53

part of a larger cooperative effort between the Federal 6. P. V. Lade and J. M. Duncan. Stress-Path
Railroad Administration and the Association of American Dependent Behavior of Cohesionless Soil. Journal
Railroads. This paper represents our views and posi- of the Geotechnical Engineering Division, Proc.,
tions and does not necessarily reflect those of the Fed- ASCE, Vol. 102, No. GT 1, 1976, pp. 51-68.
eral Railroad Administration or the Association of 7. Stresses in the Rails, the Ballast, and in the Forma-
American Railroads. tion Resulting From Traffic Loads: Summary of
Results Given in Reports 1-12. Office for Research
REFERENCES and Experiments of the International Union of Rail-
ways, utrecht, Netherlands, Rept. D 71 /RP 13, 1970.
1. R. B. Peck. First Progress Report of the Joint 8. Annual Book of Standards. ASTM, 1976.
Investigation of Methods of Roadbed Stabilization. 9. Standard Specifications for Transportation Mate-
Univ. of Illinois, Urbana-Champaign, Bulletin, rials and Methods of Sampling and Testing.
Vol. 43, No. 66, Reprint Series No. 34, 1946. AASHTO, 1974.
2. J. H. Haynes and E. J. Yoder . Effects of Repeated 10. S. B. Hudson and H. F. Waller. Evaluation of
Loading on Gravel and Crushed-Stone Base- Course Construction Control Procedures: Aggregate-
Materials Used in the AASHO Road Test. HRB, Gradation Variations and Effects. NCHRP, Rept.
Highway Research Record 39, 1963, pp. 82-96. 69, 1969.
3. R. G. Hicks and C. L. Monismith. Factors Influ- 11. British Standard 812. In Methods for Sampling and
encing the Resilient Response of Granular Materials. Testing of Mineral Aggregates, Sands, and Filters.
HRB, Highway Research Record 345, 1971, pp. British Standards Institution, 19 67.
15-31. 12. J. P. Rostron and others. Density Standards for
4. R. D. Barksdale. Repeated-Load Test Evaluation Field Compaction of Granular Bases and Subbases.
of Base-Course Materials. Georgia Institute of NCHRP, Research Results Digest 57, 1974.
Technology, Atlanta, 1972. 13. R. M. Knutson and others. Materials Evaluation
5. J. J. Allen. The Effects of Nonconstant Lateral Study: Ballast and Foundation Materials Research
Pressures on the Resilient Properties of Granular Program. U.S. Department of Transportation,
Materials. Univ. of Illinois, Urbana- Champaign, Rept. FRA-ORD-77-02, 1977.
PhD thesis, 1973.

Ballast and Subgrade Response to

Train Loads
Ernest T. Selig, Department of Civil Engineering, State University of New
York, Buffalo
Andrew Sluz, Transportation Systems Center, Cambridge, Massachusetts

Ballast and subgrade play major roles in the maintenance life of track railments)and increased damage to equipment and lading
structures because they are the source of the cumulative permanent de· unless additional track maintenance is provided or train
formation associated with the deterioration of surface and line. Ballast speed (and hence service level) is reduced. During the
is also the principal means of correcting for this deterioration, which is past few decades, traffic loads have increased and, at
caused by traffic and environmental factors. Better methods are still
needed for the prediction of the effects of the controlling parameters on the same time, economic factors have restricted the
track performance for more rational track design and maintenance plan- amount of maintenance that can be done each year. In
ning. The purpose of this paper is to provide a better understanding of practice, the maintenance cycle frequency is often
these problems and describe progress being made toward their solution. dictated by factors such as the availability of money and
The functions of ballast and subgrade are briefly discussed, and the equipment to do the required work rather than by the
mechanisms of permanent deformation are described. Newly developed amount of track deterioration. Thus, U.S. railroads
or improved methods to measure the in situ physical state of ballast are
presented, and examples of results from field tests are given. The capa- have had increasing difficulty in maintaining the high
bilities of existing analytical track structure models for the prediction service level desired. A recent estimate of the dollar
of track deterioration are assessed. New instrumentation techniques value of the maintenance deficit for all of U.S. railroads
used for measuring the dynamic and permanent strains and deforma- was reported by Ward (1) to be $10 billion.
tions in ballast and subgrade are described. Finally, the characteristics Raymond (~)has rep0rted that approximately 40 per-
of the stress, strain, and deformation in ballast and subgrade are illus- cent of the $100 million that Canadian railroads spend
trated with results of both analytical and experimental studies. on track-structure maintenance relates to ballast
maintenance alone. Therefore, it is a safe assumption
that, at least in dollar value, both the ballast and the
The type and condition of the ballast and the subgrade subgrade parts of the track substructure are important
are key factors in the performance of a track structure. in the upkeep of the service level of the track.
During the service life of a track, permanent strains Ballast maintenance is the means by which the de-
accumulate in its substructure and cause permanent de- terioration of track geometry is controlled, irrespective
formation that is visible as deterioration of surface and of the driving forces behind the geometry changes.
line. This deterioration of the track geometry leads to Whether the structural deficiency is in the ballast, the
decreased safety (including increased potential for de- subgrade, or the track superstructure (the crossties
54

