performance of geosynthetic-reinforced alternative … › 14d7 › 186629cec7d6699...geosynthetics...

Performance of geosynthetic-reinforced alternativesub-ballast material in a railway track

G. Fernandes1, E. M. Palmeira2 and R. C. Gomes3

1Federal University of Ouro Preto, Department of Civil Engineering—School of Mines, 35400-000 Ouro

Preto, MG, Brazil, Telephone: +55 31 35591558, Telefax: +55 31 35591548,

E-mail: [email protected];2University of Brasılia, Department of Civil and Environmental Engineering-FT, 70910-900 Brasilia,

DF, Brazil, Telephone: +55 61 3273 7313 ext. 217, Telefax: +55 61 3273 4644,

E-mail: [email protected] University of Ouro Preto, Department of Civil Engineering—School of Mines, 35400-000 Ouro

Preto, MG, Brazil, Telephone: +55 31 3559153, Telefax: +55 31 35591548, E-mail: [email protected]

Received 4 December 2007, revised 10 February 2008, accepted 13 February 2008

ABSTRACT: Geosynthetics have been used as reinforcement in various applications in geotechnical

engineering. This paper presents a study on the use of geosynthetic reinforcement to reduce the

consumption of good-quality sub-ballast material in a railway track in the state of Minas Gerais,

Brazil. Six test sections on the railway were instrumented and monitored for a period of 2 years.

Nonwoven geotextile and a geogrid were used as reinforcement in different positions in the sub-

ballast. The results obtained showed that the presence of geosynthetic reinforcement reduced the

strains mobilised in the sub-ballast, reduced the breakage of ballast element, and allowed the use

of a cheaper alternative material in the sub-ballast construction. The results showed the potential of

the use of alternative sub-ballast material reinforced with geosynthetic to reduce the costs of

construction and maintenance of railway tracks.

KEYWORDS: Geosynthetics, Reinforcement, Railway track, Alternative sub-ballast material

REFERENCE: Fernandes, G., Palmeira, E. M. & Gomes, R. C. (2008). Performance of geosynthetic-

reinforced alternative sub-ballast material in a railway track. Geosynthetics International, 15, No. 5,

311–321. [doi: 10.1680/gein.2008.15.5.311]

1. INTRODUCTION

Geosynthetics have been used as reinforcement in various

applications in geotechnical engineering, such as retaining

walls, steep slopes, and embankments on soft soils. How-

ever, a potential field of application of geosynthetics that

is not well explored is as reinforcement or separators in

railways. In comparison with other applications, in this

type of work the geosynthetic is subjected to very severe

conditions, owing to the mechanically aggressive charac-

teristics of the fill (ballast) and to the high load levels

transmitted by the trains. Nevertheless, the use of geosyn-

thetic reinforcement in railways can reduce the develop-

ment of permanent strains, increase track lifetime, and

reduce maintenance costs.

Geotextiles have been used in maintenance works in

railways as separators or drains in regions of soft subgrade

or shallow groundwater level, or for seepage control

(Raymond 1985; Selig and Waters 1994; Arruda Filho

1997). The geotextile can be placed at different elevations

in the track structure, such as the ballast top, between

ballast and subgrade, between ballast and sub-ballast, or

between sub-ballast and subgrade, depending on the

structural configuration of the track (Selig and Waters

1994). Benefits from the use of nonwoven geotextiles in

railway tracks have been reported by Raymond (1999) and

Salim (2004), for instance.

Atalar et al. (2001) performed laboratory model tests on

ballast with inclusions of geogrids and geotextiles, and

found reductions of track settlement due to the presence

of the geosynthetic reinforcement. Similar tests were

performed by Bathurst and Raymond (1987), Gobel et al.

(1994) and Raymond (2002), who also found the bene-

ficial effects of the geosynthetic reinforcement in reducing

track deformation. Indraratna et al. (2007) observed that

the presence of geosynthetic reinforcement reduced track

settlements and particle breakage, particularly for tracks

made of recycled ballast. Cuconatto (1997) monitored

sections of an urban railway where geotextiles and

geocells were used, and a good performance of these

sections was observed, in comparison with the traditional

track solutions. Rowe and Jones (2000) recommended the

use of a geocomposite capable of providing ballast rein-

Geosynthetics International, 2008, 15, No. 5

3111072-6349 # 2008 Thomas Telford Ltd

forcement and separation between ballast and subgrade,

and this can be accomplished by the combined use of a

geogrid and a geotextile.

The aggressiveness of the ballast has to be considered

in railway applications, because of the possibility of

mechanical damage of the geosynthetic inclusion. Selig

and Waters (1994), Raymond and Bathurst (1990), Ashpiz

et al. (2002) and Nancey et al. (2002) have reported

different levels of geosynthetic mechanical degradation

with time, depending on the characteristics of the track.

Raymond (1999) found that geotextiles with a mass per

unit area below 500 g/m2 failed after 5 years of service,

partially because of incorrect installation. Because of the

nature of the fill material, and the high load levels,

measures may have to be taken to protect the geosynthetic

layer against mechanical damage and to prolong its life-

time in railway applications.

Another possible use of geosynthetic reinforcement in

railways is in the combination of substandard or recycled

ballast material and reinforcement in order to save

construction and maintenance costs in this type of work.

