with an
150 CSR = 0.2 increment of 0.2%. The fibre length was 12 mm. The
cyclic stress ratio was set as 0.15. Table 7 summarises the
test results for the specimens with different densities.
100 Figures 11 and 12 exhibit the changes of the Nre values
and the maximum axial strain with the fibre contents,
respectively. By comparing the results of the specimens
50 with different densities, the Nre values of the maximum
axial strain of specimens with Dr = 30% are observed to
have higher Nre values and lower maximum axial strains
0 than do the specimens with Dr = 20%, indicating that the
0 5 10 15 20 25 liquefaction resistance of the sand increased with an
FL (mm)
increase in relative density. Moreover, the Nre values and
the maximum axial strains of specimens with both
Figure 8. Relationship between Nre and the fibre length
densities have similar variation trends when the fibre
contents increase from 0.2% to 0.6%, indicating that the
10 reinforcement in loose specimens (Dr = 20%) has a similar
CSR=0.1 enhancing effect on the liquefaction resistance of sand as
CSR=0.15
8
in the medium dense specimens (Dr = 30%). However,
CSR=0.2
when the fibre content increases from 0.6% to 0.8%, then
Maximum axial strain (%)

the rate of increase of the Nre values and the rate of

6
reduction of the maximum axial strains of the specimens
with Dr = 20% are observed to be higher than they are
4 for specimens with Dr = 30%, indicating that the enhance-
ment effect on the liquefaction resistance is more
2 significant for the loose specimens with Dr = 20%.

0 4.2. Results of the hollow cylinder shear tests

In the previous section, the enhancement effect on the
0 5 10 15 20 25
liquefaction resistance of sand was discussed based on the
FL (mm)
experimental results obtained from triaxial compression
Figure 9. Relationship between the maximum axial strain and the tests. Although the triaxial test is the most widely used in
fibre length geotechnical engineering, it displayed the deficiency that
the shear stress mainly occurs in the plane at an angle
of 45° from the horizontal plane; this loading condition
4.1.3. Influence of the relative density is obviously different from an earthquake loading con-
In the above tests, only medium-dense specimens with a dition, in which the shear stresses in the soil mainly
relative density of Dr = 30% were tested. To investigate the occur in horizontal planes. However, this deficiency in the
influence of the relative density of the samples, another set triaxial test apparatus can be overcome by another type of
of tests was conducted on less dense samples (Dr = 20%). shear test, that is, the hollow cylinder torsional shear test.
The less dense samples were prepared using the same In a hollow cylinder torsional shear test, the shear load
moist tamping method but with lower tamping times. can be applied along the horizontal plane. To clarify
Geosynthetics International, 2017, 24, No. 6

Downloaded by [ University of Auckland] on [24/12/17]. Copyright © ICE Publishing, all rights reserved.
 Liquefaction resistance of sand reinforced with randomly distributed polypropylene fibres 633

Table 7. Summary of the test results of specimens with different 120

Nre = 7
relative densities (at FL = 12 mm and CSR = 0.15)

EPWP (kPa)
80
Dr (%) FC (%) Nre Maximum axial strain (%) at (i) N = 15
for Dr = 20%; (ii) N = 15 for Dr = 30%
40
20 0.2 7 6.43 FC = 0
0.4 9 4.74 0
0.6 10 4.65 0 10 20 30
0.8 26 0.55 Number of cycles (N)
30 0.2 13 5.55 120
0.4 16 3.32 Nre = 9
0.6 19 2.29

EPWP (kPa)
80
0.8 27 0.42

40
30 FC = 0.2%

0
Number of cycles to liquefaction, Nre

0 10 20 30
Number of cycles (N)
120
20 Nre = 10

EPWP (kPa)
80

10 40
FC = 0.4%

Dr = 20% 0
0 10 20 30
Dr = 30%
Number of cycles (N)
0 120
0 0.2 0.4 0.6 0.8 1.0 Nre = 11
FC (%)
EPWP (kPa)

80
Figure 11. Relationship between Nre and fibre content for
different densities 40
FC = 0.6%
8
0
0 10 20 30
Dr = 20% Number of cycles (N)
Dr = 30%
6
Maximum axial strain (%)

120
Nre = 24
EPWP (kPa)

80
4

40
FC = 0.8%

2 0
0 10 20 30
Number of cycles (N)