and up), or even if the track degradation has been caused (particularly nos. 2, 4, and 7 cited above), is intended
by an overloading of the normal traffic-carrying capacity to prevent the mutual penetration or intermixing of the
of the track, the correction is usually affected by rework- subgrade and the ballast and to reduce frost penetration
ing the ballast. However, reworking of the ballast, in into the subgrade. Any free-draining sand or gravel
turn, changes its physical state and leaves it prone to materials can serve as a subballast as long as they
increased deformation and, hence, track settlement. meet the proper requirements of a filtering material.
This problem is compounded not only by the limited The mechanical properties of ballast result mostly
amount of maintenance funds but also by an insufficiency from its physical state. Physic al state is defined by
of tools for assessing the cause of the problem and (a) the in-place density and (b) the index properties of
optimizing the use of the maintenance funds. Un- the individual material particles, such as size, distribu-
fortunately, there are no uniform criteria for mainte- tion, shape, angularity, and hardness. The in-place
nance that can be applied to railroads in general. Al- density of ballast is the result of some type of compaction
though many railroads do keep some type of maintenance process. Typically, the resUlting initial density is
records, the definition of performance for any partic- created by maintenance tamping and the subsequent
ular section of track is usually dependent on the sub- density changes result from train traffic and environ-
jective evaluation of the track foreman. mental factors. Experience has shown that tamping
Some solutions to these problems are being developed does not produce a high degree of compaction, and
and becoming available as a result of research sponsored there is clearly little control of geometry when com-
by the Federal Railroad Administration (FRA), Office paction is achieved by train traffic. Therefore, con-
of ,Research and Development. Performance data are sideration is now being given to additional compaction
being generated and measurement tools are being tested during maintenance by using special machines or new
under FRA sponsorship at the facility for accelerated techniques.
service testing (FAST) track at the Transportation Test The need for more information on the subject of
Center in Pueblo, Colorado. ballast compaction has resulted in a research project
In addition to discussing the responses of ballast and at the State University of New York at Buffalo sponsored
subgrade to train loading, this paper will present some by the FRA. In this study the mechanics of ballast com-
of the ideas that are currently being developed under paction and the optimization of the maintenance process
FRA sponsorship, including concepts in performance by using compaction to improve the ballast physical
assessment and methods of measuring the physical state and reduce traffic-induced track settlement are
state of the substructure. being investigated.
Subgrade is the layer of material on which the ballast
FUNCTIONS OF BALLAST, SUBBALLAST, and subballast layers rest; it has the following functions
AND SUBGRADE and requirements.

Ballast is the selected material placed on top of the 1. It must support, without appreciable permanent
track subgrade to support the track structure. Con- deformation, the maximum dynamic, traffic-induced
ventional ballast is a coarse-sized, noncohesive, stresses transmitted through the ballast.
granular material, that usually has a uniform gradation. 2. It must resist the cyclic stresses without exces-
This aggregate layer, tamped under and around the ties, sive cumulative volume or strength reductions.
has several important functions: 3. It must be nonfrost susceptible and volumetrically
stable during cycles of wetting and drying.
1. It limits tie movement by resisting vertical, 4. It must resist softening that could cause pumping
lateral, and longitudinal forces from the train and the and penetration into the ballast.
track.
2. It reduces the stresses from train loads that are The subgrade is a very important component in the
applied to the subgrade of the roadbed and thus limits track structure and has frequently been the cause of
permanent settlement. track failure and the development of poor track. Un-
3. It provides immediate water drainage from the fortunately, in existing track, the subgrade is not in-
track structure. volved in the maintenance operation and little can be
4. It helps to alleviate frost problems. done to alter its characteristics without major track
5. It facilitates maintenance surfacing and lining reconstruction, i.e., removal and replacement of track,
operations. ballast, and subballast.
6. It retards the growth of vegetation and resists The present state of the art of track design as it
the effects of fouling from surface-deposited materials. concerns the ballast and the subgrade is mostly em-
7. It provides support for the ties and the necessary pirical, and the factors that control performance are
resilience to absorb the shock from dynamic loads. poorly understood. Reliance on past experience can
be very misleading, because not only is the experience
Traditionally, angular, crushed, hard stones and at a particular site a complex and unknown function
rocks, uniformly graded to drain freely, free of dust of many factors, but the controlling factors are often
and dirt, and not prone to cementing action have been not even adequately documented. For example, to
considered good ballast materials. However, avail- assess the reasons why a particular section of track is in
ability and economic considerations have often been the poor-track category, it is necessary to know (a)
the prime factors considered in the selection of ballast the characteristics of the ballast and the subgrade, (b)
materials. Thus, a wide variety of materials-such the maintenance history (including frequency and type
as crushed granite, basalt, limestone, slag, and of operation), (c) the environmental history, and (d)
gravel-have been used for ballast in the United States the traffic history. Usually, only the last item is
and Canada. readily available, although the second and third can
Sub ballast is material that is used as a transition layer sometimes be estimated from records. Necessary in-
between the upper layer of large-particle good-quality formation of the characteristics of the ballasts and sub-
ballast and the lower layer of fine-graded subgrade. grades of existing track, however, is practically non-
The subballast used in most new construction, in addi- existent. Even the classification of these materials is
tion to fulfilling some of the functions of the ballast in doubt, not to mention their physical state. Often,
 55

