Railway Ballast Requirements for High Speed and Heavy Haul Lines:

Hardness, Fouling, Life Cycle

Giannakos Konstantinos
Civil Engineer PhD, F. ASCE, M. TRB-AR050&AR060, AREMA, fib.,
Visiting Professor of Railways, Dpt. Civil Engineering, University of Thessaly, Greece

Loizos Andreas
Civil Engineer, PhD, Professor, Department of Transportation Planning and Engineering,NTUA, Greece

Plati Christina
Civil Engineer PhD, Laboratory of Highway Engineering, NTUA, Greece

ABSTRACT: Results from tests performed on ballast used in the Greek network, as well as a new method for
the estimation of ballast fouling as a function of ballast hardness are presented in the present paper. Stresses
that develop under the seating surface of the sleeper and that are transmitted to the ballast, influence maintenance intervals and cost. Ballast hardness influences the fouling and consequently the life-cycle of the material laid at the ballast-bed. This leads to the determination of a minimum quality for the technical specifications in relation to traffic conditions (daily tonnage). The present paper discusses these issues and results from
the research programs performed on the Greek railway network.
1 INTRODUCTION

undeflected as possible in order to prevent or to

minimize permanent vertical deformations. This
function of the ballast-bed in every railway track
implies its deterioration and the subsequent fouling
of ballast that leads to costly maintenance works.
The authors have been involved in research
programs of the National Technical University of
Athens (NTUA) for the Hellenic Railways
Organization (OSE) to develop modern Technical
Specifications for ballast. In the present paper a
portion of the investigation is presented.

During the study for the dimensioning as well as

the selection of the individual materials of a railway
track, the weak links are the ballast and the substructure. These elements of the track present residual deformations as a percentage of the deflection/subsidence, directly connected to the
deterioration of the so-called geometry of the track,
which can be nevertheless described much more
specifically as quality of the track. The smaller the
residual deformations and the slower their alteration
over time, the better the quality of the track. A railway track is a multi-layered structure (Fig. 1) consisting of a vertical succession of various materials
or layers of materials that define the final position of
the rail running table as well as the properties of the
track itself, as it reacts to the action that is created from the motion of the railway vehicle. Each
material or layer that constitutes the line can be
simulated by a combination of a spring with spring
constant ki and a damper with damping coefficient
c i.
According to the theoretical analysis of Winkler,
Timoshenko, and others (e.g. Zimmermann), that
model the track as an infinite beam on an elastic
foundation, the deflection of the track should be high
enough to distribute the acting load to a longer section of the track and thus to reduce the reacting force
at each point. This amount of deflection can be
provided by a resilient fastening and its rail pad,
since the substructure should be constructed to be as

Figure 1 Typical simulation of ballasted track as multi-layered

structure / combination of springs and dampers, with characteristic values of spring constants i (static stiffness coefficient
in kN/mm) per layer

2 BALLAST PROPERTIES REQUIREMENTS

2.1 General
In Greece a High Speed Line of mixed traffic
(maximum operational speed V250 km/h and axle
1

load 22.5 t) on the main corridor Patras - Athens Thessaloniki - Eidomeni (frontier with FYROM) is
under construction with the funding of European
Union and Greek Government. Until 1999, only
twin-block concrete sleepers were used, which were
of French technology, type Vagneux U2, U3 with
RN fastenings and U31 with Nabla fastenings. Nabla
fastenings are laid also in the TGV (High Speed
Trains) lines in France. After almost 12 years of operation extended cracking on sleepers U2/U3 and
completely fouled ballast-bed were observed. A
twenty-year research program to determine the
causes was initiated to study the sleeper -ballast
system under the specific conditions (rolling stock,
ballast quality, rail running table, level of maintenance, etc.). The research program (in which one of
the authors, Dr K. Giannakos, participated as head of
the Hellenic railway scientific team and co-ordinator
of the research and the other two as members of the
research team) was conducted by OSE Hellenic
Railways Organization with the participation of
European universities and research centres of railway organizations. It included both laboratory tests
and investigation of the phenomena that occurred on
the track. After the end of the first part of the investigation program, a part of which was for the ballast,
the requirements for ballast properties were changed
in the Greek network.

