PB263414

REPORT NO. FRA-OR&D-76-234

ON THE STRESS ANALYSIS OF RAILS AND TIES

I

Arnold D. Kerr

. Princeton University
Department of Civil Engineering
Princeton NJ 08540

SEPTEMBER 1976

INTERIM REPORT

DOCUMENT IS AVAILABLE TO THE U.S. PUBLIC

THROUGH THE NATIONAL TECHNICAL
INFORMATION SERVICE, SPRINGFIELD,
VIRGINIA 22161

Prepared for
U.S. DEPARTMENT OF TRANSPORTATION
FEDERAL RAILROAD ADMINISTRATION
Office of Research and Development
Was h i n g~o~~gj: .. ~~Q_5~D

REPRODUCED BY,
u.s. Department of Commerce
National Technicallnformalion Service
Springfield, Virginia 22161
 NOTICE
This document is disseminated under the sponsorship
of the Department of Transportation in the interest
of information exchange. The United States Govern-
mentassumes no liability for its contents or use
thereof.

NOTICE
The United States Government does not endorse pro-
ducts or manufacturers. Trade or manufacturers'
names appear herein solely because they are con-
sidered essential to the object of this report.
 TECHNICAL REPORT STANDARD TITLE PAGE:
I I. Report No. 2. Government Accession No. 3. Recipient's Catalog No.

FRA-"OR&D-76-284
4. Title and Subtitle S. Report Dot.
September 1976
ON THE STRESS ANALYSIS OF RAILS AND TIES 6. Performing Organization Code

7. Author's) 8. Performing Organization Report No.

DOT-TSC-FRA-76-l6
Arnold D. Kerr Princeton U. 76-TR-7
9. Performinl! Organizolion Nome and Address 10. Work Unit No.
Pri~ceton University* RR619/R632l
Department of Civil Engineering II. Contract or Grant No.
Princeton r NJ 08540 DOT-TSC-900
.'
13. Type of" Report and P.riod Cover.d
~--------------------------------------------------------~
12. Spo'lsoring Agency Name and Address Interim Report
U.S. Department of Transportation Sept. - Nov. 1975
Federal Railroad Administration
G)ffJ_ce of Research and Development 14. Sponsoring Agency Code
Hashington, DC 20590
15. Supplementary Notes
U. S. Department of Transportation
Transportation Systems Center
*Under contract to:
Kendall Square
Cambridqe MA 02142
16. Abstract
The report first reviews the methods presented in the literature for the
stress analysis of the railroad track components and the results of a variety of
validation tests. It was found that the equation Elw iv + kw = q yields deflec-
tions and bending stresses in the rails of longitudinal-"tie and cross-tie tracks
tha"L agree sufficiently (for design purposes) with corresponding track test re-
sults r provided the coefficients that enter the analyses are properly chosen.
'l'his is followed by a review of discussion of the methods for determining the
coefficients that enter these analyses. The report concludes with recommended
analyses and test methods for the determination of stresses in the ,rail-tie
structure.

17. Key Words 18. Distribution Statement 1

Railroad Track Document is available to the U.S. Public
Stress Analysis of Track through the National Technical Informa-
Continuously Supported Beams tion Service r Springfield, Virginia 22161
Rail and Tie Analysis

19. Security Clossif. (of Ihis report) 20. Securi ty Clossi f. (of thi s page) 21. No. of Pages 22. Price
Unclassified Unclassified 50

Form DOT E..1700.7 (B-69)

 PREFACE

This report is a partial result of a research effort whose aim is

to form a basis for the rational design, construction and maintenance

of railroad tracks. This research program is sponsored by the Federal

Railroad Administration, Office of Research & Development, with the

Transportation Systems Center as Program Manager. The present report

was prepared as part of the contract DOT-TSC-900 with Dr. Arnold D.

Kerr as Project Director and Dr. Andrew Kish of the Transportation

Systems Center as Contract Officers Technical Representative.

iii
 METRIC CONVERSION FACTORS

Approximate Conversions to Metric Measures _::: Approximate Conversions from Metric Measures

- :: 5,..., Who. Ya. Kn,. Ma'tlpl, by T. Fi.~ S~"I

S,.... Will,. Y.. knew ••111,.,. ., .e r... ~,...... -_
- ~ .. LENGTH
:;:
LENGTH millimeter. 0.04 inches in
~ em centimeters 0.4 inches ,n
.= meters 3.3 teet ft
in inches 2.5 centimeters em - GO meters 1.1 yards yet
ft feet 30 centimeter. em - tun kilcmelers 0.& miles nti
yd yards 0.9 meters m =-=:--___
mi nftles 1.6 kilometers km ::

- AREA
AREA ~
0\1 - - c::nl- square centimeters 0.16 square inches inZ
J
io square inche. 6.5 square centimet.,. en/- :: m2 square meters 1.2 squant yards .,.r
ttl sqUllr. ~••t 0.09 square meters m2 _ km1 square kilometers 0.4 square miles mil
2
/-1- y~ square yards 0.8 square meters m .. he hectares (10,000 m2) 2.5 ae",s
e:::: mil square mil.. 2.5 square kilometers aun2 _-
acr.s 0.4 hectares hi _ P'I

MASS (weight!
MASS (weight! -=-__-::..-
- 9 grams 0.035 ounce. 01
01 ounce. 28 g~ams g _ -=---:: kg kilograms 2.2 pounds Ib
Ib pounds 0.45 kilograms kg tonnes (1000 kg) 1.1 short ton,
short tons 0,9 tonnes - =_=:--__-::
IZOOO lbl ~

VOLUME _ VOLUME

tsp t.. s~. milliliter. ml - ml ~illiliter. 0.03 fI.uid oonc.. floz

Tbs tablftspoons 15 mi1lilil8ril ml - I hten 2.1 Pints II'
fI 0: fluid ounce. 30 milliliters ml _ liters 1.06 quartS qt
C cups 0.24 liters I I liters 0.26 gallon. ~'
pt pints 0.47 lite" m~ eub~c meter, 35 cub~c fae1 It 3
qt qwa,ts 0.95 liters - m cubiC meters 1.3 cubiC yards yd
gal gallons 3.8 lite" I -
ttl cubiC tN' 0.03 cubiC meters m3 -
yd' cub,c y.. d. 0.76 cubic me'or. m' TEMPERATURE (exact)

TEMPERATURE Inact) _ . c Cel,iu, 9/51'hen F.lvenhoit "F

_ temperature add 32) temperatUN
·11 fehrenheit 5/9 lafter Celsius ·c _
r
teo 'eqture subtracting temperature .,
32) - 0F'2 98.6 212

_ -40 , ,~" '1.,0 I ' , e~ , ~ , I~O, I ' I~O" , 2c;:' ,I

1
. i' iii i i i I
Q. - E -40 -20 0 20 40 eo 80 !~O
! _ _ u.c '7
 TABLE OF CONTENTS