lmowledge of the present conditions of a site based on a The tamping process used in track maintenance is
field examination is all that is possible, because past generally believed to loosen the ballast under a tie and
records are not normally available. decrease the density state that had developed over time
under traffic loading. Tamping also leaves the crib
MECHANISMS OF PERMANENT ballast very loose. Loose crib ballast is a disadvantage
DE FORMATION because it does not contribute significantly to tie lateral
resistance and because it reduces the supporting capacity
The major causes of track settlement can be grouped of the ballast under the tie by providing less lateral con-
into two general categories: (a) repeated loading from finement than does dense crib ballast. For this reason,
rail traffic and (b) environmental factors such as mois- machines to recompact the crib and shoulder ballast after
ture changes, frost action, and mechanical and chemical tamping are now being considered in the United States
weathering. In addition, the subgrade, including the and Canada to speed up the process of traffic-induced
foundation soil strata, can undergo settlement because densification and to provide higher lateral track stability
of consolidation over a period of time. Although this immediately after maintenance.
category is also important and deserves consideration, Very little direct evidence is available to support
this paper will focus on the effects of traffic loading many of these conclusions because in situ methods of
and related phenomena only. measuring the physical state of ballast have been in-
Permanent deformation of track structure results adequate. However, new or refined methods have
from four basic mechanisms of ballast and subgrade recently been developed to provide tools for the study
mechanical behavior. The first is volume reduction of the behavior of ballast. A few examples are shown
or densificatio11. caused by particle rearrangement under in Figures 1 to 6 to illustrate some of the methods and
the cyclic shear straining produced by repeated train- the resulting observations; a detailed evaluation of the
loads. The second is inelastic recovery on unloading techniques is given elsewhere (3).
or stress removal and is a function of both stress A device that measures the resistance of individual
history and stress state. The third is volume reduction ties to lateral force is shown in Figure 1. This type of
caused by particle breakdown from train loading or en- test is the only one extensively used in the past that
vironmental factors. The fourth is subgrade penetra- provides a measure of the physical state of ballast.
tion into ballast voids that allow the ballast to sink into However, it is only an indirect test of physical state, its
the subgrade. The first two apply to both ballast and primary function having been to assess lateral track
subgrade, but the third applies mainly to ballast and stability. Typical results that relate lateral force to
the fourth to subgrade. displacement are shown in Figure 2. Crib and shoulder
Normally, ballast is initially open graded, which compaction following tamping significantly increases
facilitates maintenance operations and allows free tie lateral resistance.
drainage. In service, the ballast gradation changes as a A device that measures ballast stiffness by means of
result of (a) mechanical particle degradation during the vertical settlement of a small loaded plate is shown
construction and maintenance work and under traffic in Figure 3. The 12.7-cm (5-in) diameter plate can be
loading, (b) chemical and weathering degradation from seated anywhere on the crib or shoulder surface or on
environmental changes, and (c) migration of fine the tie-bearing area after the tie is carefully removed.
particles. As the ballast degrades, it loses its open- The preliminary field results (see Figure 4) show that
graded characteristics and, in some cases, cementing there is a significant increase in stiffness in the com-
may occur, which produces a layer of undesirable pacted crib near the rail and a lesser, but still notice-
rigidity and reduces resiliency. able, increase beneath the tie.
Traffic-induced stresses at the ballast-subgrade in- A newly developed method for the measurement of the
terface may be high enough relative to the strength of in-place density of ballast is illustrated in Figure 5.
the subgrade soil that the soil is squeezed into the voids The results of or:..: set of tests (see Figure 6) show the
in the ballast. Under repeated cycles of loading, the density increase achieved by applying crib and shoulder
amo•mt of intermixing may be substantial, particularly compaction after tamping compared with only tamping
with soft soil conditions. Water is trapped in depres- the ballast.
sions that develop under the rail seat, and both drainage
and drying are impeded by the fines in the ballast. Thus, ANALYTICAL TRACK MODELS
the soft conditions and ballast fouling are extended and
the track settlement is self-perpetuating. In existing The principal fuaction of a track model is to interrelate
track, this problem occurs most frequently at rail the components of the track structure so that their com-
joints. In new construction and in major rehabilitation plex interactions in determining the net effect on the
work, this source of track settlement can be reduced stresses, strains, and deformations of the system of the
by the use of a layer of subballast or a filter fabric traffic loads is properly represented. Such a model
placed over the subgrade. provides the foundation for predicting track performance
Ballast fouling can also be caused by internal abrasion and, therefore, the technical and economic feasibility
of the ballast particles or by surface infiltration of fines. of track design and maintenance procedures. Analyses
Whatever the cause, the track-supporting capability of are complicated, however, by the fact that the physical
such ballast decreases when it is wet and the permanent states of the ballast and the subgrade, but especially
settlement under load increases. When the fouled the ballast, change with time. Because maintenance
ballast is dried, its resiliency decreases. In either life is measured in years, these long-term effects
case, maintenance operations to correct surface and must be considered. Considerable effort has been
line are inhibited. devoted to the development of track models that could
Frost heaving may occur in subgrades and ballasts realistically represent the actual behavior of a track
when fine-grained material in the track is wet and then system subjected to various loading conditions. How-
freezes. Soils display volume changes during freezing ever, more research is needed for several reasons: (a)
and thawing, and significant volume increases occur the difficulty of handling the complexities inherent to
when ice lenses develop. Differences in volume changes each component of the track structure and their interac-
in the subgrade soil over short distances along or tions under loads, (b) the lack of adequate understanding
across the track can cause rough track. of the ballast and subgrade behavior to define the model
56