Indraratna et al. (2006) reported significant contributions

of geosynthetic reinforcement for the reduction of settle-

ments and particle breakage of a recycled ballast material.

In this context, Minas Gerais is the most important state

in Brazil with regard to mining activities. Most of the

mining production is transported by trains, which subject

the track to very high loads, resulting in a need for

continuous maintenance works. Because of this, good,

expensive ballast materials are required in railway con-

struction and maintenance. However, the mining activity

in the region generates a huge amount of waste, and some

of this waste may present satisfactory mechanical proper-

ties for use in railway tracks. The combination of this

waste with geosynthetic reinforcement may provide a

cheaper feasible technical alternative to the use of more

expensive construction materials.

This paper describes the use of geosynthetic reinforce-

ment and mining waste as a solution to reduce the use of

more expensive sub-ballast materials in railways.

2. EXPERIMENTS

The experiment described in this paper was carried out in

a segment of the Vitoria–Minas railway, in the state of

Minas Gerais, Brazil. This railway is 104 years old,

898 km long, and is responsible for approximately 40% of

the load transported by railways in the country. The main

objective of the research programme was to evaluate the

potential use of geosynthetics as reinforcement and

separators in railways, with an emphasis on the substitu-

tion of part of the good-quality sub-ballast material by a

combination of mining waste and geosynthetic reinforce-

ment. For this study, experimental sections of that railway

were constructed and instrumented.

The research described in this paper was conducted

under real in-service conditions of the railway. In part, this

limited some research activities that might disrupt the

traffic of trains. However, in spite of that constraint,

relevant data on the performance of the test sections were

obtained.

Six experimental railway test sections, each 25 m long,

were constructed as part of the research programme. These

sections are localised in Line 2 of the Variante Capitao

Eduardo, Vitoria–Minas railway, between kilometres 40

and 42. This railway is intensely trafficked, with approxi-

mately 16 compositions (2 locomotives of 160 t plus 100

wagons with approximately 100 t each: 300 kN axle load)

daily. The steel sleepers are of type UIC865, with dimen-

sions 2.2 m 3 0.3 m 3 0.02 m. The rail type is TR 68.

Figure 1 shows the geometrical characteristics of the

instrumented sections, and Figure 2 shows a longitudinal

profile of the experimental test sections (sections S1 to

S6), where the location and types of geosynthetic materi-

als used can be identified. All sections were composed of

450 mm of ballast and 200 mm of sub-ballast. Experimen-

tal section 1 (the control section) was constructed follow-

ing the traditional practice adopted in the region of using

good-quality ballast material. In sections 2 to 5 an

alternative sub-ballast material incorporating geosynthetics

was employed. In section 6 the alternative ballast material

was also used, but without the presence of a geosynthetic

layer. The alternative sub-ballast material used in sections

2 to 6 was a mixture of the subgrade local soil (silty sand,

average particle diameter 0.13 mm, 50% by weight), a fine

mine waste (sandy silt, average particle diameter

0.032 mm, 25% by weight), and the traditional sub-ballast

material used in section 1 (sandy gravel, average particle

diameter 5.1 mm, 25% by weight). The same ballast

material was used in all test sections.

Laboratory investigations on the materials used in the

experimental sections included characterisation tests, soil

characteristic curve tests, direct shear tests, scanning

electronic microscopy, X-ray diffractometry, chemical

analysis, Los Angeles abrasion tests, and point load tests

(ASTM D 5731). For the control of construction of the

aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa

500

410

1080 500800 800pipe

650Ballast Sub-ballast

200

rail

Sleeper

3% 3%Instrumentationsupport box

Drainage trench

410

Figure 1. Typical test-section cross-section (dimensions in mm)

312 Fernandes et al.


sections, tests such as the in situ density sand cone test

and the in situ stiffness test were performed. The latter

consists in the determination of the soil stiffness in situ

from its dynamic response using a GeoGauge H4140

apparatus, manufactured by Humboldt Scientific Ltd. This

apparatus calculates ground stiffness based on successive

measurements of ground surface displacements caused by

the application of a constant cyclic compressive load.

Good comparisons between the results of soil modulus

from GeoGauge, cyclic triaxial tests and back-analysis

from Benkelman beam tests are reported by Fernandes

(2005). Soil samples were periodically taken from the

track for characterisation and Los Angeles abrasion tests

(for the ballast material).

The characteristics of the ballast material used in all

sections are presented in Table 1. Particle size distribution

curves of the soils are presented in Figure 3. The charac-

teristics of the sub-ballast materials used in the sections

are presented in Table 2. The results in Figure 3 and Table

2 show that the sub-ballast material used in sections 2 to 6

is considerably finer that that used in the control section

(section S1). By being finer, the sub-ballast material used

in sections 2 to 6 is less mechanically aggressive to the

geosynthetic layer. Figure 4 shows a typical view of the

ballast elements, as well as the rails and sleepers.

The characteristics of the subgrade material in each

section are presented in Table 2. Table 3 shows the

geotechnical properties of the subgrade material obtained

by field tests after compaction. The results of elastic

moduli show that the subgrades of sections S5 and S6 are

more compressible than those of sections S1 to S4.

