80 Transportation Research Record 1022

Analysis of Track Support and

Determination of Track Modulus
GERALD P. RAYMOND

ABSTRACT

A design methodology that uses an approach similar to that outlined in the

American Railway Engineering Association Manual for Railway Engineering for
calculating the stresses below cross ties of conventional railway tracks is
presented. There is discussion of the practical effects of track modulus vari-
ability and its possible optimization from the point of view of both static and
dynamic loading with particular reference to ballast and subgrade stress
levels. These stresses are then used to estimate track deformation. Track
deformation may then be used to calculate a conventional track modulus. With
such a technique it is possible to calculate the effects of ballast depth vari-
ability and compressibility of crossties on track response.

The two main requirements of a stable subgrade are where

the provision of sufficient granular or soil-modi-
f ied cover to ensure that overstressing does not oc- E elastic modulus of the beam,
cur and the provision of a granular filter blanket I second moment of area of the beam,
to prevent the piping and thus loss of subgrade z = deflection at a point x from the applied load,
fines from below the track bearing area. To ensure x = distance along the beam from the point of ap-
that over stressing does not occur, the track plication of the load, and
stresses need to be calculated. A means of calculat- U modulus of the elastic support commonly known
ing these stresses that follows an approach similar as the track modulus (note that the track mod-
to that of the American Railway Engineering Associa- ulus as used in common practice and in the
tion's Manual for Railway Engineering is outlined AREA manual relates to a single rail) •
herein. Presentation of subgrade stahility tech-
niques is beyond the main scope of this work. For a single wheel load (P) on an infinitely long
beam the solutions to this equation are well known:

zx = ().P/2U) (cos ).x + sin ).x) exp(-).x) ( 2)

STATIC DESIGN OF CONTINUOUS WELDED RAIL
(P/4).) (cos AX - sin ).x) exp(-).x) (3)
Talbot CJ) has demonstrated that the theory of a
continuous beam on an elastic support, or, more cor- where
rectly, a linear spring (Winkler) foundation gives
calculated bending moments for the rails close to ),, = [U/(4EI)]0.25 (4)
those measured in freshly tamped track. Many inves-
tigations, including extensive ones performed by the The maximum values of z and M are directly below the
Association of American Railways, have confirmed the load at x = o.
validity of this theory. The American Railway Engi-
neering Association (AREA) in its Manual for Railway ).P/(2U) (5)
Engineering (£) also endorses this method of analy-
sis (chapter 22, article 3.2, paragraph 3.2.1, sec- P/(4).) (6)
tion a):
These results are normally given in the form of the
influence chart shown as Figure 1 in which the dis-
Because of the many variables involved in tance to zero moment (X1 ) is used as a distance
tie and ballast track, the calculation of base [e.g., exhibit 1 of chapter 22, AREA manual
track stress and strain cannot be regarded (£)].
as an exact science. However, in-service
tests have shown that the track structure (7)
can be analyzed within acceptahle limits of
Alternatively, but not so commonly, a deformation
accuracy . by considering the rail as equiva-
wavelength of 2n/). is used. The rail seat loads
lent to a continuous beam resting upon uni-
are commonly calculated as
form elastic supports.
Ox = zxUS = (1/2) ).SP (cos AX
This design method results in the representative + sin ).x) exp(-).x) (8)
differential equation for the elastic beam (rail)
subject to a vertical (wheel) load (P) of where s is the spacing of crossties. For a crosstie
placed directly below the load (x = 0)

EI(d'z/dx') + Uz 0 (1) Q0 = P).S/2 (9)

Raymond 81

lp
f' t t t' 't t t, ft' ff t f' t t' 'f 't' t. f 't +' ' ' ' t '' .1
I
'', t 'ft ft ff ff ft f t
"' ~ "' p-PRESSURE PER LINEAR
1.0
-- ~""" .... ..
N
., - Ill
~ N I
M=P.~ G-x~[cos~ -s1Nx~
>- CD
~ II'! j ~~
-
\ I
0.9 ~ l '\.~ RELATIVE VALUES OF