requirements, (c) the lack of field data on track per- the track response or for multidimensional representa-
for m<1nce for validating the models, and (d) the high tions. However, the vertical behavior of the track
computer costs of running the more elaborate of the structure has received the greatest attention. The
models. following is a brief summary of the existing models
Because a railroad track is generally subjected to for vertical-response analysis of conventional railroad
three-dimensional loads, i.e., loads in vertical, lateral, track.
and longitudinal directions, various analytical models Based on the theory of a continuous beam on an elastic
have been suggested for each of these components of foundation, Talbot's work (4) was a significant contribu-
tion to understanding the behavior of a railway track
system under vehicle loading. The concept of "track
Figure 1. Apparatus for determination of lateral resistance of individual
foundation modulus" was introduced, and mathematical
ties. formulations were developed for calculation of the deflec-
tion and moment in the rail. Clarke (5, 6) has summarized
this approach to present a basis for track design pro-
cedures. However, this theory does not include several
important factors that are known to affect the stresses
and deflections in railroad track, such as longitudinal
loads from thermal stresses, a restoring moment
proportional to the rotation of the rail and ties, the
eccentricity of the vertical load on the rail head, or any
track-dynamic effects. In addition, a rather significant
limitation to the approach is that it does not adequately
model the stress-strain behavior of the ballast and the
Figure 2. Effect on 2 4 6 (MM) subgrade.
lateral resistance of crib iD
2
and shoulder compaction
after maintenance
tamping: wooden tie in
limestone ballast.
w
u
a::
0
u..
800

600
l z
0
Figure 5. Apparatus for determination of ballast density by
water-replacement method.
0
...J 400 0
<>: ¢
a:: MICROMETER
w

J
>-- 200 _..--POINT GAGE
<>:
...J
0
0 01 02 03 (IN.)
DISPLACEMENT

Figure 3. Apparatus for determination of ballast stiffness by plate-load

method.

REACTION BEAM

PLATE ON BALLAST

Figure 4. Effect on ballast bearing resistance of crib and

shoulder compaction after maintenance tamping of
limestone ballast.
200
COMPACTED

JOO
~"°'" Figure 6. Effect on ballast density of crib and shoulder
compaction after maintenance tamping of limestone ballast.
o) CRIB AREA

>--
<[ 200
COMPACTED .t 110
COMPACTED
UNDER[-_ ~ _, _ ~ __
N .
;,'; TAMPED ONLY '
<D
...J 100
TIE - - - ... , ' 1:; --
'<D...J \,,
' ? 1,
,,
>- IN[~\'
" '' ·
t: CRIB [
100 Ul
z ____,.____,,' . ~--