A nonwoven needle-punched geotextile and a biaxial

grid were used in the test sections. The nonwoven

geotextile is made of continuous filaments of polyester

and has a mass per unit area of 300 g/m2. This geotextile

was intended also to function as a separator in this

application. Its mass per unit area can be considered rather

low for railway applications. However, because of the

rather short duration of the application, a lighter and

cheaper product was chosen. As will be remarked later in

this paper, even under these conditions significant damage

was observed in exhumed geotextile specimens. The

biaxial geogrid is also made of polyester fibres, with an

HDPE cover. The main mechanical characteristics of the

geosynthetics used are summarised in Table 4. The inter-

face friction angle between the nonwoven geotextile and

the sub-ballast material used in direct shear tests was 338,

whereas for the geogrid the interface friction angle was

358.

The geosynthetic layer was installed at different posi-

tions, depending on the section considered (Figure 2). The

position of the geosynthetic in the sub-ballast layer was

varied in order to investigate its influence on the track

performance. In sections S2 and S3 the geogrid was

installed at the base and at the top of the sub-ballast layer,

aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa






450

200

S1 S2 S3 S4 S5 S6

GeogridNonwoven geotextile

Ballast

Sub-ballast

Subgrade

Figure 2. Longitudinal cross-section of instrumented test

sections

Table 1. Characteristics of the ballast material

Shape Prismatic

Unit weight (kN/m3) 35.75

Water absorption (%) 1.13

D85 (mm) 67

D50 (mm) 51

D10 (mm) 34

Fraction of fines (, 0.074 mm) 0

Unconfined compression strength (MPa) 156–166

Point load index, Is50 (MPa)a 7.9–8.3

Anisotropy index, Ia50 (dimensionless) 1.06

Los Angeles abrasion tests (%)b 10.4

Solubility in water No changes after 15 days of immersion in water

aAccording to ASTM D 5731.bBefore traffic.

0

20

40

60

80

100

Per

cent

age

pass

ing

(%)

Particle size diameter (mm)0.0001 0.001 0.01 0.1 1 10 100

Ballast Traditional sub-ballast

Alternative sub-ballast

Figure 3. Particle size distribution curves of materials of

sections

Performance of geosynthetic-reinforced alternative sub-ballast material in a railway track 313


Table 2. Characteristics of sub-ballast and subgrade materials

Section 1 Sections 2 to 6 Subgrade

Gsa 3.95 4.29 3.19

D85 (mm)b 24 0.9 5.7

D50 (mm) 5.1 0.076 0.13

D10 (mm) 0.0075 0.00081 0.006

Clay fraction (%)c 8 11 16

Silt fraction (%)c 35 31 19

Sand fraction (%)c 40 51 39

Gravel fraction (%)c 49 7 16

emind 0.26 0.11 0.878

emaxd 0.45 0.30 1.176

HRBe A-1.a A-2.4 A-5

USCe GW-GC SM SM

BPRe A-1/GIf ¼ 0 A-4/GI ¼ 1.4 A-4/GI ¼ 2

Friction angle (degrees)g 37.5 41.8 42

Cohesion (kPa)g 0 0 0

CBR (%)h 125 106 65.7

ªdmax (kN/m3)i 25.4 27.3 17.1

wopt (%)j 12.2 9.5 24.7

Expansion (%) – – 0.21

aGs: particle specific gravity.bDn (n ¼ 10, 50 or 85): diameter for which n (%) of the soil particles in weight are smaller than that

diameter.cPercentage by weight.demin and emax: minimum and maximum void ratios, respectively.eSoil classification by the Highway Research Board, Unified Soil Classification System and Bureau of

Public Roads.fGI: group index.gFrom direct shear tests.hCBR: California Bearing Ratio obtained for intermediate compaction energy.iªdmax: maximum dry unit weight obtained in compaction tests with intermediate compaction energy.jwopt: optimum moisture content.

Table 4. Geosynthetics characteristics

Nonwoven geotextile Geogrid

Mass per unit area (g/m2) 300 430

Thickness (mm) 2.7 1–2

Tensile strength (kN/m)a 15 65

Maximum tensile strain (%)a 50–70 10

Tensile stiffness (kN/m) 42–281b 520c

Grid aperture dimensions (mm) – 30 3 30

aFrom wide-width tensile tests according to ASTM D 4595.bFirst value is from tests in isolation; second value is secant tensile

stiffness at 5% tensile strain obtained in tests on geotextile confined

by alternative ballast material under confining stress of 220 kPa (test

equipment described in Mendes and Palmeira 2007).cSecant tensile stiffness for a strain of 5%.Figure 4. View of the ballast material, rail and sleepers

Table 3. Results of in situ tests on the subgrade after compaction

S1 S2 S3 S4 S5 S6

ªdmax (kN/m3) 16.59 16.76 23.28 23.30 23.28 16.50

Degree of compaction (%) 100.2 101.3 101.1 101.0 101.2 100.7

Elastic modulus (MPa)a 109.7 100.8 111.8 106.8 88.1 99.5

aFrom in situ stiffness test using GeoGauge H4140.



respectively, as shown in Figure 2. In sections S4 and S5

the nonwoven geotextile was installed at the top and at the

base of the sub-ballast, respectively.