,-- Ill~
J I I \ BENDING MOMENT

[e-~~osx.~ + SINx~j
0.8 1 I "

0.7 - - • .__ ...._ ~ ~

~
Ill
~
' RELATIVE VALUES OF
TRACK DEPRESSIONS
Z=-416:EIU3
0.6 ' l ~ RELATIVE VALUES
OF INTENSITY·
DISTANCE TO POINT
OF ZERO MOMENT
.• x, = ..!!....
4. 4./-u-
4El
I\

-- "' '
Ill\

/l •
"?
-\ OF PRESSURES
0.5
Ill
~
"'
Ill
\"!
' 0
N
\":
MAXIMUM BENDING
MOMENT Mo = P. J-fk 64U
0.4

0.3 ' Ii
I
\
' I\ .
0
N
MAXIMUM TRACK
DEPRESSION
z0 - 4 JG4EIU 3
p

"\"
0.2
I
j \
\ "\. ., MAXIMUM PRESSURE Po
= p . 64EI J;.Y;.
0.1
...... ::; -- - .,
\~
'
a> N
0
,......_d "'0• 0
0
g g 0

0 '\ • - ,·
. m 0
0
N
0
' ~ ,_ - ~ - "'-: 1...o- 0 d
- ~ - -.Y "'
0•
~
N
w -0.1 'l
>w '~ 'l I
;:::: ~ -0.2
....... ,·
N
J. i" I 1' I
<( <[ ,- 1 I I I
d> 0
0::
DISTANCE ALONG RAIL FROM LOAD POINT

E I (3) I I I
~ 45
20 40(1)60 80(2)100 120 140 160(4) 200(5) 250 I 300 350
:i::: (1.13)
I:)

~ 40
Z (I)

-
£35
u..C.BBl -
0
w 30
::::> (.7!1)
...J
~ 25
(,62) 20 40(1) 60 80(2) 100 120(3) 140 160(4) 200
DISTANCE IN INCHES FROM LOAD POINT (m)
FIGURE 1 Influence chart for track design.

As an example of the effect of track modulus on (12)

rail seat loads, Figure 2 shows the variation of the
rail seat loads in the vicinity of a single 147-kN
(33-kip) wheel load on 68-kg/m (136-lb/yd) rail for where the different values of x are the distances of
each wheel load from the position point j,
two different crosstie spacings in which the central
crosstie (crosstie 1) is directly below the load Example calculations of the maximum track deflec-
tion and maximum rail base bending stress are plo _-
(i.e., the crosstie is positioned to produce a maxi-
mum rail seat load), Examination of Figure 2 indi- ted on Figures 3 and 4 for both a single-axle and
cates that, within the limitations of values likely two coupled G-75 trucks (wheel spacing l. 78 m, 1.99
to occur in practice, variation of the track modulus m, and 1.78 m or 70 in., 78.5 in., and 70 in.) sup-
has a much greater influence on the rail seat load ported by either 45-kg/m (90-lb/yd) rail or by 68-
than does crosstie spacing. kg/m (136-lb/yd) rail. Maximum deflection occurs
below the int.erio.: ax l e s and maximum rail str ess C! s
occur below the outer axles. Examination of these
EFFECT OF MULTIPLE WHEELS OR AXLES curves show that as the track modulus decreases from
the region of about 35 to 70 MN/m2 (5 to 10 kips/
Because the preceding theory is based on elastic in. 2 ) of rail there is a rapid increase in deflec-
response, the theorem of superposition applies and tion or rail stress, and if the track modulus in-
the effect of multiple wheels or axles can be simply creases from this region of values a gradual reduc-
calculated as tion of deflection or rail stresses occurs. It may
be concluded that for statically loaded track 35 to
(10) 70 MN/m 2 (5 to 10 kips/in. 2 ) of rail may be an-
Zj
ticipated as an optimized value range of track modu-
(11) lus, at least for initial consideration.
82 Transportation Research Record 1022

KIP/in/in OF RAIL KIP/in/in OF RAIL

0 10 20 30 0 10 20
75 75
(a l
RE COMMEN DED AREA RECOMMENDED AREA
15 15
WOOD-TIE DES IGN WOOD-TIE DESIGN
·VALUE VALUE
+ j