~ 90 \1'o-' I,' . TAMPED

>- \:...,' ONLY
a::
0
b. ) UNDER TIE

.1 :p
0'-''--- - - - - - -- -- - - - --
.l1.
so ~~--------------~
 57

Meacham and others (7, 8) and Prause and others (9) dimensional qualities of track structure must be fully
have attempted to overcome some of the limitations of accounted for if the behavior of a track system is to be
the earlier beam-on-elastic-foundation approaches by successfully predicted by using finite-element models.
developing a theoretical method for the determination of The mathematical models developed for the pre-
the track-modulus value. In this method, each com- diction of track performance under dynamic loads have
ponent of track structure is represented by a series of been limited almost entirely to recoverable deforma-
elastic springs, and the spring stiffnesses are com- tions; thus, they do not adequately represent the factors
puted by considering various track parameters (such involved in maintenance-life prediction. However, even
as rail type, tie type, ballast depth, ballast type, sub- the properties associated with recoverable deformation
grade type, and tie spacing). do not fully represent the stress-state-dependent be-
The finite-beam-on-elastic-foundation approach is havior of ballast and soil under cyclic loads. Although
basically similar to the above theories, except that it there has recently been considerable study of the cyclic
considers the tie as a finite beam resting on an elastic behavior of these materials, measures such as resilient
(Winkler-type) foundation as the representation of the modulus should be designated as cyclic-index properties
response of a tie resting on the ballast. The approach rather than as behavioral properties, because they
has been extensively studied by Hetenyi (10), and various represent only a few special stress paths and cannot be
analysis methods for the solution have been presented. used without a factor that compensates for the effect of
For example, Barden (11) has considered a nonuniform stress path.
foundation modulus, andHarrison and others (12) have Currently, the approach to the prediction of per-
included both a nonuniform beam section and a non- manent deformation of track caused by ballast and sub-
uniform foundation. An approximate analytical method grade behavior is patterned after methods used in high-
was developed that makes assumptions about the distri- way flexible-pavement design (18). An elastic track
bution of wheel load over the rail and across the ties. model is used to predict the stresses in the ballast and
The vertical stress distribution with depth in the ballast subgrade from traffic loads, and repeated-load triaxial
and subgrade layers under any given tie is then com- tests are used to determine limiting the threshold stress
puted by using the Boussinesq theory. Ireland (13) has and cumulative strain as a function of confining pressure
presented a design chart for ballast-subballast depth and number of cycles of deviator stress. Repeated
selection versus cohesive strength of subgrade soil by loads are started from a zero load, increased to some
using this approach. predetermined magnitude, and then decreased to zero,
An approach has been developed at the Association thus never putting the sample in extension in the axial
of American Railroads that uses Burmister's multilayer direction. The process is repeated until either the
theory for the ballast and subgrade and a structural desired number of cycles or a limiting permanent strain
model for the rail-tie interaction. The contact between is reached. Track settlement is predicted by summing
a tie and the ballast is represented by a series of cir- the inelastic strains from the triaxial tests for the stress
cular areas that have uniform pressure. The superstruc- conditions determined from the elastic model.
ture and the substructure models were then combined
and extended to form the model termed MULTA (14). MEASURED AND PREDICTED
This is a three-dimensional model; however, theproper- RESPONSES
ties within any layer are constant and cannot be varied
with horizontal position. The nature of the recoverable deformations of ballast
Finite-element methods have also been used for and subgrade, as well as the stresses and strains in
track-structure analysis. Lundgren and others (15) these materials from traffic loading, have been pre -
have developed a two-dimensional system by assuming dieted by using the various available track analytical
plain-strain behavior of a longitudinal section of unit models. These response parameters have also been
thickness along the vertical centerline of the rail. Svec determined experimentally on actual track structures.
and others (16) used a three-dimensional model that The resulting data have been used not only to study the
represents idetailed description of the physical system. track behavior, but also to evaluate the analytical
The rail-tie system was added to the model as simple models. However, the difficulty of measuring stresses
beams, and nonlinear mechanical properties of ballast, and strains, particularly in ballast, has greatly re-
subballast, and subgrade were obtained from laboratory stricted the amount of such data that has been obtained.
tests. One feature of the procedure was the representa- The examples that follow will illustrate the general
tion of the ballast and subballast as no-tension ma- trends in both the analytical and the experimental
terials. However, the model did not have clearly studies.
defined failure criteria. Salem (19) has studied the vertical stress distribu-
Another finite-element model-ILLI-TRACK-has been tions in ballast and subgrade under statically loaded
developed at the University of Illinois (17). This is not wooden ties in a series of laboratory tests that used
a three-dimensional model but consistsessentially of various ballast depths, tie spacings, and types of
two two-dimensional models, one transverse and the ballast. Figure 7 (19) shows that chat, pit-run gravel,
other longitudinal, and uses the output from the longitu- and crushed slag ballast produce nearly the same vertical
dinal model as the input to the transverse model. This pressure below the centerline of a single tie. Figure
gives a three-dimensional effect at less computer cost 8 (19) illustrates the average vertical pressure dis-
than a three-dimensional model. Nonlinear mechanical tribution when varying depths of ballast were used at
properties for the materials were obtained in the a constant tie spacing. Figure 9 (19) illustrates the
laboratory from repeated-load triaxial tests. An in- average vertical pressure distribution on the subgrade
cremental load technique was used to affect a solution. in a longitudinal direction parallel to the tie and below
Explicit failure criteria were developed for the ballast, its centerline at a constant depth of ballast. These
subballast, and subgrade material. However, the model tests indicated that the depth of ballast needed to obtain
does not prevent tension from being transferred across a fairly uniform pressure on the subgrade equals the
the rail base into the tie plate, and further study is tie spacing minus 7 .6 cm (3 in). A comparison of
needed to determine whether the combined two- measured and calculated values also indicated that,
dimensional models accurately represent three- although the shapes of the measured and the calculated
dimensional physical conditions. Certainly, the three- curves are similar, the calculated pressures may be
58