Monitoring of the sections’ performance was achieved

by the installation of soil deformation gauges at the top

and bottom of the ballast material. The soil deformation

gauges were 100 mm long and consisted of extensible

electric resistances (strain gauges) fixed to circular plates

at the ends (Figure 5a). The deformation gauges were

calibrated in the laboratory buried in the sub-ballast

material under the same conditions expected in the field.

Temperature and moisture content gauges (CS615 Water

Content Reflectometer, manufactured by Campbell Scien-

tific Inc.) were also installed in the ballast and sub-ballast

materials of the test sections. Benkelman beam tests

(Figures 5b and 5c) were carried out on the test sections at

different times during the experiments. For these tests a

locomotive weighting 1600 kN with 200 kN/axle was

employed. A rain gauge meter allowed for measurements

of precipitation on the experimental sections.

Additional information on the materials and method-

ology used is reported by Fernandes (2005).

3. RESULTS AND DISCUSSIONS

3.1. Track stiffness

In situ stiffness tests using the GeoGauge H4140 apparatus

were carried out after the installation and compaction of

the sub-ballast material in each section to assess the

significance of the presence of the geosynthetic for the

elastic response of the system. Figure 6 shows the results

obtained; it can be seen that the elastic modulus of the

control section S1 was 165 MPa, and for the other sections

it varied between 119 MPa and 154 MPa. The differences

in the elastic moduli of the sections suggest a stiffer

response of the sections with geogrid reinforcement (S2

and S3) in comparison with those with nonwoven geotex-

tile. The sections with geogrid (S2 and S3) presented

elastic moduli 5.8% and 29.7% greater than those with

geotextiles (S4 and S5) for the geosynthetic at the sub-

ballast top and base, respectively. Although these results

may suggest a stiffer response of the geogrid-reinforced

sections, the influence of variability of the soil properties

of the test sections cannot be ignored. Table 3 shows that

the subgrade in section S5 had the lowest elastic modulus

(25% lower than that in section S1). Hence the differences

between the elastic moduli of the sections after sub-ballast

construction are likely to have been a consequence of

varying subgrade properties rather than of the presence or

absence of geosynthetic layers.

Figure 7 shows results of maximum surface deflections

obtained in Benkelman beam tests on the sub-ballast

material just after its installation and compaction. For

these tests a loaded truck with a rear axle load of 82 kN

was used. Sections S5 and S6 presented the largest surface

deflections, which is consistent with the results obtained

by the in situ stiffness test. However, for the magnitude of

(a)

(b) (c)

Figure 5. Techniques to assess track deformability: (a) installation of strain gauges in track; (b) Benkelman beam tests on

subgrade and (c) on superstructure



the displacements measured, the differences obtained are

again likely to be a consequence of soil properties

variability rather than the influence of the presence of

geosynthetics.

After the test sections were completed, Benkelman

beam tests were carried out on the superstructure (after

installation of the ballast layer, UIC865 sleepers and TR

68 rails). For these tests, a two-car locomotive with eight

axles each and a total weight of 1600 kN was employed.

Figure 8 shows the vertical displacement profiles obtained

for each section. Under the much higher load provided by

the locomotive the results obtained suggest a marked

influence of the presence of the geosynthetic layers on the

performance of the alternative sub-ballast material. Sec-

tion S6 (unreinforced alternative sub-ballast material)

presented much higher vertical displacements than the

other sections. With the exception of section S4, for all

other reinforced sections the maximum vertical displace-

ments were closer to or smaller than those obtained for

section S1 (control section, with standard sub-ballast

material, unreinforced). It is also interesting to notice that

the sub-ballast material (the alternative material) was the

same in sections S2 to S6. Thus from Figure 8 it can be

seen that the maximum vertical displacements obtained

for the reinforced sections were between 2.3 and 4.4 times

smaller than that obtained for the unreinforced section S6.

3.2. Measured strains

The mortality rate of instruments in railways is quite high,

owing to the magnitude of the loads and the severity of

the conditions to which they are subjected (Selig and

Waters 1994). Even in highway pavements, where the

conditions are less severe than in railways, the loss of

instruments can be significant. Warren and Howard (2007)

report losses of 18% of the instruments in test sections of

a highway pavement constructed under well-controlled

conditions. In this study, some instruments were damaged

during the monitoring period, which limited the amount of

data acquired. Nevertheless, the measurements made allow

inferences to be drawn on the effectiveness of geosyn-

thetics used in combination with the alternative sub-ballast

material. Figure 9 shows the variation of horizontal strain

at the bottom of the sub-ballast layer with the number of

train axles that passed on sections S1, S3, S4 and S6. The

measurements were made under the sleeper and on the

vertical passing on the rail centreline (Figure 9). For up to

600 000 passages of train axles the performance of all

sections was similar. Beyond this, sections S3 and S4

aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa






0

50

100

150

200E

last

ic m

odul

us (

MP

a)

S1 S2 S3 S4 S5 S6

Figure 6. System stiffness from in situ tests after construction

of sub-ballast







0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Max

imum

sur

face

def

lect

ion

(mm

)