1. INTRODUCTION . . . . . 1

2. THE STRESS ANALYSIS OF THE LONGITUDINAL TIE TRACK SUBJECTED

TO VERTICAL LOADS 3

3. THE STRESS ANALYSIS OF THE CROSS-TIE TRACK SUBJECTED TO

VERTICAL LOADS . . . . . . . 9

4. DETERMINATION OF THE PARAMETERS IN EQ. (3) FOR THE CROSS TIE

TRACK ........ . 21

4.1 THE LOAD PARAMETER q 21

4.2 THE BENDING RIGIDITY EI IN THE VERTICAL PLANE 21

4 .3 THE BASE PARAMETER " ... '" " • lit .. ~ ., ., , 22

5. THE STRESS ANALYSIS OF RAILS SUBJECTED TO NON-CENTRAL AND

LATERAL LOADS . , 31
6. THE WHEEL RAIL CONTACT STRESSES 32
7. EFFECT OF TRAIN SPEED ON TRACK STRESSES 33
8. CONCLUSIONS AND RECOMMENDATIONS 34
REFERENCES . . . . . 36
APPENDIX - REPORT OF INVENTIONS 41

v
 LIST OF FIGURES

Figure
1 A typical railroad track in the U.S. 2

2 Equilibrium position of deformed beam 2

3 Continuously supported beam subjected to a load q(x). 4

4 To the derivation of the bending rigidity of a
longitudinal-tie track 4
5 To influence line method 7
6 Comparison of deflections for longitudinal-tie track. 7

7 Proposed analytical models for the determination

of largest bending moment in rails 10

8 Comparison of bending moments . . . . 12

9 Comparison of stresses for cross-tie track 14

10 To the stress analysis of cross-ties . . . 19

11 Idealized beam model to demonstrate the effect of

the fasteners on the bending rigidity of the rails 23

12 -- To the derivation of expressions for the determina-

tion of the base parameters . . . . . . 23

vii/viii
 1. INTRODUCTION

After the introduction of metal rails, during the 19th century,

two types of track were in use: the longitudinal-tie track and the

cross-tie track. Whereas in the longitudinal-tie track the two metal

rails are continuously supported by longitudinal ties, in the cross-

tie track these rails are supported discretely by cross-ties which are

spaced at a prescribed distance from each other [1]. *

During the second half of the past century the longitudinal-tie

track exhibited various deficiencies and its: use diminished. As a

consequence, in the past several decades the cross-tie track has become

the dominant mode of track construction.

When the cross-tie track was introduced, the wheel loads were very

small and the tie spacing relatively large. For example, around 1800,

the tie spacing was about 1.8 meters [2]. As the wheel loads progres-

sively increased the rail and tie cross-sections increased and the tie

spacing decreased. According to E. Winkler [3], in 1875 the tie spacing

on main lines was about 0.9 meters. A view of a typical track currently

in use in the USA, with even smaller tie spacings,is shown in Fig. 1.

Although the development of the railroad track, up to the present,

was mainly intuitive,based on the trial and error approach, since the

second half of the 19th century railroad engineers have been attempting

to analyze the stresses in the track components.

The purpose of this report is to critically review these analyses

and the related test results, in order to establish which of the proposed

methods are suitable for the analysis of tracks currently in use and the

ones to be built in the future.

*Number in brackets refer to references.

1
 FIG. 1. A TYPICAL RAILROAD TRACK IN THE U.S.

L. undeformed

Q_X_iS___
q(X)

deformed
beam

FIG. 2. EQUILIBRIUM POSITION OF DEFORMED BEAM

2
 2. THE STRESS ANALYSIS OF THE LONGITUDINAL-TIE TRACK
SUBJECTED TO VERTICAL LOADS

In 1867, E. Winkler [4] analyzed the stresses in the

rails of a longitudinal-tie track by considering the rails as a con-

tinuously supported beam. The differential equation for the bending

theory of an elastic beam

(1)
was established by this time., In this equation w(x) is the vertical

deflection at x, EI is the flexural rigidity of the rail and tie, q(x)

is the distributed vertical load, and p(x) is the continuous contact

pressure between the ties and base, as shown in Fig. 2. For the base

response Winkler proposed the relation

p(x) =k w(x)

where k is the base parameter. This is the origin of the well-known

Winkler foundation model. The resulting track equation

is a fourth order ordinary differential equation. It represents the

r~sponse of a beam which is attached to a spring base, as shown in Fig. 3.

In Refs. [3] and [4] Winkler presented a solution of equation (3)

for the special case of a beam of infinite extent subjected to equi-

distant concentrated loads of the same intensity. In order to simplify

the obtained results, Winkler also considered the case of increasing

load spacing and obtained expressions which are the solution for a

single concentrated load (It appears that Winkler did not realize that,

since he refers to them as "approximate formulas" [3]).

In the analysis of the longitudinal-tie track, Winkler stipulated that

(4)

3
 x

FIG. 3. CONTINUOUSLY SUPPORTED BEAM

SUBJECTED TO A LOAD q (x)

q (xl

FIG. 4! TO THE DERIVATION OF THE BENDING RIGIDITY

OF A LONGITUDINAL TIE-TRACK

4
where E I and E I are the flexural rigidities of the rail and tie
r r t t
with respect to ~heir centroidal axes.

The above relation may be derived by assuming that, although the rail

and longitudinal-tie press against each otber, the friction forces in the

contact area are negligible (The calculated maximum bending stresses and

deflections will thus be larger than the actual ones). Noting that at

each point x the vertical displacements of the rail and longitudinal

tie are the same, namely

and that at each x the contact pressures, p*(x), are equal but of opposite

sign, as shown in Fig. 4, the differential equations for rail and tie

may be writtenJrespectivelyJas

iv
E I w = q(x) p~x)
r r (6)

where Ir and It are the moments of inertia of rail and tie with respect

to the respective centroidal axes. Adding these two equations, we

obtain

iv
w = q(x) - p(x)

Comparing eq. (7) with eq. (1) it may be concluded that for the longi-

tudinal-tie track the EI to be used in eq. (1) or (3) is (ErIr + EtI t ) ,

which agrees with eq. (4).

In 1882, J. W. Schwedler [5] in discussing bending stresses in the

rails of a longitudinal-tie track, presented the solution of equation (3)

for the case when the infinite beam is subjected to one concentrated

force P

5
 w(x) = ~~ n(x) (8)

and the corresponding expression for the bending moment

d 2w P
M(x) = -EI dx2 = 40 ll(X)

where

and n(x) = e"""(3x [cos (i3x) + sin (j3x)] (10)

f3x
Jl(x) = e- [cos (/3x) - sin (f3x)]

Schwedler also used the above expressions as influence functions to deter-

mine the effect of several wheel loads. For example, according to this

method, for the three wheel loads PI' P , and P shown in Fig. 5 the
2 3
deflections and bending moments at point 0 are

3
w = ~L
2k n=l

and (n)

3
M= h n=l
~
)'3

For the determination of bending stresses in the rail and the

longitudinal-tie, note that the bending moment at x is

(12)

where Mr and M are the corresponding bending moments in rail and tie
t
and that because of eg. (5)

M (x)
r
= -Er I r w"(x) (13)

= -Et I t w"(x) (14)

6
 FIG. 5. TO INFLUENCE LINE METHOD

7t 7t
-.?35 em
0

·E
E
.£ 2
en
c
0
-.:. 3
u

-4
~
OJ
'0
- - - analysis

FIG. 6. COMPARISON OF DEFLECTIONS FOR LONGITUDINAL-TIE TGACK

and hence
M
_T_=
ET I T

Elimination of M from eg. (12),by using eg. (15),yields

t

E I
M =~M (16)
r EI

and similarly

The largest normal bending stresses in the rails and ties are then

obtained from the well known stress formulas

E Mc
r r
(18)
EI

MtC Et M ct
t
(0) = --= EI'
t max It

where EI is given in (4).

In 1888, H.Zimmermann [6] published a book which contained solu-

tions of eg. (3) for many special cases of interest for the analysis

of a railroad track. Zimmermann like Schwedler, utilized the obtained

solutions to analyze the longitudinal-tie track as well as the ties

of the cross-tie track. Of interest is the presented comparison of

the deflection curves for a longitudinal-tie track caused by two

loads of seven tons each, obtained analytically and from a test, which

are reproduced in Fig. 6. The close agreement between the measured and

calculated ordinates pointed to the conclusion that the linear bending

theory for a beam on a linear Winkler base was sufficient for the analy-

sis of the longitudinal-tie track.