I
610mm (2ftl TIE SPACING
z
50
I 68 kv/m (1361b/yd) RAIL
SINGLE WHEEL LOAD OF
I
.>< I
z 50 I 147 kN U13 KIP)
0
<[
0
...J
II
10
0
<[
0
I 10

0.. ...J
~ :.: 2
w
Vl I 147 kN (33KIP) ~
w
P• 147kN (33KIP)

~ 25
...J
'IW~'tw','IJl,~llllJl"IJl,"'411
21234567
Vl
...J 25 ~'lFlV'W1WZW'
2123456
5 5
<i
ll'.'.

50 100 150 200 50 100 150 200

TRACK MODULUS MN/m/m OF RAIL TRACK MODULUS MN/m/m OF RAIL

FIGURE 2 Effect of track modulus variation on rail seat loads resulting from a single static wheel or axle load: (a) 460-mm
(18-in.) crosstie spacing and (h) 610-mm (24-in.) crosstie spacing.

KIP/in/in OF RAIL KIP/ in/in OF RAIL

0 10 20 30 0 10 20 30 3
;:.....,....-- - - - - - r " - - - - - - - . r -- ----...,--.o.3 ;..,-~~~~~~~~~~~~~~~~~~~.,.--.

NE 20
7 WOOD TIE DESIGN MODULUS
E 'z
E :::;:
I STATIC LOADING I
STAT IC LOADING
z 6 ...J
0 --SINGLE. 294kN {66kip) AXLE. <[ --SINGLE 294kN (66kip) AXLE.
I- ll'.'. 15
<[ - - - INTERIOR AXLE-COUPLED G75 - - - EXTERIOR AXLE-COUPLED G75
:::;: 5 0.2 u.. 2
TRUCKS-294kN (6611.ip)AXLE.S 0 TRUCKS-294kN l66kip)AXLES
ll'.'.
0
u.. ~ w
Vl ~ N
w 4 45kg/m (901b/yd) RAIL :r <[ 45kg/m(901b/yd) RAIL c
o .. u (D

:.: z I- 10 '
0..
u <[ ><'.
<[
ll'.'.
3 Vl
I- Vl
0.1 w
ll'.'.
:::;: 2 1-
::J
:::;:
Vl
:::;:
5 - - - --
x
<[ ::J
:::;: :::;:
x
<[

0 :::;: 0 '---~""-".c.JCJ."""'L----'-----'------'---' 0
0 50 100 I 50 200 50 100 150 200
TRACK MODULUS - M N/m/m OF RAIL TRACK MODULUS- MN/m/m OF RAIL

FIGURE 3 Effect of trnck modulus variation on maximum track FIGURE 4 Effect of track modulus variation on maximum stress
deformation. at base of rail.

CALCULATIONS OF STATIC SUBGRADE STRESSES area. The area of the crosstie bearing surface per
rail seat (~) may be calculated according to the
Similar calculations were made using Equation 12 for AREA manual as
the determination of the rail seat loads obtained
from multiple axles. The rail seat load may then be Ab {per rail seat) b(l - P.rl { l
used to calcu1ate tie bearing pressures and subgrade
stresses. Sufficient evidence is available in the - [C{P. - P.r)/t0.75]) (13)
technical literature to show that where soils are where
unable to resist tensile forces the vertical stress
is given by superposition of Boussinesq's solution b width of crosstie,
for stresses within a semi-infinite elastic solid l length of crosstie,
with surface loading. Below each rail seat the tr distance center to center between rails,
crosstie can be assumed to produce a rectanqularly t thickness of crosstie, and
loaded area on the ballast. This permits the use of c constant = 0.04 (0.018) if dimensions are in
the solutions developed by Love (.:l_) who extended millimeters (inches) (note constant is de-
Boussinesq's solution to a rectangularly loaded pendent on units).
Raymond 83