considerably different than the measured data. thickness increases. However, this decrease is a re-
Analytical predictions of track responses have been sult of stress attenuation with depth. Because the rail-
made by using MULTA for a particular range of track seat load and the maximum pressure at the bottom of the
parameters . This analysis assumed uniform properties tie increase as the ballast depth increases, at a com-
under the tie, which is usually not the case, and the mon depth in the ballast, the stresses should actually
ballast was assumed to be much stiffer than the subgrade. increase as the thickness of the ballast layer increases .
The following general trends were observed (14): 4. The maximum vertical stress on the subgrade
surface and the stresses in the subgrade decrease
1. The maximum bending moments at the center of
the tie decrease as the ballast depth increases . How-
ever, the maximum rail-seat bending moments in- Figure 10. Typical gauge layout at FAST track.
crease by a small amount (approximately 5 percent) INSIDE RAIL OUTS IDE RAIL
when the ballast depth increases from 31 to 91 cm (12
to 36 in).
2. The vertical rail displacement and the rail bending
moment decrease and the rail-seat load increases as the
ballast depth increases.
3. The deviatoric and bulk stresses at the middepth
of the ballast decrease rapidly as the ballast-layer

Figure 7. Vertical pressure distribution below

centerline of a single tie.
SECTION B-B PLAN VIEW

- - 200 kPa
'!;:

I-
(/)
0
6
CRUSHED SLAG
Q l ::;:
ct
.J
~ \. PIT RUN u
.J 12 GRAVEL
0 GRANITE BALLAST °"
ct
ID "'
LL
0
18 J
J: 24 TIE LOAD = 20,000 LB s"
I- (15CM)
~
w EXTENSOMETERS - ...._A- 3
Cl 30
0 10 20 30
VERTICAL PRESSURE (LB / SQ IN)
40 50
~ (311~·, ~
SECTION A-A ~ ~

Figure 8. Average vertical pressure distribution on

subgrade at different depths of ballast.
Figure 11. Dynamic response of inductance-coil instruments in FAST
'!;:
track section shown in Figure 10.
0
(/)

'
ID
~
w
er
:::>
(/)
(/)
w 10.0
15 0 COMPRESSIVE
STRAIN 1
( (IN./IN)
o.oooelJlDJ!J!L
0 .0004
er 0 - - - --
~
a) SUBBALLAST STRAIN (COILS A-2)
.J 5.0
ct
(,)
DOWNWARD

~:~:~
i= 00 J_
er 315 21 10.5 o.o 105 21 31.5 DEFLECTION,
w I MM
> d (IN)
DISTANCE FROM CENTER - LINE OF MIDDLE TIE (IN )
b) SUBGRADE DEFLECTION
t
(EXTENSOMETER A-3)
COMPRESSIVE
Figure 9. Average vertical pressure distribution STRAIN, 0 ,004 t /"'\ ,,_...._ r---. ,.-'\
below a single tie on subgrade at ballast depth of E (IN/IN) o ~-\...J---V~'-­
45.7 cm. cl BALLAST STRAIN (COILS A-4)

Q (RAIL SEAT LOAD)

'!;:
i
COMPRESSIVE
10t (\M (\ ,. ., . A j_
0
(/)
STRESS,
<:r (PSI) oLJ_U _ LJ_J/__
__V__L 5-P kPo
0
'.J
l
ID d) SUBGRADE STRESS (GAU GE B-4)
2
w 0
EXTENSION
er
-0.0~~
~
::> 4 STRAIN,
(/)
(/) 0 f (IN./IN)
w
a:
~
6 "'
__l
e) BALLAST STRAIN (COILS B-6)
COMPRESSIVE
.J
ct
(,)
8 STRAIN, o.oo~L;y\_twf\ _ M_A _._f\_
i= 10 f ( IN./ INJ · 0 .0001 ~- L.J . L./ . VI VV
a:
w Note: 1 cm - 0.4 In. f) SUBBALLAST STRAIN (COILS B-7)
>
 59