8
 3. THE STRESS ANALYSIS OF THE CROSS-TIE TRACK
SUBJECTED TO VERTICAL LOADS

The development of the analyses of rails for a cross-tie track was

more involved. It started by considering the rail as a beam resting

on discrete rigid, then elastic supports, and then as a continuousZy

supported beam.

In 1875, E. Winkler [3] presented an analysis of bending stresses

in the rails, by considering each rail as an infinitely long elastic

beam which rests on an infinite number of discrete rigid supports, as

shown in Fig. 7(1). For the shown load distribution he found that the

largest possible bending moment is

M = 0.1888 Pa (20)

Realizing the shortcoming of the Winkler assumption of rigid

supports, Zimmermann [6] presented a bending stress determination con-

sidering the rail as a finite elastic beam on four discrete eZastic

supports (in order to simplify the analysis), as shown in Fig. 7(11).

The obtained expression for the largest bending moment, which takes

place under the load P, was given as

8y + 7 Pa
M = 4y + 10 4 (21)

where y is the parameter of the discrete elastic support.

Schwedler [5] proposed to analyze the rail by considering it as a

beam over eight elastic supports sub,jected to one. concentrated force.

For the largest moment Schwedler obtained a similar expression to the

one shown in eq. (21) .

F. Engesser [7J analyzed the rail by considering it

as an infinite beam on equidistant elastic supports subjected to a

9
 20 1.88 a 1.88 a 20
-I· ": ·1- -I- -I- -I •
JP . IP lP !P !P
A A A A Aa a
a A A ZS ZS ZS
a • I ~·I·z. I_ a a a a
-I- a
-j.-
a
-1-- '0 • I.. -I· a -I· ·1 - ·1- ·1·
(I) ·Winkler model

30 30
-I- ·1·

{III) Engesser model

no rna no rna
·1- -I- .. I • -I-
P I•
A A A
t A
IZS lS
. ~P
2S ZS
IP
ZS
,!Pl
ZS ZS
-I- 30 _1. ·1
0 20 .1. .1.
0 30 _1_0.1_. 20 • ~o·l- 30 _I 0'1_

(IV) VMEV model

FIG. 7. PROPOSED ANALYTICAL MODELS FOR THE DETEID1INATION

OF 1ARGEST BENDING MOMENT IN RAILS

10
periodic arrangement of forces~ as shown in Fig. 7(111). For the largest

moment, Engesser obtained the expression

19y + 4 Pa
M = 3y + 1 24

A similar approach was also utilized by a number of other inves-

tigators (many of these papers appeared in the journal Organ fUr die

Fortschritte des Eisenbahnwesens). These and related results are

discussed by R. Hanker in Section B.V.3 of reference [8].

An analysis of a beam on many discrete supports was at the time

rather cumbersome, since it involves the solution of many simultaneous

algebraic equations. It was therefore natural that attempts were made

to analyze the bending stresses in the rails by assuming that also for

a cross-tie track the rails respond like a continuously supported beam.

Early investigators who adopted this approach are:A, Flamache [9] in

1904, S. Timoshenko [10] in 1915, and the ASCE-AREA Special Committee

on Stresses in the Railroad Track [11] in 1917. The tendency of steadily

increasing wheel loads, which was countered by a steady decrease of the

cross-tie spacings, enhanced the justification of the "continuity"

assumption.

The use of the "continuity" assumption, in conjunction with eq. (2),

for the cross-tie track prompted a number of studies to determine whether

this assumption is justified for the track parameters in use.

One approach was to analyze the track ,as apeam on discrete elastic

supports, then as a beam on a continuous Winkler base, and then compare

the obtained results. Such a comparison was performed, for example, by

G. S. Gough [12], E. Czitary [13], A. Wasiutynski [14] and A. D. dePater [15].

11
 -- -
M
In'pL

~ r:J ~
,
7..- ~-
~~
~...t li\ t---
- --- ---
14' ;---- '--- !6j

~~\ / ~I I V o
~
r-

the numbers lndicnte the \oc:ltion or the IUpports

~~ 1Jt '/
1 - discontinuous method.
--- continuous m~thod. .
1"\ ' i1' 1/
JI'-
~t

\ lLu-
"I'l~

(a) Bending moments along the beam; load applied between two
supports

Itt
Fi"
D.I

~]1.1___ ~- r
~r
")
I 1;;'>, - t--J
I?I I
- [9
-~ ~
l y V J~
~\

'\ r.£ (.V

\

~l! ~I
~t1.

~<r

(b ) Benc'ling moments along the bea..rn; loa,C! a,pplied oyer support

FIG. 8. COMPARISON OF BENDING MOMENTS [15]

12
Graphs from Ref. [15] which compare the bending moment distributions,

are reproduced in Fig. 8. Note the good agreement of the shown results.

A more recent comparative study, with error estimate, was presented by

H. Luber [16]. Related results were published by C. B. Biezeno [17],

C. Popp [18-19], G. Hutter [20], and J. P. Ellington [21]. Gough [12]

and Ellington [21] also studied the effect of a missing tie. According

to [21] when the base is relatively soft, for a concentrated load over

the missing tie)the increase in the largest possible bending moment at

the missing tie is about 30%. The increase of the largest possible

bending moment for a relatively rigid base is over 100%.

Another approach was to compare the results based on equation (3)

with corresponding test results obtained using an actuaZ track. For

examples of this approach refer to the studies by the ASCE-AREA Special

Committee on Stresses in Railroad Track [11] [22]. One of the compari-

sons from Reference [11] is reproduced in Fig. 9. Note that also in

this study the results for the bending moments show good agreement.

Because of the agreement found in such comparative studies, and

the absence of a better (and simple) analytical approach, the validity

of the "continuity" assumption, in conjunction with the Winkler

hypothesis (2) was accepted by a n~ber of railroads as a basis for the

analysis also of cross-tie tracks [23- 2 4].

The acceptance of this method was not universal, however, and

many railroads had their own methods of track analysis. To demonstrate

this point let us consider the corresponding developments at the rail-

roads of central Europe.

Since World War I an attempt was made by the central European

railroads (Verein Mitteleuropaischer Eisenbahnverwaltungen,VMEV)

13
 __ £,.~rhucllt"l. 00- ... Co.nllUSit". -_. An:\Irtk:4l.

FIG. 9. COMPARISON OF STRESSES FOR CROSS-TIE TRACK [11]

14
to standardize the analyses of the track, by comparing the available

analyses with test results. As part of this effort, for several years

stresses were measured in the rails of the German and Dutch tracks

which were caused by a train of a prescribed composition [25]. The

obtained results were then compared with the calculated stresses based

on the bending moment formulas of Winkler, Zimmermann, and van Dijk

(who like Zimmermann used discrete elastic supports, but took into

consideration the effect of adjoining wheel loads). On the basis of

this comparison it was concluded that the axle-load spacing is an

important parameter,that the use of discrete elastic supports led to

too high stresses, and that the analysis based on rigid discrete supports

yielded stress values which in the mean agreed with the measured ones.

Based on these conclusions, the Technical Committee of VMEV, at

its meeting held September 16-18, 1930 in MUnster Germany, recommended

an analysis of the rails, developed by the Dutch railroads, which is

based on a weightless beam which rests on rigid discrete supports with

possibility of lift-off and is loaded by a periodic load distribution,

as shown in Fig. 7(rv). The recommended expression for the bending

moments under a load, which is always located in midspan, is

M = 12mn - 7(m + n) + 4 p (23)

16 [3mn _ (m + n)] a

where m, n are wheel set separation parameters shown in Fig. 7 tt~.