160/170 km/h) show that the combination of the results of the aforementioned two tests give the specification for the total ballast strength. These two tests
lead to a coefficient of total hardness of ballast, as
derived from a double entry diagram (see Giannakos, 2010b) cited in both French and Greek regulations, and adopted after the common research program (Giannakos, 2004). This coefficient depicts the
ballast behavior with the time passing. The Deval
Wet attrition Coefficient (D.H.) is entered on the
horizontal axis and the Los Angeles Abrasion
Coefficient (L.A.) is entered on the vertical axis. The
intersection point of the two parallel lines to the axes
drawn from D.H. and L.A. is located between two
consecutive trapezoidal lines so the coefficient of
instant hardness of ballast DRi, and the coefficient of
total hardness of ballast DRG is calculated as the
average of twelve samples (twelve DRi). The
relative quantity of powder, created by impact
stressing and abrasion, is proportional to the
coefficient of total hardness (DRG according to
regulations). This coefficient provides the "measure" of life-cycle of the ballast laid on track (Loizos
et al., 1992-1993).
2.3 Greek Standards for Ballast

2.2 Ballast requirements for High Speed Lines

and Greek regulations
Ballast material should be produced from solid
rocks with angular grains. Its durability should
secure drainage of rain water, absorb vibrations as
damper and provide the best conditions for
maintaining the geometry of the track after tamping.
Selig & Waters (1994/2000), describe the properties
that ballast should fulfill. Due to these demands
ballast grains should provide both the relevant
strength for:
(a) the Deval Wet attrition test, (Selig & Waters,
1994/2000, and SNCF, 1985 and modifications), replaced more recently by microDeval attrition test.
The Deval test characterizes the ballast resistance in
the abrasive degradation between the ballast grains,
particularly in the area between the sleepers (see
relevantly Giannakos, 2004) and
(b) the Los Angeles Abrasion test, (Selig & Waters, 1994/2000, and SNCF, 1985 and modifications), that characterizes the value of the ballast
strength in impact loading (shock).
Experience from French network in both conventional lines (maximum operational speed up to 200
km/h) and High Speed lines (maximum operational
speed up to 300 km/h) and the experience from the
Greek network (maximum operational speed up to

Figure 2 Definition of Railway Ballast Hardness DRi according

to the Greek and French Regulations through Los Angeles coefficient (LA) on the vertical axis and Deval Wet (humide) coefficient (DH) on the horizontal axis. DRi is defined at the intersection of the two lines of LA and DS parallel to the axisand the area among the trapezoidal scale (from 5 to 32).

The Ballast Hardness DRi is defined, according

the Greek and French Regulations before the enactment of the European Standards for railway ballast,
through the Los Angeles coefficient and the Deval

ing to the methods cited in the American literature

and Giannakos (2004) method. Finally a comparison
between the loads on track imposed by Heavy-Haul
traffic with slow speeds and High Speed traffic with
normal axle loads and 300km/h or 186.45mph yields
results more adverse for the case of High Speed
lines, in many cases (see also a more detailed analysis in a forthcoming TRR 2011 issue, (Giannakos,
2011).

coefficient, via a diagram of double input (Fig. 2) in

accordance to French railways' regulation. The hardness of ballast was investigated in the past also in
Greece in relation to the concrete ties (Loizos et al.,
1992-1993).
Ballast hardness was measured in laboratory according to: (a) Los Angeles (L.A.) test per ASTM
C131, (b) L.A. test per the French Normes NFP
18573, (c) Deval test and (d) the total hardness DRi
derived from the combination of L.A. (French) and
Deval tests (French and Greek regulations).

3 BALLAST-BED STRESS AND

DEFORMATION
3.1 General

2.4 European Standards (E.N.) for Ballast

requirements

For a given quality of ballast material, as far as

the part of the deformations caused by the ballast
and the earthworks of the track are concerned, the
correct combination and usage of heavy track machinery is planned for the accomplishment of the
geometry of the track according to the level prescribed in the regulations. In modern railway track
construction, on High Speed lines, for the layers underneath the ballast a very well-executed construction is required: crushed stone material in the upper
layer, 100% Modified Proctor or 105 % Proctor
compaction. This specification implies a substructure almost undeflected (with minimal contribution
of the subgrade to the total deflection) scoping to the
diminishing of the permanent deformations, since
residual deformations are a percentage of the actual
deflection of the railway track caused by the passing
of the loads (Hay, 1982) and they originate mainly
from the substructure and ballast. This leads to an
almost proportional deterioration of the geometry of
the track. This means that the deflections should be
kept, if possible, almost zero. However, the need of
significant deflection development, in practice, in
order to achieve a distribution of the acting load to
the adjacent sleepers contradicts the requirement to
minimize track deflection. This significant deflection should be offered by the fastening and its rail
pad (Giannakos, 2011). As for the issue of ballast
fatigue, the existing literature assumes a uniform
distribution of stresses under the sleeper and without
further details uses the mean value of stress on the
ballast-bed.
But in reality, the seating of the sleepers is supported on discrete points, points of contact with the
grains of the ballast as well as points of contact
among the grains of ballast, (Fig. 3) and the resulting
necessity to calculate the stress per grain of ballast
cannot give comparative results to the rest of the literature. So it is possible to use the mean value of
pressure not as an absolute quantity, but comparatively and in combination with the possibility it covers. Dr. J. Eisenmann, (1988) also shares this view