Figure 12. Cumulative substructure strain and

displacement in FAST track section shown in 0.02 TAMPING7 1-1--~~...__~~l ~
O ~~-IO~O~~
Figure 10. L

r-
-0.01

'
I '
10 JOO + BALLAST STRAIN
-010 BALLAST STRAIN LONGITUDINAL

l VERTICAL
v
SUBGRADE DEFLECTION

IL ~~~~~~~~~~~
Q3 K
(76MM)
r
L ~~~...__~~~~~...__~
01 10 100 0.1 10 100
(9 x10 11 N)
TRAIN LOADING (LOG MGT)

rapidly as the ballast-layer thickness increases. This The current project at FAST is providing important
trend is also largely a result of attenuation of stress new information on this subject, however. Cumulative
with depth. ballast and subballast strain and subgrade deflection
have been measured as a function of total traffic load
The most extensive track response measurement for a variety of track conditions. A typical set of re-
program undertaken to date is that being conducted at the sults is shown in Figure 12 for one track section.
FAST track. This program includes strains in the Strain measurements of this type have not previously
ballast and subballast, vertical stress at the subballast- been available. The slopes of all of these curves de-
subgrade interface, and vertical deformation of the crease rapidly as the traffic increases but permanent
subgrade surface relative to an anchor point approxi- subgrade settlement was still continuing to accumulate
mately 305 cm (10 ft) below this surface. A typical significantly even after 91 million gross metric tons
layout is shown in Figure 10. The strain measure- (100 million gross tons) of loading.
ment method in particular is new and provides im-
portant data not previously available. This instru- CONCLUSIONS
mentation is described in detail elsewhere (20).
A typical set of dynamic records is shown in Figure A general understanding of the functions and behavior
11 and illustrates the elastic response when a three- of ballast, subballast, and subgrade has been achieved.
car train passes slowly over the instrumented wooden Analytical models exist for track structure that predict
tie section. The observations that can be made from elastic response under train loading. A beginning has
these records include the following: been made in obtaining needed field data on dynamic
and permanent strains in the substructure. Field test
1. The permanent strain and deformation from one methods are available to investigate the ballast physical
pass of the train is negligible compared with the elastic state, and data are being obtained on operating track.
components. Continued development of analytical tools for the
2. The 119-metric ton (131-ton) hopper cars produce prediction of stress and strain (both elastic and in-
larger responses than does the 119-metric ton locomo- elastic) that are consistent with material behavior and
tive (because of the higher axle loads). track-structure boundary conditions is necessary. It
3. The variation in stress, strain, or deformation is especially important to account for stress-state-
as each individual axle in a group passes over the gauge dependent material behavior and the effects of the
is small compared with the group average, indicating mechanisms that cause permanent deformation.
that the rail is distributing the axle loads over distances Simultaneously, the collection of field data on track per-
greater than the axle spacing. formance should be encouraged with emphasis on the
4. The vertical strain in the ballast is mostly nega- behavior of ballasts and subgrades. The FAST track is
tive (extension) beneath the center of the tie at the currently the principal source of such information.
centerline of the track. The extension and compression
strains beneath this point in the subballast are approxi- REFERENCES
mately equal.
5. The subgrade deflection is always downward 1. E. J. Ward. Technology and Reviving the Rail-
relative to the unloaded track position, and the sub- roads. Technology Review, July-Aug. 1976.
ballast strains beneath the rail are essentially only 2. G. P. Raymond and 0. Svec. Selection and Per-
compressive. formance of Railroad Ballast. Paper presented
6. The ballast strains are extensional at the mid- at Symposium on Railroad Track Mechanics,
point of the cars because of the spring-up of the rail. Princeton Univ., NJ, April 21-23, 1975.
However, part of this extension could be a result of 3. E. T. Selig, T. S. Yoo, and C. M. Panuccio.
the lifting of the tie from the ballast because the top Mechanics of Ballast Compaction: Vol. 1-Field
part of the strain gauge was attached to the tie rather Methods for Ballast Physical-State Measurement.
than to the ballast surface. Transportation Systems Center, U.S. Department
of Transportation, Cambridge, MA, Interim Rept.,
Analytical models that directly predict permanent Nov. 1977.
ballast strain and cumulative track settlement from 4. A. N. Talbot. Stresses in Railroad Track. Repts.
traffic loading have not been developed. Also, very of Special Committee on Stress in Railroad Track,
few experimental data are available from the field. Proc., AREA, Vol. 19, 1918, pp. 873-1062; Vol.
60