S1 S2 S3 S4 S5 S6

Figure 7. Results from Benkelman beam tests on sections

after construction of sub-ballast

0

1

2

3

4

0 1 2 3 4 5 6 7 8 9 10

S1

S3

S6

S2

S5

S4

Distance (m)

Def

lect

ion

(mm

)

Figure 8. Vertical displacement profiles from Benkelman

beam tests on superstructure

aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa

1.00 10� 5

6.00 10� 51.10 10� 6

1.60 10� 62.10 10� 6

2.60 10� 6

0

2000

4000

6000

8000

10000

12000

14000

16000

Number of axles

Str

ain

()

µε

S1S3S4S6

Strain gauge

Figure 9. Horizontal strains at bottom of sub-ballast



presented less horizontal strain than the control section

S1. Section S6 (unreinforced alternative sub-ballast mate-

rial) presented the highest horizontal strains at the base of

the sub-ballast layer.

Figure 10 presents the variation of vertical strains at the

base of the sub-ballast layer with the number of train

axles. Not much difference in the results can be seen up to

100 000 train axles. As the traffic increases, section S6

tends to present the largest vertical strains, followed by

section S1, although the latter shows signs of reduction on

the rate of strain with traffic intensity. The same trend of

reduction of the rate of strain with the number of train

axles is observed for section S3 after the passage of

1 500 000 train axles.

3.3. Ballast degradation

Samples of the ballast of the test sections were collected

for testing during the monitoring period. Particle size

distribution analyses and Los Angeles abrasion tests were

conducted on the samples. Figure 11 shows the results of

particle size distribution analyses after 600 days of

monitoring, corresponding to a passage of 2 120 000 train

axles or 75 860 070 t on the test sections. The particle size

distribution of the ballast material before traffic is also

shown, for comparison. The results show some level of

degradation of the ballast material caused by the traffic of

heavy trains. In comparison with the ballast particle size

distribution curve before traffic, the results obtained for

sections S1 and S6 were the poorest. Among the rein-

forced sections, section S2 (geogrid at the sub-ballast

base) presented the greatest particle size reduction.

The breakage level of the ballast grains can be quanti-

fied by the breakage index Bg (Marsal 1973), which

expresses the percentage by weight of particles that have

undergone breakage, and is defined as

Bg ¼Xn

1

˜Wki � ˜Wkfð Þ (1)

for values of ˜Wki � ˜Wkf . 0, where Bg is the

percentage by weight of the ballast that has undergone

breakage, ˜Wki is the initial fraction of the sample weight

corresponding to a given range of ballast particles dimen-

sions before breakage, ˜Wkf is the final fraction of the

sample weight corresponding to a given range of ballast

particles after breakage, and n is the number of ranges of

particle dimensions for which ˜Wki � ˜Wkf . 0.

The variations of Bg with time for the experimental

sections are shown in Figure 12. Up to 90 days of train

traffic the values of Bg for the test sections were similar.

After 600 days section S1 presented the greatest values of

Bg, followed by section S6: both of these were unrein-

forced sections. Not much difference was observed among

the values of Bg for sections S3, S4 and S5 after 600 days’

traffic. The result obtained for section S2 was slightly

higher than those for the other reinforced sections after

600 days’ traffic. However, the authors believe that this

difference lies within the scatter of results from this type

of evaluation, because of the test limitations and the

possible variability of ballast properties. These results

suggest that the presence of the geosynthetic layers had a

beneficial effect in reducing ballast breakage. This can in

part be due to the increased lateral confinement of the

railway track because of the presence of the reinforcement,

as suggested by Lackenby et al. (2007).

The fouling index (Selig and Waters 1994) provides an

evaluation of the level of ballast contamination by fines,

and is defined as the sum of the percentages of ballast

masses passing through sieves number 4 (4.8 mm) and

200 (0.074 mm) in a particle size analysis. Section 1

aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa

1.00 10� 5

6.00 10� 51.10 10� 6

1.60 10� 52.10 10� 6

2.60 10� 6

0

2000

4000

6000

8000

10000

Number of axles

Str

ain

()

µε

Strain gauge

S1S3S4S6

12000

Figure 10. Vertical strains at bottom of sub-ballast

100

80

60

40

20

010 100

Particle diameter (mm)

Per

cent

age

finer

(%

)

S1

S3

S6

S2

S5

S4

Ballast beforetraffic

20 40 60 80

Figure 11. Particle size distribution curves of ballast material

after 600 days’ traffic

0 100 200 300 400 500 600 700Time (days)

0

10

20

30

40

50

60

S1

S2

S3

S4

S5 S6

Bg

(%)

2.1No. of axles ( 10 )� 6

0.3 1.40.7

Figure 12. Breakage index against time



presented the highest fouling index value (0.42%) after

600 days’ traffic, although the results in Figure 11 yield

values of fouling index below 1% for all sections. Accord-

ing to Selig and Waters (1994), contamination of the

ballast can be considered negligible for fouling index

values below 1%.

Figure 13 shows the results obtained in Los Angeles

abrasion tests on samples of the ballast material collected

from the railway track, according to the Brazilian standard

NBR 6465 for this type of test, which uses 12 steel

spheres with a mass of 6 kg. The results of the tests varied

between 6.4% and 8.3%. This variation of results is likely

to be within the scatter of test results for the ballast used,

and suggests that the different types of sub-ballast material

and the presence of geosynthetics did not have a signifi-

cant effect on the ballast strength as measured by the Los

Angeles abrasion test.