For additional details of the recommendations the rearler is referred to

Ref. [25] p. 120. For a discussion of this method of analysis refer to

Hanker [8J p. 39.

As part of t~e above effort, preliminary tEsts were also conducted

for the determh!ation of tlle bas(; parameter k "which enters eq. (3),

15
in order to determine whether eq. (3) is suitable for the analysis

of the cross-tie track, and thus whether the "standard" track analy-

sis of VMEV could be based on eq. (3). Since no conclusive results

were obtained, the Technical Committee recommended (also in 1930)

that member railroads conduct tests to determine k using a standard

test.

The recommended test consisted of loading vertically one tie, which

was separated from the rails by removal of the fasteners, and then by

recording its vertical displacement due to the load. Two "point" loads

were generated by a loaded freight car (about 16 tons) which was equipped

with two hydraulic pumps mounted between the two wheel sets. The

pumps, when activated, pressed against the tie ~

lifting up the car; thus

exerting about 16 tons on the tie ([25] p. 121).

A total of 385 tests were conducted on the lines of the German,

Dutch, and Swiss railroads. The Technical Committee of VMEV could not

detect definite effects of the various types of ballast, of the condi-

tion of the ballast, or of the type of the ties on the vertical tie

displacements. It did notice a strong effect on the tie response by the

type of sub-base, but could not establish a tendency based on sub-base

properties. On the basis of these findings, the Technical Committee

of VMEV, at its meeting in Stockholm May 28-30, 1935, passed a resolu-

tion to recommend to its member railroads not to use eq. (3) for the

analysis of the railroad track [25].

It appears that the main problem with the above study (and the

VMEV resolution) was that the test set-up used to obtain the base para-

meter k, which loads only one tie, is conceptually incorrect.

16
 The first shortcoming is that the base parameter k depends on the

size of the loading area [27] [28]. Thus, the loading with one tie

does not yield the same coefficientkas the loading by a row of closely

spaced ties encountered in an actual track. The conceptual difficul-

ties encountered by Driessen ([2~ p. 125), who observed that the adja-

cent ties when separated from the rails although unloaded also displaced

vertically, are closely connected to this phenomenon.

The second shortcoming is that the material properties of the

ballast and sub-soil, because of their granular character, vary locally.

Thus the loading of only one tie, at different locations along the

track, will necessarily show a wide scatter in the obtained data. This

is very apparent from the test data presented by Driessen ([25] P.123).

For a critical discussion of some of the arguments presented by

Driessen [25],who favored the VMEV decision, refer to R. Hanker ([29]

Section IV). It should be noted that the VMEV decision was made in

spite of the findings of the extensive study by the ASCE-AREA Committee

[11] [22] discussed previously and the opinion of many central European

track experts [30-37], who favored the use of eq. (3) for the analysis

of rail stresses.

In 1937, A. Wasiutynski [14] published the results of an extensive

experimental program performed on a main line track by subjecting it to

moving locomotives and by comparing the obtained results witb thone based

on eq. (3). The presented graphs show general agreement between the

measured and calculated deflections and bending moments for the rails;

thus confirming the findings of the ASCE-AREA Committee [11] [22] that

eq. (3) is suitable for the rail analysis of the cross-tie track.

In the course of the following decades the use of the analysis

17
based on eq. (3) found general acceptance, as evident from the writings

by S. Timoshenko and B. F. Langer [38], C. W. Clarke [39], G. Sauvage [40],

J. Eisenmann [41], H. C. MeachEm, R. H. Prause and J. D. Waddell [42]

and the Association of American Railroads [43] as well as from the books

by W. W. Hay [44], M. Shrinivasan [45], R. Hanker [8], A. Schoen [46],

G. Schramm [47], G. M. Shakhunyants [48], M. A. Frishman and co-authors
[49], and V. V. Basilov and M. A. Chernyshev [50]. However, as pointed

out by Schoen ([46] p. 258), the simple formulas (20) and (21), in spite
of their known deficiencies, are still being used by a number of rail-

roads for the determination of the largest bending moment in the rails

of a cross-tie track.

A shortcoming of track analyses based entirely on eq. (3) was

suggested by the observation that, for example, in front of a locomo-

tive over a certain interval the track lifts off the ballast. Because

of the separation of the rail-tie frame from the ballast, in this domain

eq. (3) is not valid, since k = O. Problems of this type were recently

solved by Y. Weitsman [51].

Once the "continuity" of rail support is adopted, the determina-

tion of the force the raiZ exerts on a tie is very simple; i t is the

contact pressure integrated from half span to half span; or approxi-

mately, the pressure ordinate at the tie multiplied by the center to

center tie spacing. The determined largest force, F ,that each rail
max
could exert on a cross-tie caused by the anticipated wheel loads of a

moving train, are then used for the stress analy.sis of the cross-tie,

as shown in Fig. 10.

A cross-tie anaZysis based on eq. (3) was presented by Zimmermann

[ 6) in 1888. A major shortcoming of this analysis i~) the assumption

18
 Fmax Fmax
l I
I I
Fmax
PI =1])

I" L/2 "', L/2 .:1 ~ L/2 _I. L/2 .. I

(I) Assumed pressure distribution for determination of Mmax at rail seat

l
!IFmax rmax
F
J
I I
2Fmax
P2- ...- P2 = L*b

-I
(II ) AssUmed pressure distribution for determination of M at midspan
max

FIG. 10. TO THE STRESS ANALYSIS OF CROSS-TIES

19
that the tie rests on a uniform linear Winkler base, which is not the

case in the field. In order to prevent "end bound" or "center bound"

ties, the ballast is usually tamped under each rail seat. Thus, the

resulting contact pressure distribution is often as shown in Fig. 10

(For actual test results refer to [11], Second Progress Report, 1920).

Because of the continuously varying contact pressure distribution,

caused by changtng ballast properties due to the moving trains and

environmental factors, the design analysis of ties is often based on

the simplifying assumption that

/"
the contact pressure distribution is

uniform and extends over a distance L or L*, as indicated in Fig. 10

(Distributions I and II). The values Land L* are based on experi-

ence ([52] p. 285, [39] p. 159, [42] I p. 52). Because of the uncer-

tainty in the tie support conditions during the tie service life, this

method, although very simple, yields an upper bound on the expected

tie bending stresses and thus seems sufficient for tie design purposes.

20
 4. DETERMINATION OF THE PARAMETERS IN EQ. (3)
FOR THE CROSS-TIE TRACK

The utilization of eq. (3), for the stress analysis of the rails

and the determination of the forces the rails exert on the cross-ties,

requires the knowledge of three entites: the load parameter q, the

bending rigidity EI, and the base parameter k.

4.1 THE LOAD PARAMETER q

The load parameter is determined from the geometry and the axle

loads of the locomotives and cars to be used on the track under consid-

eration. Thus, once the anticipated rolling stock and admissible train

speeds are established, the parameter q, which enters eq. (3), is known.

Since q usually consists of a large number of concentrated forces

(wheel loads), the resulting deflections and bending moments are deter-

mined using the influence function method, as indicated in eq. (11).

4<2 THE BENDING RIGIDITY EI IN THE VERTICAL PLANE

The bending rigidity of the rail-tie structure in the vertical

plane is usually assumed to be the product of E for rail steel multi-

plied by the 2 moments of inertia of each rail with respect to the

horizontal axis which passes through their centroids.