Recently, the European Committee for Standardization published the European Standard for Railway
Ballast (EN 13450, 2002), applicable in all Member
States of E.U., and the relevant Greek Organization
ELOT, adopted it with the publication of the Greek
Standard (ELOT 13450, 2003). The European
Standards, as most of the E.U. directives, are derived
as a compromise among the national regulations of
the Member States, and as such the EN 13450 only
takes into account the L.A. coefficient. It should be
noted that the non-uniformity of support with swinging sleepers (Hay, 1982) is unacceptable for good
tracks. In High Speed lines as well as in conventional lines the track must be of excellent quality and
consequently situation of a non well seated track with voids under the sleeper seating surface and
swinging sleepers, permitting abrasion among the
ballast grains and the sleeper- is "forbidden". After
the enforcement of EN 13450 there was a need of
finding the correspondence between the older and
newer regulations as well as the correlation between
ballast hardness and ballast life-cycle on track, according to existing literature from older tests and
measurements. For this purpose an investigation
program was undertaken (Loizos et al., 2006-2007).
2.5 Heavy Haul Railways
There is a general discussion in railway engineering cycles and academics that in Heavy Haul railroads (wheel loads 17.69 t or 39,000 lb and maximum speed 60 mph or 96,6 km/h) the actions/loads
per sleeper are higher than in the High Speed lines
of mixed traffic (wheel loads 11.25 t or 24,800 lb
and V 250 km/h or 155.34 mph). This discussion
could lead to the false conclusion that Heavy Haul
generates much more severe actions on sleepers, ballast and substructure. It has to be underlined that all
theoretical methods in international literature are
based on exactly the same theoretical approach
based on Winkler's theory also adopted by
Zimmermann. According to Eisenmann (1984) the
theoretical calculation gives results close to the average of the measurements on track under operation.
In the present paper an analysis is presented accord3

stating that the mean value of pressure is a criterion

for the stressing of the ballast on track.

3.3 Giannakos (2004) method

The mean stress on the upper surface of formation (ballast) can be calculated by the following
equations:

QNSM
QSM
p Asubsidence Qwheel Q
C (3)
hTR
2

Figure 3: Ballast grains in the ballast bed and transmission of

stresses and actions.

There is no uniform support of the sleeper on the

ballast, nor uniform compaction of the ballast and
the ground and there are faults on the rail running
table, imperfections on the wheels etc. A decisive
coefficient in determining the dimensioning of the
superstructure is the maximum value which is calculated, based on probability laws, from the mean
value and standard deviation. In order to calculate
the value of the load that stresses the sleeper, the triple value of the standard deviation (Giannakos &
Loizos, (2009)) is taken (probability of occurrence P
= 99.7%) and for the ballast twice the value of the
standard deviation, of the dynamic component of the
load (probability of occurrence P = 95.5%).
3.2 AREMA method
In AREMA (2005) the following equation is
given for the mean value of stress p ballast on ballastbed:
IF

Qwheel 1
Astat
Q A
100

total stat
pballast
Leff tie btie
Ltie e btie
(1)
Where: Qwheel = Static Wheel Load, Qtotal = Total
Wheel Load static and dynamic, Lsleeper = length of
the sleeper i.e. 8-6 or 2590 mm, e = gauge of the
track (~1500 mm), Leff-sleeper effective length of the
sleeper, bsleeper = width of the sleeper at the seating
surface, IF = impact factor
1 4 3 total
(2)

E
2 2
Where: the distance between the sleepers, total=
total static stiffness coefficient of the track, E,I the
modulus of elasticity and the moment of inertia of
the rail.
A stat

total
Ab

2

Asubsidence

(4)
1
2 2

3
3
E J hTR

(5)