The same sub-ballast (alternative) material was used in

sections S2 to S6, S2 to S5 being reinforced sections and

S6 unreinforced. Thus the influence of the presence of the

reinforcement can be assessed by comparing the results

obtained, using section S6 as a reference. Figure 14

presents the results obtained for maximum vertical dis-

placements from Benkelman beam tests on the super-

structure, ballast breakage index (Bg) and abrasion tests

(Los Angeles abrasion test results after 600 days’ traffic)

for sections S2 to S5. In this figure, the raw results

obtained for sections S2 to S5 were normalised by the

corresponding results obtained for section S6. The results

in Figure 14 suggest that the presence of geosynthetic

reinforcement in the alternative ballast material reduced

the maximum vertical displacements in Benkelman beam

tests by between 57% and 78% with respect to the value

obtained for the unreinforced section S6. This range is

consistent with the reduction of strains at the sub-ballast

base due to the presence of reinforcement (Figures 9 and

10). Values of Bg were reduced by between 16% and 47%,

whereas the maximum reduction observed for Los Angeles

abrasion test results was 23%. Bearing in mind the

expected scatter of test results in full-scale experiments

under in-service conditions, these results indicate a sig-

nificant effect of the presence of the reinforcement in the

alternative sub-ballast material on the relative increase on

track stiffness and reduction in ballast breakage. However,

less influence on the increase in ballast resistance to

abrasion can be inferred, in comparison with the situation

for the unreinforced alternative sub-ballast material.

The results in Figure 14 do not allow a definite conclu-

sion to be drawn on the best position for the reinforcement

layer in the sub-ballast. However, these results are consis-

tent with those obtained from large-scale laboratory tests

performed by Brown et al. (2007), which showed little

difference between the performance of ballast reinforced

with geogrid at the base or at the middle of the ballast

layer thickness. In the field, reinforcement at the base of

the ballast is the most practical solution.

3.4. Geotextile mechanical degradation

Geotextile specimens were exhumed from the track for

wide-width tensile tests to evaluate the effects of mechani-

cal damage. Tests were carried out on virgin specimens,

on specimens after spreading and compaction of the sub-

ballast, and on specimens exhumed after 90 days’ traffic

(N ¼ 3.18 3 105). The wide strip tensile tests were per-

formed according to ASTM D 4595. It was difficult to

obtain authorisation from the railway company to exhume

samples of the geosynthetics installed in the railway track,

because the time required for such exhumations inter-

rupted train traffic and disrupted the normal activities of

the railway. The tests were therefore limited to specimens

of the geotextile used in the research programme. Figure

15 shows samples of the damage in the exhumed speci-

aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa


aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa




0

2

4

6

8

10

Los

Ang

eles

abr

asio

n re

sult

(%)

S1 S2 S3 S4 S5 S6

Figure 13. Results of Los Angeles abrasion tests on ballast

material

aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa




0

0.2

0.4

0.6

0.8

1.0

Nor

mal

ised

val

ue

S2 S3 S4 S5

Normalised maximum vertical displacement fromBenkelman beam tests on superstructure

Normalised values of after 600 days’ trafficBg

Normalised Los Angeles abrasion test results

Figure 14. Comparison between performance of sections S2

to S5 and section S6



mens of the geotextile. These damages were characterised

by small cuts and holes.

Figure 16 shows the range of wide-width tensile tests

performed on the exhumed specimens of the nonwoven

geotextile. The maximum tensile force in tests on virgin

geotextile specimens varied between 14.8 kN/m and

16.7 kN/m. Some reduction of the average tensile strength

and an increase in the scatter of the test results can be

observed after compaction of the sub-ballast and after

90 days’ train traffic, which corresponded to 3.18 3 105

passages of loaded axles. The average strength and scatter

of the test results obtained after 90 days’ traffic were

similar to those obtained just after spreading and compac-

tion of the sub-ballast material. This suggests that, up to

90 days’ traffic, most of the damage to the geotextile

occurred during installation and compaction of the sub-

ballast.

In-soil tensile tests were also performed on virgin and

damaged geotextile specimens, using the equipment de-

scribed in Mendes et al. (2007). In these tests the

alternative sub-ballast material was used to confine the

geotextile specimens. Figure 17a shows a schematic view

of this equipment, which allows the execution of in-soil

tensile tests on geotextile specimens under confinement

by the soil only (no friction between soil and geotextile).