The determination of EI was the subject of a major controversy in

the early 1930's. Nemcsek [30, 35] and Janicsek [32, 36] were of the

opinion that the cross-ties contribute to the rigidity of the track and

added the I of the ties divided by the tie spacing to the I of the

rails, whereas Saller [31, 34] and Hanker [37] argued that the effect

of the ties is negligible. This controversy ended without agreement

21
among the authors [53].

The inclusion of the I of the cross-ties, as done by Nemcsek and

Janicsek, is definitely not correct. However, the close spacing of

the ties and the rigidity of some of the fasteners currently in use,

may contribute to an increase of the "effective" bending rigidity of

the track, which in turn may have an effect on w(x). To illustrate

this phenomenon, consider the strongly exaggerated situation, shown

in Fig. 11, in which a beam is periodically "rigidized". The effect

of the rigid parts on the global response of the beam is obvious: It

leads to smaller deflections and an increased "effective" rigidity.

Because of the close tie spacing currently in use on main lines,

it may be advisable to measure the rail deflections and stresses in

an actual track, in order to establish whether the "rigidization" of

the rails in the fasteners noticeably affects them. As part of this

test program one could also study the effect of the tie resistance to

rotation about their long axes on the track response, as discussed by

Hanker [29] and Kerr [1]. These rotational resistances do occur, but

are not taken into consideration in eq. (3). As shown by Kerr [1],

the corresponding differential equation is

where p is the rotational proportionality coefficient.

4.3 THE BASE PARAMETER

The Winkler assumption for the base response

p(x) = kw (x)

22
 ___ r
FIG. 11. IDEALIZED BEAM MODEL TO DEMONSTRATE THE EFFECT OF
THE FASTENERS ON THE BENDING RIGIDITY OF THE RAILS

(I)
~por 2P

:IIJJllllllllll~11111111117illll JIIII"l)k
_I. a ·1· a .1_ a _I_ a .1 c a _I_ a -I"

( It) . f'" 2P

. FIG. 12. TO THE DERIVATION OF EXPRESSIONS FOR THE

DETERMINATION OF THE BASE PARAMETERS

23
is an approximation. A major shortcoming o~ this representation is

the absence o~ shear interactions between the vertical base elements.

These questions were discussed by A. D. Kerr [54] in 1964. Refer also

to [27] and [28].

Because of the simplifying assumptions implicit in relation (2),

the k coefficient which enters eq. (3) is not a true constant, but

depends on the size of the loading area. As pointed out previously,

this was one of the reasons why the Zoading of onZy one tie, as sug-

gested and practiced by VMEV, did not yield meaning~ul results. In

this connection note also the more recent test results reported by

F. Birmann [55], which were obtained utilizing this approach. Accor-

ding to H. Nagel [56], this test method is being used by the DB for
the determination of the efficiency of track compaction.

In view of the approximate nature of eq. (2) and of the governing

eq. (3), the method for the determination of k should be such that the

analytically obtained quantities (like rail deflections and/or the

stress distribution in the rails) should represent the corresponding

actual quantities as closelyoopossible. To achieve this objective,

first of all a test for the determination of k should involve a rela-

tively long section of track subjected to vertical loads, similar to

the actual situation in the field. In the following, three methods

for the determination of k are discussed which utilize the entire track

and in which the rails are loaded vertically as shown in Fig. 12. The
~

difference between these methods is the way the k-value is computed

~rom the obtained test data.

One method for the determination of k was used by the AREA-

24
ASCE Special Committee ([11] First Progress Report, 1918) and by

Wasiutynski [14]. Although the agreement found between analytical

and test results for rail deflections and stresses appears satisfac-

tory, questions may be raised regarding the validity of the used

,
method for calculating k from the test data.

To demonstrate this point, let us derive the formulas used in

[11] and [14] for the determination of k, and thus establish the

assumptions they are based on. For this purpose consider the track

subjected to two loads P, as shown in Fig. 12. The corresponding

averaged deflection of the two rails over each tie caused by 2P, is

denoted by w. To avoid future misunderstandings, in the following

m
the foundation modulus for the two rails is denoted by k and the one

for only one rail by k .

r

Consider the two rails as a beam which rests on discrete linearly

elastic springs with spacing a as shown in Fig. 12(1). Assuming

that K is the spring constant for the two rails and K for one rail,
r

it follows from vertical equilibrium that

co 0..

2P = KL: w
n
or P = Kr ~ Wn (24 )
n=-"> n=-oa

Thus, the spring constants, which are assumed not to vary along the

track, are

2P p
K = and K
r
= ---- 00
(25)
L:
n=-tlQ
w
n

where ware the measured (averaged) rail deflections at each tie

n
and

K
r
= K/2

25
 Next, consider the two rails as a beam which rests on a

continuous Winkler base, as shown in Fig. 12(11). Then vertical

eQuilibrium yields
...
J f
ao

2P = p(x)dx =k w(x)dx or (26)

-00 -co

Thus

2P p
k = ----::::---- and k =
j
-eo
w(x)dx
r

and

k
r
= k/2 (21)

In order to find the dependence between K and k, or K and

r
kr , we eQuate the corresponding right hand sides of (24) and (26)
since, for a given test, the P values are the same. This results in

and K
""
~w (28)
rf..J n
f);-oO

or rewritten, noting that the tie spacing a: is constant,

....
f.... krJ....
00
ao K (]O

= ~L:wa a L:wn a
= -E
k w(x)dx and w(x)dx (29)
a n n~-f!:l
n=-G'J
- -

The integral in eQ. (29) is the area formed by the deflection

curve. When the tie spacing is so small that

.....
f
-""
w(x)dx =n~. .wna (30)

then the eQuations in (29) reduce to

K
r
k = aK and k
r = a
(31 )

26
This is the relationship introduced by Timoshenko [10], also presented

by M. Hetenyi [57] p. 27, whose derivation seems to be missing in the

literature (Hanker [37] p. 93; Saller [53] p. 97).

Eliminating K and K from (31), using (25), yields

r
2P P
k = 00
and k = (32 )
r 00

L: wn a L: wn a
n=-oo n=-oo

which are equivalent to the equations in (27), when eq. (30) is valid.

According to the above procedure two loads P are placed on the

track, as shown in Fig. 12, and the deflections w of the rails over

each tie are measured. Then the K or K value is obtained using (25)
r
and the corresponding k or k value using (32). Note that for tests
r
which use more wheel load, the above derivations are valid by replac-

ing P with the sum of the used loads.

Determining the rail foundation modulus in this manner, the ASCE-

AREA Sppcial Committee [11]* and Wasiutynski [14] found good agreement

between the measured rail deflections and rail stresses and the cor-

responding values based on eq. (3). The result of one such comparison

is shown in Fig. 9.

Nevertheless a question may be raised regarding the generalvali-

dity of eq. (25) and eq. (32) from a conceptual viewpoint. For this

purpose, consider a long straight track which, when subjected to loads

2P, deflects as shown in Fig. 12. According to (32), the corresponding

k-value is
P
k =
Al

*)According to [11], 1st Progress Report, Section 47, the expre3l:iion

for the determination of k is the same as the one shown in (32),
except that the denominator is multiplied by the number of ties in
the track test section. In view of the above derivations, this is
incorrect.
27
where Al is the area formed by the straight and deflected rail axes.

Next, imagine the rails replaced by much lighter (or much heavier

rails), without disturbing the base, and then subjected again to 2P.

It is easy to realize that the area formed by the straight and the

new deflection curve, A , will generally.not be equal to A . Thus,

2 l
the calculated k-value will not be the same, although the ties, ballast,

and sub-base (whose properties k represents) are identical for both

cases. It appears that for eq. (32) to yield reasonable k-values, in

the test the contact shape and the EI of the structure (here the rail-

tie system) should be as close as possible to the one to be analyzed.