Where: Ab = the sleeper seating surface (for

monoblock sleepers the central non-loaded area
should be subtracted), Q=Component of the load
due to cant deficiency, (QNSM) standard deviation
of the dynamic component of the load due to NonSuspended Masses, (QSM) standard deviation of
the dynamic component of the load due to Suspended Masses (Giannakos, 2010 a).
It must be noted that even French and German literature cites that the measurements on track indicate
that the dispersion of results due to the dynamic
loading should be taken into consideration, entering
the calculation through coefficients depending on
the probability of the occurrence of various parameters (Eisenmann, 1980, Eisenmann, 1988). A
smaller coefficient of probability of occurrence
(95.5% with t=2 or even 68.3% with t=1) is used for
the formation of the track (Eisenmann, 1988).
4 RELATION BETWEEN BALLAST LIFECYCLE AND FOULING
4.1 Ballast Fouling as a function of Ballast
Hardness
The ballast-bed acts as a damper to the loads applied on the track through the friction between its
grains and of its resilience. In order to maintain
these properties the voids of its structure should
not be filled either by up-moving soil (pumping) or
by the small particles or powder created by the attrition induced from loading. When the voids between
the grains of the ballast-bed are filled beyond a certain point, commonly referred to as fouling, a wellcompacted and polluted ballast-bed is created that
loses its resilience and it cannot be tamped or maintained. Moreover, it forms an undeflected seating,
increasing the actions undertaken by the track and
reducing the possibilities of well-performed tamping. In this case the track geometry is not maintained. For these reasons it is very important to find

where: percentage = the percentage of fines in a

determined area under the sleeper's seating surface
constituting the polluted area of ballast-bed.
According to the measurements performed at the
laboratory of the SNCF and the experience in
France, when the percentage of fines is >12.5%, the
case where the ballast is considered to be completely
fouled (no sufficient tamping work can be performed) and needs replacement or cleaning or rising
of the track of at least 15 cm or approximately 6
inches, t = hours of operation of Vibrogir test,
DRG/DRi =the global or instant hardness of ballast
according to the French (and Greek) regulations for
railway ballast derived from the coefficients Los
Angeles and Deval Wet as shown below, which
is extracted from the Greek ballast regulations. It is
obvious that the fouling due to ballast's wear is different from the case of coal dust ballast fouling
(Tutmuluer et al., 2008). The increase of the fines
and the reduction of void spaces lead to a more
compacted ballast reducing its operation as damper
and worsening its capability to keep in high level its
"maintenability". The coefficient is determined by
the following equations:

a relationship between the ballast quality and the

life-cycle of ballast on track and assess the impact of
ballast fouling.
During a research program from 1988-2008, The
Hellenic Railways Organization investigated the
matter of ballast quality and life-cycle. The authors
of the present paper participated, over a long period
of time either in collaboration or separately in this
investigation program.
During the investigation performed at the French
Railways (SNCF) laboratories, tests on the Greek
and French ballast were executed. Cyclic load test
results for hard limestone ballast, as well as for
metamorphic, eruptive and magmatic ballast performed in the Vibrogir device (the device is described thoroughly in Giannakos, 2010 b) already
existed at the SNCF (e.g. Lecocq, 1988). The experiments, as scheduled and performed, simulate a
track under real conditions with a circulation of 22.5
tons per axle and over 200 km/h speed. For the
ballast fouling the number of cycles simulates the
number of tonnage passing over the track. This test
is scheduled and approved in the French State Railways (for speeds V>300km/h). Tamping influence is
not measured in this test. A ballast hardness of
DRG=12.2 was used as an average for the Greek ballast and DRG=16.3, 19.2, and 21 for the French ballast (metamorphic, eruptive and magmatic). For the
ballast material box test have also performed and
their results are described in Giannakos, (2010 c).
The following equation was derived from laboratory
test results and can be used to calculate the lower
values of hardness (Giannakos, 2010 b):

10
percentage

DRG

(6)

t 0.8

35.00

LA(EN) = 0,761 LA(FR) - 0,004

R2 = 0,92
30.00

Los Angeles (EN)

25.00

20.00

15.00

10.00

DRG 16.3

2 106 DRG4.28

for

DRG 16.3

and

(7)

4.2 Correlation between French and English

coefficients for Los Angeles and Deval

5.00

0.00
0.00

for

This equation fits the experimental results for

U41 twin-block sleepers of the French Railways laid
and in TGV lines with a maximum speed 300 km/h
and is applicable for all conditions (different type of
sleepers and fastenings and different ballast quality).
Although it is not a linear relation it could be approximated as such (see relevantly Selig & Waters,
1994/2000). A test of 100 h in the Vibrogir device
is equivalent to a circulation/passing of 360,000,000
t or 396.9 billion lb, which is equivalent to the fatigue that a track panel undergoes during a period of
12 years with a traffic of 120,000 t/day or 132.3 million lb/day, corresponding to the limit Tf2 between
the groups of traffic UIC 1 and 2 according to the
code 714 R of the International Union of Railways
(UIC, (1989)). For more details about hours in Vibrogir and the test the interested reader should read
(Giannakos, 2010 b).