The average results obtained for different levels of

confinement are presented in Figure 17b, and show a

reduction of the geotextile secant tensile stiffness of

between 12% and 36% in the strain range 2–5%, depend-

ing on the confining stress value considered. In this strain

range the secant tensile stiffness of damaged geotextile

specimens varied between 220 kN/m and 460 kN/m, de-

pending on the confining stress. The latter value is not so

significantly smaller than the geogrid tensile stiffness

(520 kN/m), which may in part explain the close perform-

ance of the test sections reinforced with geotextile and

with geogrid. These significant values of geotextile stiff-

ness may have also been a consequence of impregnation

of the geotextile by particles of the sub-ballast material

during its spreading and compaction. Mendes et al. (2007)

found significant increases of secant tensile stiffness in

Figure 15. Typical mechanical damage to exhumed geotextile specimens

Aftersub-ballastcompaction

0

5

10

15

20

Tens

ile s

tren

gth

(kN

/m)

Virgin After 90days’ traffic

Figure 16. Ranges of wide-width tensile test results

JackLoad cell

ClampGeotextile (100 mm 200 mm)�

Rollers

Pressurised bag

100

1100

Soil

Displacement transducer

(a)

0 5 10 15 20Strain (%)

0

200

400

600

800

1000

Sec

ant t

ensi

le s

tiffn

ess

(kN

/m)

Virgin - 220 kPaExhumed - 50 kPa

Exhumed - 100 kPaExhumed - 220 kPa

(b)

Figure 17. In-soil tensile tests on virgin and exhumed

geotextile specimens: (a) in-soil tensile test apparatus

(Mendes et al. 2007); (b) Secant tensile stiffness against

tensile strain



in-soil tensile tests on similar geotextiles using a confining

granular material (glass beads) with particle size range

similar to that of the alternative sub-ballast material used

in the test sections.

The experimental technique used in the in-soil tensile

tests does not simulate the stretching and pre-tensioning

of the geotextile that are likely to occur under field

conditions during spreading and compaction of the fill

material. Careful installation of the geotextile in the field

is required in order to minimise mechanical damage and

geotextile folding, wrinkles or loose spots. The beneficial

effect of impregnation of the geotextile by soil particles

will depend on the particles’ dimensions and geotextile

opening sizes.

4. CONCLUSIONS

This paper has presented the performance of railway test

sections where different forms of sub-ballast reinforce-

ment were employed using geosynthetics. The main

objective of the research was to evaluate the use of a

reinforced low-quality, cheaper sub-ballast material as a

substitution to the one traditionally used in a railway

subjected to intense traffic and high loads due to the

transportation of iron ore. The main conclusions of the

research programme are summarised below.

The testing techniques and experimental methodology

employed were satisfactory. Research under in-service

conditions is very important in investigating the mechan-

isms that influence the performance of the track. However,

the constraints imposed by the need to maintain full-time

railway operation limit the amount of information that can

be acquired. The mortality of instruments in this type of

work is usually high owing to the severity of the

conditions to which the instruments are subjected (high

loads, vibrations, and soil particle aggressiveness).

The results of in situ tests and strain measurements

showed that the presence of the geosynthetic reinforce-

ment reduced the compressibility of the system. Reduced

breakage of the ballast material and greater abrasion

resistance were observed in the test sections with geosyn-

thetics. This can be attributed to the greater lateral

confinement provided by the reinforcement layer.

Similar track performance was observed for the two

types of geosynthetic tested. However, significant mech-

anical damage was observed in the geotextile used after

600 days in service. This suggests that the long-term

performance of the geotextile might be severely compro-

mised by increasing mechanical damage, and that the use

of a thicker, more durable geotextile would have been

more appropriate.

The results obtained validated the use of an alternative,

low-cost, sub-ballast material in combination with geosyn-

thetic reinforcement. This solution can be attractive for

the mining industry in situations where reasonably good-

quality mining wastes are plentiful, and conventional track

construction materials are scarce or expensive. Neverthe-

less, further research is required for a proper understand-

ing of the beneficial effects of geosynthetics in railway

tracks, with particular reference to the development of

design methods.

ACKNOWLEDGEMENTS

The authors are indebted to the following institutions that

supported the research activities described in this paper:

University of Brasilia, Federal University of Ouro Preto,

CVRD-Vale do Rio Doce Mining Company, CNPq-

National Council for Scientific and Technological Develop-

ment, and CAPES-Brazilian Ministry of Education.

NOTATIONS

Basic SI units are given in parentheses.

Bg breakage level (dimensionless)

CBR California Bearing Ratio (dimensionless)

D10 diameter for which 10% in mass of particles

are smaller than that value (m)





emax maximum void ratio (dimensionless)

emin minimum void ratio (dimensionless)

GI group index (dimensionless)

Gs particle specific gravity (dimensionless)

Ia50 anisotropy index (dimensionless)

Is50 point load index (Pa)

N number of axles (dimensionless)

n number of ranges of particle dimensions for

which ˜Wki � ˜Wkf . 0 (dimensionless)

wopt optimum moisture content (dimensionless)

˜Wki initial fraction of sample weight

corresponding to given range of ballast

particle dimensions before breakage

(dimensionless)

˜Wkf final fraction of sample weight corresponding

to given range of ballast particle dimensions

after breakage (dimensionless)

ªdmax maximum dry unit weight (N/m3)

REFERENCES

Arruda Filho, A. (1997). Flumitrens: permanent rail track. Seminar on

Technological Update for Permanent Track Maintenance, Rio de

Janeiro, RJ, Brazil (in Portuguese).

Ashpiz, E. S., Diederich, R. & Koslowsi, C. (2002). The use of

spunbonded geotextile in railway track renewal St Petersburg-

Moscow. Proceedings of the 7th International Conference on

Geosynthetics, Nice, France, Vol. 3, pp. 1173–1176.