The other two methods for the determination of k or k also utilize

r
the test set up shown in Fig. 12, but determine the value by comparing

a measured deflection or a strain with the corresponding value based on

eq. (3). This is a proper approach, since the criterion for deter-

mination of k should be that the analytical results represent the

actual ones as closely as possible. These two methods are described in

the following.

In one method [38] the deflections of both rails under the loads

2P are measure& The average of these two deflections, wm, is sub-

stituted in eq. (8). Placing the origin of x at the wheel load, thus

x =0 at P, it follows that
2P
4 k
4E(2I r ) P1h~r'
w
m
= 2k
or w
m
= 2~
(34)

where F is Young's modulus for rail steel and I is the moment of

r
inertia of a rail with respect to the horizontal axis which passes

through its centroid. Comparing the equations in (34) it follows that

28
k = 2kr , in agreement with eq. (27'). Solving each equation in (34)

for k, the only unknown, we obtain

p4
1 1
k = -2 and k
EI
4 r = "4 4
w w
r m m

The other method for the determination of k is based on the mea-

surement of the strain at the bottom of each rail under the load, by

means of strain gauges. The average value of the two measurements,

e: , multiplied by E for rail steel yields the corresponding stress,

m
a= Ee:. Noting that the bending moment M(O) is given in eq.(9),
m m
and that the bending stress at the bottom is a = M(O)z o /1, it follows

that

2P P
a = 4 or
am=~k
(36)

m 4Y4E(2I )' 4 r
r 4EI r

where Zo is the vertical distance between the centroid of the rail

and its bottom surface. Solving for k, the only unknown, we obtain

k = and

Note that when eq. (35) is used, k is determined by equating the

deflection obtained analytically with the measured value at only one

point, namely at P. When eq. (37) is used, k is determined by, equat-

ing the analytical expression and the measured value of the bending

29
strain also at only one point. Thus, in both approaches k is deter-

mined by equating only one analytical and test quantity at one point.

In view of the good agreement between the deflection curves and ben-·

ding moments, based on eq. (3) and the corresponding test data of

actual tracks shown in [llJ and [14], the determination of k by equat-

ing only one analytical and teat quantity at one point may be suf-

ficient.

From the above discussion it follows that the value of k or kr

could also be determined from a least square (or any other suitable)

fit of the measured and calculated deflections, or stresses, or other

quantities of special interest which can be easily measured.

It appears, that the method for the determination of k which uses

relations (35) or (37) or both, because of its simplicity and non-

destructive character, may be the most suitable one also for the deter-

mination of the efficiency of track compaction machinery, if the value

of k is a suitable indic~tor.

In connection with the above discussion of the loads, bending

rigidities, and rail foundation moduli, note that when both rails are

considered as a track-beam then eq. (3) becomes

(38)

whereas when only one rail is considered then the corresponding dif-

ferential equation is

where ~rack = 2 ~ail

30
 5. THE STRESS ANALYSIS OF RAILS SUBJECTED TO
NON-CENTRAL AND LATERAL LOADS

The vertical force of an actual railroad wheel does not act on

the rail centrally. Furthermore a wheel of a moving train exerts

on the rail also lateral forces. Corresponding stress analyses and

test results were presented by S. Timoshenko [58] and by S. Timoshenko

and B. F. Langer [38]. A more recent discussion of these stresses is

contained in a paper by J. Eisenmann [59] and in an ORE Report [60].

It may be of interest to note, however, that to date many rail-

roads do not require such analyses. For example, according to Schoen

[46], Vol. I, p. 250, the effect of the lateral forces is taken into

consideration by choosing a low value for Gall in the bending analysis

for vertical wheel loads. Schrarrrrn n 47l, p. 58) used a similar approach,

by estimating these additional stresses using test data as a guide.

Hay [44] describes a similar procedure (p. 417) as do Frishman and

~

co-authors [49].

31
 6. THE WHEEL-RAIL CONTACT STRESSES

In the wheel-rail contact region, where as much as 16 tons are

transmitted from the wheel to the rail over a very small area, the

stresses deviate considerably from those obtained from the bending

theory of beams discussed above. The occurrence of "shelling" rail

failures prompted many analytical and experimental studies of this

problem. For an early study refer to N. M. Belyaev [61]. For more

recent results refer to N. G~ssl [62], J. Eisenmann [63], and G. C.

Martin and W. W. Hay [64]. For an extensive up-to-date survey of

analyses and tests dealing with rail-wheel contact stresses refer

to B. Paul [65].

These studies are of utmost importance, in view of the continu-

ously increasing car loads and the resulting increase of rail fail-

ures. If it can be proven to the railroad community that

(1) the condition of a main line track for the usual train speeds

has a relatively small effect on the contact stresses, and

(2) that these increasing stresses are directly related to the

increasing number of rail failures

then, unless a major metallurgical breakthrough is achieved and realiz-

ing the practical limitations on the wheel size, this may lead to a

restriction of the axle loads. This in turn may induce the car

builders to increase the number of axles per car or to limit the

weight of cars and locomotives.

32
 ,. EFFECT OF TRAIN SPEED ON TRACK STRESSES

According to test results (for example [14J), the forces a moving

train exerts on the rails of a well maintained track are very little

affected for train speeds of up to about 50 km/h (30 mph). With increas-

ing train speeds, however, the forces increase noticeably. A number of

analytical papers were published in this area. However, to date this

problem because of its complexity (since it involves the dynamic inter-

action between the train and track), is not yet solved. A review of

the published tests and analytical results would require a separate

study.

It should be pointed out, however, that many railroads when analyz-

ing the track components for design purposes, take into account the

effect of the train speed by multiplying the static forces by a "speed

coefficient". For example, according to Schramm([4,J,p. 49), the bending

moment due to dynamic loads is

M = a. M (40)
dyn static

where a. is the train spee~ coefficient

for v < 1,0 km/h (41)

In the above equation v (in km/h) is the largest admissible train speed.

According to the above formula

([ 39], Part ,'.

33
 8. CONCLUSIONS AND RECOMMENDATIONS

The above study leads to the conclusion that, to date, eq. (38)

or (39) is the most suitable one (also because it is very simple) for

the determination of bending stresses in the rails and for the

determination of the forces the rails exert on the cross-ties due

to vertical loads. The extensive test results presented by the ASCE-

AREA Special Committee [11] and by Wasuitynski [14] prove that the

obtained accuracy is sufficient for many design purposes. The coef-

ficients which enter into (38) and (39) are defined in the previous

sections.

For the determination of the k or k value the actual track should

r
be used by loading it with a sZowZy moving (about 10 km/h) locomotive

or loaded car with known axle loads. The speed of the moving vehicles

should be high enough to avoid the occurrence of non-elastic deforma-

tions and low enough not to create noticeable inertia effects in the

track, effects not represented in eq. (3). The calculation of k from

the obtained test data should be based on a comparison of the test

data and the corresponding results based on eq. (3) of the type shown

in (35) and (37).

In conclusion it should be noted that the stress analysis of a

track, as presently prescribed by many railroads, usually involves

the determination of the following quantities for an anticipated load

environment:

• largest bending stress 0(1) in a rail (due to wheel loads and

max
temperature change),

• largest bending stress 0(2) in a cross-tie,

max
 • largest contact pressure cr(3) between the rail and the cross-tie
max
or when tie plates are used between the tie plates and cross-tie,

• largest contact pressure cr(4) between the cross-tie and the ballast,
max
and

• largest contact pressure cr(5) between. the ballast and subsoil.

max

The obtained values cr are then compared with the corresponding

max
allowable stresses crall' which are usually prescribed in the codes of

the various railroads. The design criteria used are

The above analyses have to be conducted taking into consideration

the anticipated wear of the rail heads, because rail wear decreases the

bending rigidity of.the rail. Additionally, it has to be noted that in

a new or newly renovated track, the ballast properties may differ sub-

stantially from those of a track compacted by traffic, and hence for the

same loads the track response may differ. The same applies to the base

properties during the winter and summer periods, because the track is

exposed to the seasonal weather conditions and field tests show that

the ballast and subsoil characterisitics are affected by it.