21, 1920, pp. 645-814; Vol. 24, 1923, pp. 297- Study of Track Response and Performance.
453; Vol. 26, 1925, pp. 1081-1245; Vol. 31, 1930, Battelle Columbus Laboratories, OH; U.S. Depart-
pp. 69-336; Vol. 35, 1934, pp. 66-308. ment of Transportation, Transportation Systems
5. C. W. Clarke. Track Loading Fundamentals: 1. Center, Cambridge, MA, Sept. 1977.
Railway Gazette, Vol. 106, No. 2, Jan. 11, 1957, 15. J. R. Lundgren, G. C. Martin, and W. W. Hay. A
pp. 45-48. Simulation Model of Ballast Support and the Modulus
6. C. W. Clarke. Track Loading Fundamentals: 3. of Track Elasticity. Univ. of Illinois, Urbana-
Railway Gazette, Vol. 106, No. 6, Feb. 8, 1957, Champaign, Transportation Series, Vol. 4, Sept.
pp. 157-160 and p. 163. 1970.
7. H. C. Meacham and others. Study of New Track- 16. 0. J. Svec, G. P. Raymond, K. Van Dalen, P. N.
structure Designs. Battelle Memorial Institute, Gaskin, and K. R. Davies. Analytical and Experi-
Columbus, OH; U.S. Department of Transportation, mental Investigation of a Rail Track Structure.
Summary Rept., Aug. 20, 1968. Paper presented at 2nd Symposium on Applications
8. H. C. Meacham and others. Studies for Rail Ve- of Solid Mechanics, Mc Master Univ., Hamilton,
hicle Track Structures. Battelle Memorial In- Ont., June 1974.
stitute, Columbus, OH; Office of High Speed Ground 17. Q. L. Robnett, M. R. Thompson, R. M. Knutson,
Transportation, U.S. Department of Transportation, and S. D. Tayabji. Development of a Structural
Rept. DOT-FR-9-0021, April 1970. Model and Materials Evaluation Procedure. Ballast
9. R. H. Prause and others. Assessment of Design and Foundation Materials Research Program, Univ.
Tools and Criteria for Urban Rail-Track Structure. of Illinois, Urbana-Champaign; Federal Railroad
Battelle Columbus Laboratories, OH; U.S. De- Administration, Rept. DOT-FR-30038, May 1975.
partment of Transportation, Final Rept. DOT-TSC- 18. D. L. Heath, M. J. Shenton, R. W. Sparrow, and
UMTA-74-5, April 1974. J. M. Waters. Design of Conventional Rail Track
10. M. Hetenyi. Beams on Elastic Foundations. Univ. Foundations. Proc., Institution of Civil Engi-
of Michigan Press, Ann Arbor, 1946. neers, London, Vol. 51, Feb. 1972, pp. 251-268.
11. L. Barden. Distribution of Constant Pressure 19. M. T. Salem. Vertical Pressure Distribution in
Under Foundations. Geotechnique, Vol. 13, No. 3, the Ballast Section and on the Subgrade Beneath
Sept. 1962, pp. 181-198. statically Loaded Ties. Department of Civil Engi-
12. H. B. Harrison. General Computer Analysis of neering, Univ. of Illinois, Urbana-Champaign,
Beams on Elastic Foundations. Proc., Institution PhD thesis, 1966.
of Civil Engineers, London, Vol. 55, Ft. 2, Sept. 20. E. T. Selig. Soil Strain Measurement Using In-
1973, pp. 605-618. ductance Coil Method. In Special Technical Pub-
13. H. 0. Ireland. Railroad Subgrade stresses. Proc., lication 584: Performance Monitoring for Geo-
AREA, Bulletin 641, Jan.-Feb. 1973, pp. 382-386. technical Construction, ASTM, Aug. 1975, pp.
14. R. H. Prause and J. C. Kennedy. Parametric 141-158.

Techniques for Evaluating Effects of

Track and Vehicle Wear on
Freight-Car Performance
C. Thomas Jones, ENSCO, Alexandria, Virginia
Donald E. Gray, Federal Railroad Administration

Track and vehicle wear affect the dynamic performance and therefore Finally, statistical processing of wheel-rail forces is used to obtain lateral-
the economic performance of the railcar-track system. A multiphase to-vertical force ratios and lateral wheel forces as functions of the track
test program has been designed to determine the relationship between section. The instrumentation and data-processing techniques designed
the dynamic performance of freight vehicles and track condition, vehicle- for this program proved effective in evaluating freight-car dynamics.
component wear, and variations in track structure. The first part of Evaluation of the effects of variations in track structure on vehicle dy-
this program has been completed, i.e., the development of test, instru- namics led to the following conclusions: (a) track containing unsup-
mentation, and analysis techniques and the determination of their ap- ported bonded joints produced the highest car-body accelerations; (b)
plications to a baseline dynamic-performance test. The test methodology curves greater than 4 degrees and discrete events such as turnouts pro-
involves dynamic testing of a high -travel car and a reference or low-travel duced high accelerations and wheel forces; and (c) variations in track
car. Two test tracks at the Transportation Test Center were used, the and roadbed such as ballast-shoulder width and depth, spiking patterns,
facility for accelerated services testing track and sections of the railroad tie material, and rail anchor type had little if any effect on the dynamic
test track. The instrumentation for each test vehicle included precision response of the vehicle.
accelerometers to measure accelerations on the car body, bolsters, and
trucks and instrumented wheel sets to measure lateral and vertical forces
on the wheels. The analysis of the acceleration data is based on the use
of six degrees of freedom, or rigid-body modes, for each primary mass
(car body and truck). Statistical processing of the computed modal data
The dynamic performance of the railcar-track system
is used to determine the effects of track structure and condition on vehi- has a direct effect on the economics of railroad opera-
cle performance. T ransmissibillty between truck and car body is calcu- tions in terms of lading damage and maintenance costs.
lated to determine the effect of component wear on vehicle performance. This performance changes with accumulated use as a