ASTM D 4595. Test Method for Tensile Properties of Geotextiles by the

Wide-Width Strip Method. ASTM International, West Conshohocken,

PA, USA, 11 pp.

ASTM D 5731. Standard Test Method for Determination of the Point

Load Strength Index of Rock and Application to Rock Strength

Classifications. ASTM International, West Conshohocken, PA,

USA.

Atalar, C., Das, B. M., Shin, E. C. & Kim, D. H. (2001). Settlement of

geogrid-reinforced railroad led to failure due to cyclic load.

Proceedings of the 15th International Conference on Soil Mech-



anics and Geotechnical Engineering, Istanbul, Vol. 3, pp. 2045–

2048.

Bathurst, R. J. & Raymond, G. P. (1987). Geogrid reinforcement of

ballasted track. Transportation Research Record, No. 1153, 8–14.

Brown, S. F, Kwan, J. & Thom, N. H. (2007). Identifying the key

parameters that influence geogrid reinforcement of railway ballast.

Geotextiles and Geomembranes, 25, No. 6, 326–335.

Cuconatto, J. M. (1997). A Study on Railway Ballast Contamination and

the Use of Geosynthetics for Control. MSc thesis, COPPE/Federal

University of Rio de Janeiro (in Portuguese).

Fernandes, G. (2005). Behaviour of Geosynthetic Reinforced Railway

Tracks Constructed with Fine Grained Soils and Mining Wastes.

PhD thesis, Graduate Programme of Geotechnics, University of

Brasilia, Brasilia, Brazil, 253 pp. (in Portuguese).

Gobel, C. H., Weisemann, U. C. & Kirschner, R. A. (1994). Effectiveness

of a reinforced geogrid in a railway subbase under dynamic loads.

Geotextiles and Geomembranes, 14, No. 2, 91–99.

Indraratna, B., Khabbaz, H., Salim, W. & Christie, D. (2006).

Geotechnical properties of ballast and the role of geosynthetics in

rail track stabilisation. Ground Improvement, 10, No. 3, 91–101.

Indraratna, B., Shahin, M. A. & Salim, W. (2007). Stabilisation of

granular media and formation soil using geosynthetics with special

reference to railway engineering. Ground Improvement, 11, No. 1,

27–43.

Marsal, R. J. (1973). Mechanical properties of rockfill. Embankment

Dam Engineering: Casagrande Volume, R. C. Hirschfeld and S. J.

Poulos (eds), John Wiley & Sons, New York, pp. 109–200.

Lackenby J., Indraratna B., McDowell G. & Christie D. (2007). Effect of

confining pressure on ballast degradation and deformation under

cyclic triaxial loading. Geotechnique, 57, No. 6, 527–536.

Mendes, M. J. A., Palmeira, E. M. & Matheus, E. (2007). Some factors

affecting the in-soil load–strain behaviour of virgin and damaged

nonwoven geotextiles. Geosynthetics International, 14, No. 1, 39–50.

Nancey, A., Imbert, B. & Robinet, A. (2002). Thick and abrasion

resistance geotextile for use under the ballast in railways structures.

Proceedings of the 7th International Conference on Geosynthetics,

Nice, France, Vol. 3, pp. 1191–1194.

NBR 6465. Los Angeles Abrasion Tests on Aggregates. Associacao

Brasileira de Normas Tecnicas (Brazilian Association for Technical

Standards), Rio de Janeiro, 4 pp. (in Portuguese).

Raymond, G. P. (1985). Research on railroad ballast specification and

evaluation. Track Design and Construction: Transportation Re-

search Record, No. 1006, 1–8.

Raymond, G. P. (1999). Railway rehabilitation geotextiles. Geotextiles

and Geomembranes, 17, No. 4, 213–230.

Raymond, G. P. (2002). Reinforced ballast behaviour subjected to

repeated load. Geotextiles and Geomembranes, 20, No. 1, 39–61.

Raymond, G. P. & Bathurst, R. J. (1990). Test results on exhumed railway

track geotextiles. Proceedings of the 4th International Conference

on Geotextiles, Geomembranes and Related Products, The Hague,

The Netherlands, Vol. 1, pp. 197–202.

Rowe, R. K. & Jones, C. J. F. P. (2000). Geosynthetics: innovative

materials and rational design. Proceedings of GeoEng 2000,

Melbourne, Australia, Vol. 1, pp. 1124–1156.

Salim, M. D. W. (2004). Deformation and Degradation Aspects of Ballast

and the Role of Geosynthetics in Track Stabilisation. PhD thesis,

University of Wollongong, 285 pp.

Selig, E. T. & Waters, J. M. (1994). Track Geotechnology and

Substructure Management. Thomas Telford, London, UK, 750 pp.

Warren, K. A. & Howard, I. L. (2007). Sensor selection, installation, and

survivability in a geosynthetic-reinforced flexible pavement.

Geosynthetics International, 14, No. 5, 298–315.

The Editor welcomes discussion on all papers published in Geosynthetics International. Please email your contribution to

[email protected] by 15 April 2009.



performance of geosynthetic-reinforced alternative … › 14d7 › 186629cec7d6699...geosynthetics...

Documents