DRG
10 1

0.745 DRG0.36

5.00

10.00

15.00

20.00

25.00

30.00

35.00

40.00

The Hellenic Railways Organization (OSE)

applied the new European Standard (E.N. 13450) for
railway ballast. For this purpose, the N.T.U.A. performed a research program (Loizos et al., 20062007) to contribute to the determination of a relation
between the EN and the French regulations for bal-

45.00

Los Angeles (FR)

Figure 4 Relation between the Los Angeles coefficient according to French regulations (LA-Fr) and the Los Angeles coefficient according to English regulations (LA-En).

last, in order to establish a relation between ballast

hardness and life-cycle as derived from past tests
(Giannakos, 2010 b).
The correlation between Los Angeles coefficient
according to French regulations methodology
LA(Fr) and the Los Angeles coefficient according
to English regulations methodology LA(En) is
given by the Eq. (8):
LA( En) 0.761 LA( Fr ) 0.004
(8)
This correlation (Loizos et al., 2006) is depicted
in Fig. 4 as derived from the N.T.U.A. laboratory
measurements.
The relation between the Deval dry coefficient
according to French regulations methodology
DE(Fr) and the Micro-Deval coefficient according
to English regulations methodology MD(En) is
given by:
MD( En) 0.889 DE ( Fr ) 16.668 (9)
This relation (Loizos et al., 2006-2007, (14)) is
depicted in Fig. 5 as derived from the NTUA laboratory measurements.
These two equations (8) and (9) permit the correlation between the ballast quality of the old Greek
(and French) technical specification determining the
hardness DRi and the ballast quality of the new
technical specification for railway ballast according
to the European Standard EN 13450. These equations permit the use of the above relations predicting
the life-cycle of ballast on track in relation to the
hardness DRi.
TABLE 1: Classification of Railway ballast according to EN 13450
New Specification
Old Specification
K1
DRi 16
K2
DRi 14
K3
DRi 12
K4
DRi 8
K5
LA=35
K6
LA=40

16.00

14.00

Micro-Deval (EN)

12.00

10.00

8.00

6.00

4.00

2.00

3.00

4.00

5.00

6.00

7.00

8.00

According to the Greek new technical specification for railway ballast, in conformity with EN
13450, the following classification is valid as depicted in Table1 with the relevant correlation to the
hardness DRi:
5 RELATION
BETWEEN
QUALITY AND LIFE CYCLE

BALLAST

Due to the fact that concrete sleeper types have

different surfaces, the above relations and the following equation are combined (Giannakos, 2010 b):
t1 4 F1

t2
F2

(10)

where: ti = hours in Vibrogir for two different

types of sleepers, Fi = the seating surface of the
sleeper.
Combining equations (6), (7) and (10) yields the
results presented in Table 2. The table presents the
ballast life-cycle as a function of ballast hardness
and sleeper seating surface and can be used for every
ballast quality (e.g. for eruptive as well as for limestone ballast) and/or concrete sleeper type. The results of Table 2, for the case of U31 twin-block concrete sleeper (187,200 mm2), can be verified in
practice on the Greek Railway network for ballast
maintenance (renewal of the ballast bed or track elevation by adding 15 cm of new layer of ballast on
the top of the existing).
The results of Table 2 lead to a decision of
adopting as minimum hardness for high speed lines
network ,class K1, or DRi=16, as in the Greek
Railways also happened.
TABLE 2 Years of Ballast Life-Cycle on Tracks
Daily Tonnage
30.000t
40.000t
of the Line
Concrete
Seating
Seating Seating
Seating
Sleeper type Surface Surface Surface Surface
Twin-block
187,200 243,600 187,200 243,600
mm2
mm2
mm2
mm2
DRi 8
5
6
4
5
DRi 12
18
21
13
16
DRi 14
28
33
21
25
DRi 16
42
49
31
37

MDE(EN) = -0,889 DE(FR) + 16,668

R2 = 0,61

0.00
2.00

4.3 New classification of railway ballast according

to E.N. 13450 and correlation to DRi

9.00

10.00

11.00

12.00

Deval (FR)

Figure 5 Relation between the Deval dry coefficient according

to French regulations DE(Fr) and the Micro-Deval coefficient
according to English regulations MD(En).