0 Figure 13. Time histories of the EPWP of specimens with

0 0.2 0.4 0.6 0.8 1.0
different fibre contents in hollow cylinder shear tests (at
FC (%)
FL = 12 mm, Dr = 30%, and shear stress = 35 kPa)
Figure 12. Relationship between the maximum axial strain and
fibre content for different densities
contents. With an increase in the fibre content, the number
of cycles required to obtain liquefaction was found to
whether this difference in the loading condition will increase. Figure 14 displays the stress-strain relationship of
influence the reinforcement effect of fibres, we conducted specimens with different fibre contents during the first six
another series of tests using the hollow cylinder shear test loading cycles. With increasing fibre content, the maximum
apparatus. The fibre contents varied from 0 to 0.8% in axial strain decreases from 6.5% (FC = 0) to 0.15%
increments of 0.2%. The fibre length was 12 mm. The (FC = 0.8%). Table 8 summarises the test results in the
relative density was 30%. The torsional shear stress was set hollow cylinder shear tests. According to the results,
as 35 kPa. Figures 15 and 16 illustrate the variation in the Nre values
Figure 13 displays the time histories of the excess pore and the maximum shear strain with fibre content, respect-
water pressure (EPWP) of the specimens with different fibre ively. In the hollow cylinder shear test, the Nre values and
Geosynthetics International, 2017, 24, No. 6

Downloaded by [ University of Auckland] on [24/12/17]. Copyright © ICE Publishing, all rights reserved.
 634 Ye, Cheng, Liu, Zhang and Lu

40 40

FC = 0
FC = 0.2%
20 20
Shear stress (kPa)

Shear stress (kPa)

0 0

–20 –20

–40 –40
–8 –6 –4 –2 0 2 4 6 8 –8 –6 –4 –2 0 2 4 6 8
Shear strain (%) Shear strain (%)

40 40

FC = 0.4% FC = 0.6%
20 20
Shear stress (kPa)

Shear stress (kPa)

0 0

–20 –20

–40 –40
–8 –6 –4 –2 0 2 4 6 8 –8 –6 –4 –2 0 2 4 6 8
Shear strain (%) Shear strain (%)

40

FC = 0.8%
20
Shear stress (kPa)

–20

–40
–8 –6 –4 –2 0 2 4 6 8
Shear strain (%)

Figure 14. Stress-strain relationship of specimens with different fibre contents in hollow cylinder shear tests (at N = 6, FL = 12 mm,
Dr = 30%, and shear stress = 35 kPa)

Table 8. Summary of the test results in hollow cylinder shear tests reducing the liquefaction susceptibility of sand, both in
(at FL = 12 mm, Dr = 30% and shear stress = 35 kPa) hollow cylinder shear tests and in triaxial tests.
FC (%) Nre Maximum shear strain (%) at N = 6

0 7 6.5 5. CONCLUSIONS
0.2 9 2.9
0.4 10 1.1 Undrained cyclic triaxial compression shear tests and
0.6 11 0.70 hollow cylinder shear tests were conducted to investigate
0.8 24 0.15
how randomly distributed fibres influence the liquefaction
resistance of Fujian sand. From the experimental results,
the following conclusions are drawn.
the maximum shear strain were found to exhibit the same
variation trend as that of the triaxial tests, which indicates (1) From the triaxial tests, it was found that the number
that the presence of fibres has a significant effect on of cycles causing liquefaction increased with an
Geosynthetics International, 2017, 24, No. 6

Downloaded by [ University of Auckland] on [24/12/17]. Copyright © ICE Publishing, all rights reserved.
 Liquefaction resistance of sand reinforced with randomly distributed polypropylene fibres 635

30 strain exhibited the same variation trend with fibre

content as that of the triaxial tests, which indicates
Number of cycles to liquefaction, Nre

that the presence of fibres has a significant effect in

reducing the liquefaction susceptibility of sand, both
20 in the hollow cylinder shear tests and in the triaxial
compression tests.

The findings and conclusions in this study are obtained

10 from limited tests on one kind of sand and synthetic
fibres, and only two relative densities, three cyclic stress
ratios, and three fibre lengths were considered. Real soils,
reinforcements, and seismic loadings are more complex
0 than the simple conditions considered in this study.
0 0.2 0.4 0.6 0.8 1.0 Nevertheless, the results of the experiments performed
FC (%)
under these limited conditions indicated that the soil
Figure 15. Relationship between Nre and fibre contents in hollow
reinforcement technique with randomly-distributed fibres
cylinder shear tests can greatly reduce the liquefaction potential of soil.