Alternatively, the area is assumed equal to some contact pressure or vertical stress at zero depth
tamper influence distance on either side of the for a single axle load may be seen to increase as
rail. The AREA method results in rail seat bearing the fourth root of the track modulus (i.e., an in-
areas the width of the crosstie and a length deter- crease of track modulus by a factor of 16 doubles
mined by the formula (Equation 13) that results in the contact stress). With depth the difference in
879 mm (34.6 in) for a crosstie 178 mm (7 in.) deep vertical stress decreases slowly such that, even at
by 2.55 m (8 ft 6 in.) long and 975 mm (38.4 in.) 0.8 m (32 in.), the effect of track modulus variation
for a crosstie 178 mm (7 in.) deep by 2.70 m (9 ft 0 on subgrade stresses from the single axle is consid-
in.) long. For purposes of illustration two rail erable.
seat bearing lengths of 914 mm (36 in.) each below Figure 6 shows similar calculations, using the
each of two rails spaced 1.50 m (60 in.) centerline same crosstie size and spacing, for two coupled G-75
to centerline apart have been assumed. This is con- trucks having 294-kN (33-ton) axle loads. Interac-
sidered the approximate distance influenced by the tion between axle loads is sufficient under the
tamper tines. lower values of track modulus to cause a consider-
Figure 5 shows the solution obtained using dif- able increase in the contact pressure or vertical
ferent track moduli of the underlying maximum verti- stress at zero depth. This reduces the dependence of
cal support stress resulting from a single axle load either maximum contact pressure or maximum subgrade
on 68-kg/m (136-lh/yd) rail and 279-mm (11-in.) stress on the variation of track modulus. Usinq the
crossties spaced at 610-mm (24-in.) centers. The two highest track moduli interaction actually caused

2
lb/in
10 20 30 40

10

(/)
w
I
u
z
20
-279mm (llinl WIDE TIE 610mm(24inl C,C
2/914mm (36inl BEARING LENGTHS
- STATIC LOADING OF TWO 68 kg/m
(1361b/ydl RAIL l.5m (60inl CC
- SINGLE 294kN (66kipl AXLE
30

100 200 300

MAXIMUM VERTICAL STRESS IN TRACK SUPPORT - kPa
FIGURE 5 Effect of track modulus variation on maximum vertical stress in
track support resulting from a single static axle load.

lb/in 2
0 10 20 30 40

TRACK MODULUS ./
(/)
w
Cl::
f-
w 0 .2
:::;:;
I
10
w
(/)
<( co
w
'°w 0.4 J:
u
j:: z
279mm (llin) WIDE TIE 610mm (24in) C. C.
;: 2/914mm (36in) BEARING LENGTHS 20
0
__J
w STATIC LOADING OF TWO 68 kg/m
'°J: 0.6 (1361b/ydl RAIL I.Sm (60in) C. C.

f- INTERIOR AXLE-COUPLED G75

n.
w TRUCKS-294kN (66Kip) AXLES

~
Cl
30

100 200 300

MAXIMUM VERTICAL STRESS IN TRACK SUPPORT - kPo

FIGURE 6 Effect of track modulus variation on maximum vertical stress in

track support resulting from two coupled G- 75 trucks.
84 Transportation Research Record 1022