The above is a simple method of track analysis which for standard

track designs yields reasonable results. This is to be expected because

for standard tracks the design and the various elements are similar (by

definition) and the stipulation of a crall value, for each case n, is

based on closely related test results. However, with increasing wheel

loads and train speeds more sophisticated analyses may be neede·d for

the analysis of the existing tracks as well as for the planned new ones.

35
 REFERENCES

[1] Kerr, A. D., "The stres sand stability analys es of railroad tracks",
Journal of Applied Mechanics, No. 12, 1974. (FHA Report, 1974)

[2] Haarmann, A., "Das Eisenbahn Geleise" (The railroad track, In

German), Vol. I, Geschichtlicher Teil, W. Engelmann, Leipzig, 1891.

[3] Winkler, E., "Der Eisenbahn-Oberbau" (The railroad track, In

German), Third Edition, Verlag von H.. Dominicus, Prag 1875.

[4] Winkler, E., "Die Lehre von der Elastici tat und Festigkeit" (Elas-
ticity and Strength, In German), Verlag von H. Dominicus, Prag,
1867. § 195.

[5] Schwedler, J. W., "On iron permanent way", PY'oc. Institution of

CivilEngineers~ London, 1882, pp. 95..,.118.

[6] Zimmermann, H., "Die Berechnung des Eisenbahnoberbaues" (The

analysis of the railroad track, In German), Verlag W. Ernst &
Sohn, Berlin, 1888.

[7] Engesser, F., "Zur Berechnung des Eisenbahnoberbaues" (To the

analysis of the railroad track, In German), Organ fUr die Fortschritte
des Eisenbahnwesens, 1888.
[8] Hanker, R., "Eisenbahnoberbau" (The railroad track, In German),
Springer-Ver'lag, Wien, 1952.

[9] Flamache, A., "Researches on the bending of rails", Bulletin

International Rail"Jay Congress (English Edition), Vol. 18, 1904.

[10] Timoshenko, S., tlK voprosu 0 prochnosti rels" (To the strength of
rails, In Russian), Trans. Institute of Ways of Communication~
St. Petersburg, Russia, 1915.

[11] ASCE-AREA Special Committee on Stresses in Railroad Track. First

Progress Report in Trans. ASCE, Vol. 82, 1918; Second Progress
Report in Trans. ASCE~Vol. 83,1920; Third Progress Report in
Trans. ASCE, Vol. 86, 1923; Fourth Progress Report in Trans. ASCE,
Vol. 88, 1925; Fifth Progress Report AREA Bulletin, Vol. 31, 1929.

[12] Gough, G. S., "On stresses in railway-track, by an extension of

the theorem of three moments and some deductions therefrom",
The Institution of Civil Engineers, London, Selected Engineering
Papers, No. 147,1933.
[13] Czitary, E., "Beitrag zur Berechnung des Querschwellenoberbaues"
(Contribution to the analysis of the cross-tie track, In German),
Organ fUr die Fortschritte des Eisenbahnwesens, Vol. 91, H. 8, 1936.

36
[14] Wasiutynski, A., "Recherches Experimentales sur les Deformations
Elastiques et Ie Travail de la Superstructure des Chemins de Fer"
(Experimental research on the elastic deformations and stresses
in a railroad track, In French), AnnaZes de Z'Academie des Sciences
Techniques a Varsavie, Vol. IV, Dunod, Paris, 1937.

[15] dePater, A. D., "Calculation for beams of infinite length with

elastic supports", BuUetin InternationaZ RaiZway Congress Associ-
ation (English Edition), Vol. 25, No! 5, 1948.
[16] Luber, H., "Ein Beitrag zur Berechnung des elastisch gelagerten
Eisenbahnoberbaues bei vertikaler Belastung" (A contribution to
the analysis of the elastically supported railroad track subjected
to vertical loads, In German), Dr. Ing. Dissertation, Technical
University of Munich, West Germany, 1961.

[17] Biezeno, C. B., "Zeichnerische Ermittlung der elastischen Linie

eines federnd gestUtzten, statisch unbestimmten Balken" (Graphical
determination of the elastic line of a statically undetermined
beam resting on spring supports, In German), Zeitschrift fUr
angewandte Mathematik und Mechanik, Vol. 4, 1924.
[18] Popp, C., "Zur Berechnung des Durchlaufbalkens auf elastisch
senkbaren, gelenkig angeschlossenen StUtzen (The analysis of a
continuous beam on elastically yielding supports, In German),
ArchivfUr Eisenbahntechnik, Folge 4, 1954.
[19] Popp, C., "Zur Berechnung des Balkens auf elastischer Unterlage"
(To the analysis of beams on an elastic foundation, In German),
Archiv fUr Eisenbahntechnik, Folge 5, 1954.
[20] Hutter, G., "Die Lastverteilende Wirkung von Durchlauftr~gern
infolge elastischer StUtzung" (The load distributing property
of continuous beams due to elastic supports), Dr. Ing. Disser-
tation, Technical University of Munich, West Germany, 1955.

[21] Ellington, J. P., "The beam on discrete elastic supports",

BuZZetin InternationaZ RaiZway Congress Association (English
Edition), Vol. 34, No. 12,1957.

[22] Gruenewaldt, C. V., "Amerikanische Oberbau-Untersuchungen"

(American track investigations, In German), Organ fUr die
Fortschritte des Eisenbahnwesens, Vol. 84, H. 7,1929.
[23] Yamada, T., and Hashiguchi, Y., "The construction of modern track
to carry heavy loads at high speeds, and method of modernizing
old track for such loads and speeds. Facing points which can be
taken at high speeds", BuUetin InternationaZ Rail-way Congress
Association (English Edition), Vol. XIX, No.3, 1937.
[24] Gelson, W. E., "Primary stresses in railway tracks", BuUetin
InternationaZ RaiZway Congress Association (English Edition),
Vol. XXI, No.8, 1939.

37
[25] Driessen, Ch. H. J., "Die einheitliche Berechnung des Oberbaues
im Verein Mitteleuropaischer Eisenbahnverwaltungen" (The standar-
dized analysis of the track in the Union of Central European
Railroads, In German), Organ fUr die Fortschritte des Eisenbahn-
wesens, Vol. 92, Heft 7, 1937.

[26] Engesser, F., "Zur Theorie des Baugrundes" (To the theory of foun-
dations, In German), Zentralblatt der Bauverwaltung, 1893.

[27] Terzaghi, K., "Evaluation of coefficients of subgrade reaction",

Geotechnique, Dec. 1955.

[28J Kerr, A. D., "A study of a new foundation model", Acta Mechanica,
Vol. I, No.2, 1965.

[29 J Hanker, R., "Die Entwicklung der Oberbauberechnung" (The evolu-

tion of the track analyses, In German), Organ fUr die Fortschritte
des Eisenbahnwesens, Vol. 93, Heft 3, 1938.

[30J Nemcsek, J., "Zur Frage der Oberbauberechnung" (To the question
of track analysis, In German), Organ fUr die Forschritte des
Eisenbahnwesens, Vol. 85, Heft 5, 1930.