8
ACKNOWLEDGEMENTS
This work was supported by the National Natural Science
Foundation of China (No. 41472249), the State Key
Maximum shear strain (%)

6
Laboratory of Geo-Hazard Prevention and Geo-
Environment Protection (No. SKLGP2016K019), and
4
the Fundamental Research Funds for Central Universities.

2 NOTATION
Basic SI units are given in parentheses.

0
0 0.2 0.4 0.6 0.8 1.0
CSR cyclic stress ratio (dimensionless)
FC (%)
Dr relative density (dimensionless)
emax maximum void ratio (dimensionless)
Figure 16. Relationship between the maximum shear strain and emin minimum void ratio (dimensionless)
fibre contents in hollow cylinder shear tests EPWP excess pore water pressure (Pa)
FC fibre content (dimensionless)
FL fibre length (m)
increase in the fibre content, whereas the maximum
Nre cycle number at which the EPWP first reached
axial strain during cyclic loading decreased with an
the effective confining stress (dimensionless)
increase in the fibre content. The experimental results
ρs density of soil particles (kg/m3)
clearly showed that fibre reinforcement has a
σ′1 vertical effective stress in the triaxial test (Pa)
significant effect in reducing the liquefaction
σ′3 horizontal effective stress in the triaxial test (Pa)
susceptibility.
(2) For the same fibre content, the increase in the
fibre length helps to increase the number of cycles
causing liquefaction and reduce the deformation
of specimens during cyclic loading, indicating that REFERENCES
longer fibres lead to the better performance of Altun, S., Göktepe, A. B. & Lav, M. A. (2008). Liquefaction resistance of
fibres in soils. sand reinforced with geosynthetics. Geosynthetics International, 15,
(3) When the fibre content varies from 0.2% to 0.6%, No. 5, 322–332.
the reinforcement in loose specimens with Dr = 20% Boominathan, A. & Hari, S. (2002). Liquefaction strength of fly ash
reinforced with randomly distributed fibers. Soil Dynamics and
has a similar enhancing effect on the liquefaction Earthquake Engineering, 22, No. 9, 1027–1033.
resistance as that of the medium dense specimens Che, J. Z. (2012). Experimental Study on Liquefaction Bahavior of
with Dr = 30%. However, when the fibre content Reinforced-soil, Masters thesis, Taiyuan University of Science and
increases from 0.6% to 0.8%, the enhancing effect Technology, Shanxi, China (in Chinese).
on the liquefaction resistance is more significant for Cristelo, N., Cunha, V. M., Dias, M., Gomes, A. T., Miranda, T. &
Araújo, N. (2015). Influence of discrete fibre reinforcement on the
the loose specimens with Dr = 20%. uniaxial compression response and seismic wave velocity of a
(4) In the hollow cylinder shear test, the experimental cement-stabilised sandy-clay. Geotextiles and Geomembranes, 43,
results of the Nre values and the maximum shear No. 1, 1–13.

Geosynthetics International, 2017, 24, No. 6

Downloaded by [ University of Auckland] on [24/12/17]. Copyright © ICE Publishing, all rights reserved.
 636 Ye, Cheng, Liu, Zhang and Lu