a reduction in these maximum values from those cal- ciple," which says that two different distributions
culated for a single axle. of force, having the same resultant acting on a
The effect of variation of both track modulus and small part of an elastic body, will produce the same
crosstie spacing is shown in Figure 7. In this fig- stress except in the immediate neighborhood of the
ure is plotted the effect of using a 229-mm (9-in.) loaded part. Thus, provided the ballast is clean and
crosstie at 458-mm (18-in.) centers versus a 279-mm of full section so it can function effectively,
(ll-in.) crosstie at 610-mm (24-in.) centers on changes in crosstie spacing or the size of the bear-
track having moduli of 14 MN/m 2 (2,000 lb/in 2 ) ing area are unlikely to affect the magnitude of the
of rail and 224 MN/m 2 (32,000 lb/in. 2 ) of rail subgrade stresses but would affect the stresses in
with 68-kg/m (136-lb/yd) rail. If no change is made the ballast. These conclusions are verified in Fig-
in track modulus, only about a 10 percent change in ures 8 and 9 that show the distribution of support
contact stress occurs for a change in specified stresses for a single axle using different crosstie
crosstie size and spacing. On the other hand it is spacings and crosstie sizes, respectively. As pre-
clearly evident from Figure 7, and from the maximum viously shown, multiple axles decrease the differ-
rail seat loads shown in Figure 2, that much higher ences shown by single axles and have been omitted
support stresses result from a change in specified from these figures.
track modulus. This effect is more pronounced for a At much greater depths than shown on Figure 5
single axle than for coupled G-75 trucks. (such greater depths are of little practical signif-
In the case of a single axle on 68-kg/m (136- icance) the effect of modulus change would also be
lb/yd) rail changing the modulus from 14 MN/m 2 negligible. As far as normally anticipated ballast
(2,000 lb/in.') of rail to 224 MN/m 2 (32,000 depths are concerned, track modulus variation is
lb/in. 2 ) of rail causes the crosstie-ballast in- much more significant in affecting the magnitude of
terface stress, for the same crosstie spacing, to the subgrade stresses than variation in crosstie
approximately double (i.e., 100 percent increase). spacing or bearing area within the normally used
Interaction of axles is more pronounced on softer 1 imi ts of these variables. This is not to suggest
track because the rail is more effective in spread- that changes in crosstie spacing and bearing area do
ing the load thus reducing the effect of track mod- not affect the track modulus but rather that their
ulus change on support stresses for the coupled G-75 effect on any change in track modulus (which is not
trucks. Changing the track modulus but retaining the given by the theory) has a greater influence on sup-
same rail and crosstie size and spacing reduces the port stresses than their direct effect from theoret-
increase interface stresses to an increase of ap- ical stress distribution theory assuming no change
proximately 30 percent for the 229-mm (9-in.) cross- in track modulus.
ties and approximately 25 percent for the 279-mm Similar calculations, done to show the effect of
(11-in.) crossties. changing rail mass, are shown on Figure 10. Although
In the event of changing from crossties 229 mm (9 there is some effect on the subgrade and ballast
in.) wide at 458-mm (18-in.) spacing to crossties stresses for a single axle on the softer track, in-
279 mm (11 in.) wide at 610-mm (24-in.) spacing, the teraction of wheel loads for the coupled G-75 trucks
effect of modulus change from 14 MN/m 2 (2,000 reduces this difference to a negligible amount. The
lb/ in. 2 ) of rail to 224 MN/m 2 (32,000 lb/in. 2 ) of effect of subgrade stress reduction is much more ap-
rail on the interface stress for a single static parent from a rail mass increase when the track is
axle would be an increase of approximately 120 per- extremely stiff. This reduction occurs both for the
cent. Axle interaction for the two coupled G-75 single and the multiple axle case with maximum
trucks would reduce this increase to approximately stresses being less for multiple axles than for sin-
35 percent. gle axles on extremely stiff track.
A point worth noting from Figure 7 is that as the
depth increases the support vertical stresses re- ESTIMATE OF TRACK MOOULUS
sulting from similar wheel loadings and the same
track modulus tend toward similarity. This, of The solutions developed by Love <1> for a rectangu-
course, is to be expected from "Saint Venant's prin- lar loaded area allow for the calculation of not

lb/in 2
0 10 20 30 40

TRACK MODULUS
If) =14 MN/m/m
w
er
1-
w
::;:0.2
229mm (9in) WIDE TIE 10
w 45Bmm (IBin) C.C.
If)
<1 21914 mm (36 in) BEARING Vl
CD LENGTHS w
w 0.4 J:
u
1- 279mm (llin) WIDE TIE 610mm(24inl z
:;: C.C. -21914 (36inl BEARING LENGTHS
0 20
_J
w STATIC LOADING ON TWO 6Bkg/m
CD 0.6 (136 lb/yd) RAILS l.5m (60in) C.C .
I
I-- ALL AXLES 294kN (66Kip)
CL
w - - SINGLE AXLE
0
- - INTERIOR AXL - COUPL 0 75 30
O.B.__~~..___...._~~__.__~~~-T-R~U-C_K_S~_....~
, 7~8~m'--'--~l=
.9=9~m~~l.7~B~m=>-_,
0 100 200 300
MAXMUM VERTICAL STRESS IN TRACK SUPPORT-kPa
FIGURE 7 Example of effect of major increase in track modulus on
maximum vertical stress in track support.
Raymond 85