[31J Saller, H., "Einheitliche Berechnung des Eisenbahnoberbaues"

(Standardized track analysis, In German), Organ fUr die Fortschritte
des Eisenbahnwesens, Vol. 87, Heft 1, 1932.
[32J Janicsek, J., "Zur Frage der einheitlichen Berechnung des Eisen-
bahnoberbaues" (To the question of the standardized analysis of
the railroad track, In German), Organ fur die Fortschritte des
Eisenbahnwesens, Vol. 88, Heft 9, 1933.

[33J Saller, H., "Einheitliche Berechnung des Eisenbahnoberbaues"

(Standardized track analysis, In German), Organ fUr die Fortschritte
des Eisenbahnwesens, Vol. 88, Heft 9, 1933.

[34 J Saller, H., "Zur Frage der einhei tlichen Berechnung des Eisen-
bahngleises" (To the question of the standardized analysis of the
railroad track, In German), Organ fUr die Fortschritte des
Eisenbahnwesens, Vol. 88, Heft 20, 1933.

[35J Nemesdy-Nemcsek, J., "Prlifung von Oberbauberechnungsverfahren an

Hand von Spannungsmessungen" (Validation of methods of track
analysis using stress measurements, In" German), Organ fur die
Fortschritte des Eisenbahnwesens, Vol. 89, Heft 11, 1934.

[36J Jaky-Janicsek, J., "Zur Frage der einheitlichen Berechnung des

Eisenbahnoberbaues" (To the question of the standardized analysis
of the railroad tracks, In German), Organ fUr die Fortschritte
des Eisenbahnwesens, Vol. 89, Heft 11, 1934.
[37 J Hanker, R., "Einhei tliche Langtr8:gerberechnung des Eis enbahnoberbaues [,
(Standardized analysis of the railroad track, In German), Organ fur
die Fortschritte des Eisenbahnwesens, Vol. 90, Heft 5, 1935.

38
[38] Timoshenko, S., and Langer, B. F., "Stresses in railroad track",
Tr'ans. ASME, AppUed Mechanics, Vol. 54, 1932.

[39] Clarke, C. W., "Track loading fundamentals", Parts 1 to 7, The

Railway Gazette, 1957.

[40] Sauvage, G., "La Flexion des Rails" (The bending of rails, In
French), Doctoral Thesis, University of Paris, France, 1965.

[41] Eisenmann, J., "Beanspruchung der Schiene als Trager" (The

stressing of the rails as a beam, In German),Eisenbahntechnische
Rundschau, H. 8, 1969.
[42] Meacham, H. C., Prause, R. H., and Waddell, J. D., "Assessment of
design tools and criteria for urban rail track structures", Vol.
I and II, UMTA Report No. UMTA-MA-06-0025-74-4, April 1974.

[43] AAR report, "Track structures research program. Mathematical

modelling", AAR Technical Center, Sept. 1975.

[44] Hay, W. W., "Railroad Engineering", J. Wiley & Sons, New York,
1953.

[45] Shrinivasan, G., "Modern Permanent Way", Somaya Publications,

Pvt. Ltd. Bombay, 1969.

[46] Schoen, A., "Der Eisenbahnoberbau" (The railroad track, In German),

Vol. I, Transpress VEB, Berlin, East Germany, 1967.

[47] Schramm, G., "Oberbautechnik und Oberbauwirtschaft" (Track tech-

nology and economics, In German), Third Edition, Otto Elsner
Verlagsgesellschaft, Darmstadt, West Germany, 1973.

[48] Shakhunyants, G. M., "Zhelezodorozhnyi Put I " (The railroad track,

In Russian), Tr'anszheldor'izdat, Moscow, 1961.

[49] Frishman, M. A., Voloshko, Yu. D., and Shardin, N. P., "Raschety
Puti na Prochnost i Ustoichivost" (Analysis of the track for
strength and stability, In Russian), DIIT, Dnepropetrovsk, 1968.

[50] Basilov, V. V., and Chernyshev, M. A., "Spravochnik Inzhenera-

Puteitsa" (Handbook of a Track Engineer, In Russian), Vol. I,
Part IX, Izdatelstvo Transport, Moscow, 1972.

[51] Weitsman, Y., "On foundations that react in compression only",

Journal of Applied Mechanics, Dec. 1970.
[52] Mayer, H., "German prestressed concrete sleepers; A brief history
of their evolution", BuUetin International Railway Congr'ess
Association (English Edition), Vol. XXXIV, No.4, 1957.
[53] Concluding remarks by H. Saller, J. Jaky-Janicsek, and J.
Nemesdy-Nemcsek to controversy presented in Refs. [30] to [37].
Organ fUr die Fortschritte des Eisenbahnwesens, Vol. 90, H. 5,
1935.

39
[54] Kerr, A. D., "Elastic and viscoelastic foundation modelS",
Journal of Applied Nechanics, 1964.
[55] Birmann, F., "Neuere Messungen an Gleisen mit ver~chiedenen
Unterschwellungen" (New measurements on tracks with different
support conditions, In German), Eisenbahntechnische Rundschau,
Heft 7, 1957.

[56] Nagel, H., "Messverfahren zur PrUfung der Gleisbettung" (Measuring

methods for track testing, In German), Eisenbahntechnische
Rundschau, Heft 7, 1961.

[57J Hetenyi, M., "Beams on Elastic Foundation", The University of

Michigan Press, Ann Arbor, 1946.

I58] Timoshenko, S., "Method of analysis of statical and dynamical

stresses in rail", Proc. 2nd International Congress for Applied
Mechanics, ZUrich, Switzerland, 1927.

[59J Eisenmann, J., "Stress distribution in the permanent way due to

heavy axle loads and high speeds", Proceedings American Railway
Engineering Association, 1969.

[60] ORE Question D71 "Stress distribution in the rails", Interim

Report No.2, Utrecht 1966.

[61J Belyaev, N. M. IlPrimenenie teorii Hertza k podschetom mestnykh

napryazhenii v tochke soprikasaniya kolesa i relsa" (Application
of the theory of Hertz for the analysis of local stresses in the
contact region of wheel and rail, In Russian), Vestnik Inzhenierov,
Vol. III, Nr. 12, 1917. Reprinted in collected papers by N. M.
Belyaev "Trudy po Teorii Uprugosti i Plastichnosti", Gos. Izd.
Tekh-Teor. Lit. Moscow, 1957.

[62J Gossl, N., "Die Hertzsche Flache zwischen Rad und Schiene bei
Zugkraftbeaufschlagung und ihre Auswirkung auf die ausnutzbare
Haftung" (The Hertz contact area between wheel and rail as affected
by the traction forces and its effect on the available adhesion,
In German), Eisenbahntechnische Rundschau. H. 4,1955.

[63J Eisenmann, J., "Schienekopfbeanspruchung-Vergleich zwischen Theorie

und Praxis" (The stresses in the rail head -- A comparison between
theory and practice, In German), Eisenbahntechnische Rundschau,
H. 10, 1967.

[64J Martin, G. C., and Hay, H. H., "The influence of wheel-rail contact
forces on the formation of rail shells", ASME paper 72-HA/RT-8,
Hinter Meeting, 1972.

[65J Paul, B., "A review of rail-wheel contact .stress problems",

Railroad Track Mechanics and Technology, Proceedings ofasym-
posium, Pergamon Press, 1976.

40
 APPENDIX:

REPORT OF INVENTIONS

After a review of the work performed under this phase of

the contract, it was determined that no technical innovation,

discovery, or invention has been made. The purpose of this

project was to write a survey of the bending stress analyses

of rails and ties and a discussion of the related tests to

be conducted.

41