Eskisar, T., Karakan, E. & Altun, S. (2016). Effects of fibre reinforcement Maher, M. H. & Gray, D. H. (1990). Static response of sand reinforced
on liquefaction behaviour of poorly graded sands. Procedia with randomly distributed fibers. Journal of Geotechnical
Engineering, 161, 538–542, https://doi.org/10.1016/j.proeng.2016. Engineering, ASCE, 116, No. 11, 1661–1677.
08.688. Maher, M. H. & Woods, R. D. (1990). Dynamic response of sand
Estabragh, A. R., Soltannajad, K. & Javadi, A. A. (2014). Improving reinforced with randomly distributed fibers. Journal of Geotechnical
piping resistance using randomly distributed fibers. Geotextiles and Engineering, ASCE, 116, No. 7, 1116–1131.
Geomembranes, 42, No. 1, 15–24. Maheshwari, B. K., Singh, H. P. & Saran, S. (2012). Effects of
Falorca, I. M. C. F. G. & Pinto, M. I. M. (2011). Effect of short, reinforcement on liquefaction resistance of Solani sand. Journal
randomly distributed polypropylene microfibres on shear strength of Geotechnical and Geoenvironmental Engineering, ASCE, 138,
behaviour of soils. Geosynthetics International, 18, No. 1, 2–11. No. 7, 831–840.
Fatahi, B., Le, T. & Khabbaz, H. (2013). Small-strain properties of soft Noorany, I. & Uzdavines, M. (1989). Dynamic behavior of saturated
clay treated with fibre and cement. Geosynthetics International, 20, sand reinforced with geosynthetic fabrics. Proceeding of
No. 4, 286–300. Geosynthetics’89 Conference, North American Geosynthetics
Hamidi, A. & Hooresfand, M. (2013). Effect of fiber reinforcement Society, San Diego, CA, USA, pp. 385–396.
on triaxial shear behavior of cement treated sand. Geotextiles Noorzad, R. & Amini, P. F. (2014). Liquefaction resistance of Babolsar sand
and Geomembranes, 36, 1–9, https://doi.org/10.1016/j.geotexmem. reinforced with randomly distributed fibers under cyclic loading. Soil
2012.10.005. Dynamics and Earthquake Engineering, 66, 281–292, https://doi.org/
Huang, Y. & Yu, M. (2013). Review of soil liquefaction characteristics 10.1016/j.soildyn.2014.07.011.
during major earthquakes of the twenty-first century. Natural Olgun, M. (2013). Effects of polypropylene fiber inclusion on the
Hazards, 65, No. 3, 2375–2384. strength and volume change characteristics of cement-fly ash
Ibraim, E., Diambra, A., Wood, D. M. & Russell, A. R. (2010). Static stabilized clay soil. Geosynthetics International, 20, No. 4, 263–275.
liquefaction of fibre reinforced sand under monotonic loading. Omarov, M. (2010). Liquefaction Potential and Post-liquefaction Settlement
Geotextiles and Geomembranes, 28, No. 4, 374–385. of Saturated Clean Sands and Effect of Geofiber Reinforcement,
Ishihara, K. (1993). Liquefaction and flow failure during earthquakes. Doctoral thesis, University of Alaska, Fairbanks, AK, USA.
Geotechnique, 43, No. 3, 351–415. Ple, O. & Le, T. N. H. (2012). Effect of polypropylene fiber-reinforcement
Krishnaswamy, N. R. & Isaac, T. N. (1994). Liquefaction potential of on the mechanical behavior of silty clay. Geotextiles and
reinforced sand. Geotextiles and Geomembranes, 13, No. 1, Geomembranes, 32, 111–116, https://doi.org/10.1016/j.geotexmem.
23–41. 2011.11.004.
Krishnaswamy, N. R. & Isaac, T. N. (1995). Liquefaction analysis Tafreshi, S. M., Khalaj, O. & Dawson, A. R. (2013). Pilot-scale load tests
of saturated reinforced granular soils. Journal of Geotechnical of a combined multilayered geocell and rubber-reinforced
Engineering, ASCE, 121, No. 9, 645–651. foundation. Geosynthetics International, 20, No. 3, 143–161.
Li, J. & Ding, D. W. (2002). Nonlinear elastic behavior of fiber-reinforced Vercueil, D., Billet, P. & Cordary, D. (1997). Study of the liquefaction
soil under cyclic loading. Soil Dynamics and Earthquake resistance of a saturated sand reinforced with geosynthetics.
Engineering, 22, No. 9, 977–983. Soil Dynamics and Earthquake Engineering, 16, No. 7, 417–425.
Liu, J., Wang, G. H., Kamai, T., Zhang, F. Y., Yang, J. & Shi, B. (2011). Viswanadham, B. V. S., Rajesh, S., Divya, P. V. & Gourc, J. P. (2011).
Static liquefaction behavior of saturated fiber-reinforced sand in Influence of randomly distributed geofibers on the integrity of
undrained ring-shear tests. Geotextiles and Geomembranes, 29, clay-based landfill covers: a centrifuge study. Geosynthetics
No. 5, 462–471. International, 18, No. 5, 255–271.
Lovisa, J., Shukla, S. K. & Sivakugan, N. (2010). Shear strength of Yetimoglu, T. & Salbas, O. (2003). A study on shear strength of sands
randomly distributed moist fibre-reinforced sand. Geosynthetics reinforced with randomly distributed discrete fibers. Geotextiles and
International, 17, No. 2, 100–106. Geomembranes, 21, No. 2, 103–110.

The Editor welcomes discussion on all papers published in Geosynthetics International. Please email your contribution to
[email protected] by 15 June 2018.

Geosynthetics International, 2017, 24, No. 6

Downloaded by [ University of Auckland] on [24/12/17]. Copyright © ICE Publishing, all rights reserved.