2
l b /i n
0 30 40 50

{/)
w
Q'.
I-
~ 0.2
I 10
w
{/)
<l
co 229mm (9in) WIDTH TIE
Cl!
2/914mm (36 in) BEARING LENGTHS w
~ 0.4 :r
u
~ ~
0 279mm (II in) WIDTH TIE 20
--'
w 21914mm (36in) BEARING LENGTHS
co
:r 0.6
I- STATIC LOADING OF
0.. TWO 68kg/m (136 lb/yd) RAIL I.Sm (60in)C.C.
w
0
SINGLE 294 kN (66 Kip) AXLE
30
0 .8
0 100 200 300
MAXIMUM VERTICAL STRESS IN TRACK SUPPORT - kPo

FIGURE 8 Example of effect of crosstie spacing on maximum vertical stress in track

support.

0 10 20

Cf)
w
a:
f-
w
:::!: 0.2
I
w 10
Cf)
<! Cf)
(IJ
w
w J:
f:= 0.4 DIF'F'ERENT WIDTH TIES 610mm (24 in) C.C. u
2/914mm (36in) BEARING LENGTHS z
3::
0
...J STATIC LOADING OF' TWO 68 kg/M 20
w
(IJ (1361b/yd) RAIL l.5m(60in) c.c.
J: 0.6
f-
a... SINGLE 294kN (66 Kip) AXLE
w
0
30
0.8
100 200 300 400 500
MAXIMLJM VERTICAL STRESS IN TRACK SUPPORT kPa
FIGURE9 Example of effect of crosstie size on maximum vertical stress in track support.

only the vertical stress but also the six stresses single axle loading because many field measurements
(three normal and three shear) that define the state of track modulus are obtained using a long loaded
of stress at any point. Knowing the pseudoelastic flatcar with two central loading jacks one above
parameters of the support soil allows the calcula- each rail. The jacks are used to apply, in incre-
tion of the strain at that point, which, by numeri- ments, the equivalent of a single axle load of dif-
cal integration, allows the calculation of support ferent magnitudes. Because of voiding below the
deformation. Addition of such deformations to those crosstie, the modulus is calculated from the defor-
produced by the rail seat load on the crossties via mation recorded between ahout 36 kN (8 kips), the
the tie plates allows the total deformation to be weight on an unloaded truck wheel, and 147 kN (33
estimated for a given set of conditions. Because kips) on each jack.
different track moduli produce different calculated The results shown in Figure 11 were obtained us-
support stresses and rail seat loads, a series of ing 305 mm (12 in,) of clean ballast overlying 610
calculations of total deformations using different mm (24 in.) of clean subballast overlying a heavily
track moduli is required. These results should then compacted silt or clay subgrade, The exact compres-
be compared with the deformations predicted by the sibility of the subgrade can be expected to vary
"beam on elastic support" theory to determine the throughout the year from negligible in extremely dry
track modulus that defines identical deformations. hot weather to a maximum during extremely wet or
It should be clearly understood that the deformation thaw conditions if subjected to freezing. Figure 11
experienced by the rail at the rail seat includes represents calculations for extremely dry subgrade
not only the deformation of the track support but conditions. It may be seen that track on such a sup-
also that of the crosstie and its accessories. An per t constructed with concrete cross ties and stiff
example of the results of such a series of calcula- pads would have a track modulus of about 140 MN/m2
tions is shown in Figure 11. These results are for a (20,000 lb/in. 2 ) of rail. Use of the softer