6 CONCLUSIONS
Conditions for the Greek railway network and the
research programs performed led to relations
between the ballast hardness and the fouling of the

Tutumluer E., Dombrow W., Huang H., 2008, Laboratory

Characterization of Coal Dust Fouled Ballast Behavior,
AREMA 2008 Conference, proceedings
U.I.C. International Union of Railways, 1989, code 714R,
Classification of lines for the purpose of Track Maintenance, 3nd edition, January 1989

ballast-bed, at the limit of either the replacement of

the ballast or the elevation of the track, that is the
life-cycle of the material in the ballast-bed under
determined conditions of traffic and passing
tonnage. The relations lead to the determination of a
minimal quality of ballast material for the
requirements of a railway network. In this paper this
case was presented mainly in relation to High Speed
-but also to Heavy Haul-railways as far as the
degradation of ballast both in shock and abrasive
wear is concerned.
REFERENCES
AREMA, 2005, Manual for Railway Engineering
Eisenmann J., 1988, Schotteroberbau Moglichkeiten und
Perspektiven fur die Moderne Bahn, Der Oberbau, Basis
der neunen Bahn, Hestra Verlag 3/1988
Eisenemann Josef, Mattner Lothar, 1984, Auswirkung der
Oberbaukonstruction auf die Schotter - und
Untergrundbeanspruchung, Eisenbahningenieur 35, 1984,3.
Eisenmann Josef, 1980, Verjormungsverhalten der
Schiene.Auswirkungen auf die Oberbaubeanspruchung,
Lagestabilitat und Storungsfunktion, ZEV-Glasers Annalen
Berlin, 02/1980
ELOT 13450, Greek Standard, Aggregates for Railway Ballast, Athens, Greece, 2003.
EN 13450, European Committee for Standardization, Aggregates for Railway Ballast, Brussels, Belgium, 2002.
Giannakos K., 2011, Heavy Haul Railway Track Maintenance
and Use of Resilient versus Stiff Fastenings, TRB - 90th
Annual Meeting, Washington D.C., January 23-27,
accepted to be published in TRR-2011.
Giannakos K., 2010 a, The Use of Strain Attenuating Tie Pads and
its Influence on the Rail Seat Load in Heavy-Haul Railroads,
JRC-2010, Urbana-Champaign, USA, April 27-29.
Giannakos K., 2010 b, Loads on track, Ballast Fouling and
Life-cycle under Dynamic Loading in Railways, Journal of
Transportation Engineering, ASCE, Volume 136, Issue 12, December 2010, p. 1075-1084.
Giannakos K., 2010 c, Stress on ballast-bed and deterioration
of geometry in a Railway Track, Journal of Civil Engineering
and Architecture (ISSN1934-7359), David publishing company, published in English and Traditional Chinese, Chicago,
IL, USA, issue 6, June.
Giannakos K., Loizos A., 2009, Evaluation of actions on concrete sleepers as design loadsInfluence of fastenings, International Journal of Pavement Engineering (IJPE), Volume 11, Issue 3 June 2010 ,pages 197 213.
Giannakos K., 2004, Actions on the Railway Track, Papazissis
publ. Athens, Greece, www.papazisi.gr
Hay W., 1982, Railroad Engineering, John Wiley & Sons,
p.415.
Lecocq C., 1988, La Degradation du Ballast, Memoire,
Conservatoire National des Arts et Metiers, Chaire de
Contructions Civiles.
Loizos A. and researchers, 2006-2007, Research Program,
Contribution to the development of the draft of New Technical Specifications of Railway Ballast according to the
European Standard E.N. 13450, N.T.U.A., Faculty of Civil
Engineering, Sector of Transportation and Infrastructure,
Athens.
Loizos A., Abakoumkin K., Trezos K., and Lymberis K, 19921993, Normal gauge line monoblock sleepers made of prestressed concrete, NTUA/ Athens
Selig E., Waters J., 1994/2000, Track Geotechnology and Substructure Management, Thomas Telford, 1994, reprinted
2000, p. 7.2-7.5, p.8.48 fig 8.56.
SNCF, Direction de l' Equipement, 1985 and modifications,
Specification Technique pour la Fourniture de ballast et de
gravillon, Paris.