.. .
86 Transportation Research Record 1022

lb/in 2

Vl
w
a:
I-
~ 0 .2
I
w 10
Vl
<I
CD 229mm WIDTH TIE 610mm (24in)C.C.
Vl
w 2/914mm (36in) BEARING LENGTH w
I-
.4 :x:
u
:;: 279mm WIDTH TIE 610mm (24in) C.C. z
0
....J 2/914mm (36inl BEARING LENGTHS 20
w
CD
STATIC LOADING OF TWO RAILS I.Sm (60in)
:x: 0 .6 C.C.
l-
a.
w
a SINGLE "294kN (66Kip) AXLE
INTERIOR AXLE -COUPLED G75 TRUCKS
L78m l.99m l. 78m 30
0.8 .___ __....,__-"<LL---'--------~'------~--l..--~-....l
0 100 200 300
MAXIMUM VERTICAL STRESS IN TRACK SUPPORT-kPo
FIGURE 10 Example of effect of rail mass on maximum vertical stress in track support.

KIP/in/in OF RAIL (LOG SCALE) An alternative method of calculating the deforma-

10 tions from the support alone is to use Stein-
7.5 ..--; - - - -- - - -- - - - - . - - - - - -..., 0 .3 brenner' s solutions for the settlement of the corner
of a rectangular loaded area extended to deal with a
layered foundation. Steinbrenner's original solution
CV)
<l :::> .:.:: fl) assumes a homogeneous isotropic elastic layer of
.a.5 u~ constant elastic parameters; thus an extension to
oo ~5 deal with layers of different elastic parameters
~ Li.JO ~c
E ~ ~::i; co must be used to obtain the same results as are ob-
E 5.0 °'en ~t ~::i; 0 .2 tained by the use of Love• s solution. Both these
I ~3 E~ o techniques assume that the stresses, but not the
z
0
f=g Jil-- ~ Vl
w strains, are given by homogeneous elastic theory ir-
oO ~ :r
i= :x: respective of any variations in the soil properties.
<[ g:11 + ~-?- u When the stress distribution has been obtained, the
::2 ;!: "" ?--- z
a: ?-f=-· ..... deformations may be calculated from
0 t ?-' "" ""

· If
ES:,::~~ i
LL
w 2.5 0 .1 5z = !{ [az - vi(ax
a
+ ay) ]/Ei}dz (14)
COMPRESSION I
BALLAST/WOOD - I- - where
INTERFACE - - O 6 m1" PAD
_,., - NW(ioo CROSSTIE .
cQ MP RESSlO
COMPRESSION OF CLEAN GRANULAR SUPPORT
a = normal stress,
vi pseudo Poisson ' s ratio of the soil layer, and
0 ..__..'-'--- - - - - - - ' - - -- ' - - - - - - J 0
Ei pseudoelastic modulus of the soil layer.
10 50 100
TRACK MODULUS - MN/rn/rn OF RAIL (LOG SCALE) The integration is usually done numerically by
FIGURE 11 Example of estimating track modulus on the basis of dividing the foundation into a relatively large num-
pseudoelastic properties of track and support components. ber of soil layers and then summing the resulting
layer deformations. Note that the pseudoelastic
properties of granular soils vary with degree of
grooved pads recently installed on the u. s. North- compaction and the properties of cohesive soils with
east Corridor reduces the track modulus to about 69 moisture content or, more correctly, soil suction.
MN/m 2 (10,000 lb/in.2> of rail. The compressi- In addition, an underlying soft cohesive soil tends
bility of hardwood crossties is similar to that of to prevent compaction of an overlying granular soil;
the softer Northeast Corridor pads; however, the thus the selection of appropriate pseudoelastic pa-
major compression observed in tests on wood crosstie rameters requires considerable judgment. The main
track has been the elastic compression and rebound point, however, is that once the selection has been
of the ballast penetration of wood or ballast-cross- made the calculations are relatively simple. Results
tie interface. The magnitude of this interface com- from typical calculations showing the deformation
pression is shown on Figure 11 from which it is seen between the base of the crosstie and the given depth
that wood crosstie track has a considerably lower are shown in Figure 12 for a granular layered de-
track modulus than does concrete crosstie track. The posit in a loosely compacted, and in a densely com-
curves plotted on Figure 11 suggest a track modulus pacted, state making the somewhat simple assumption
of 24 MN/m 2 (3,500 lb/in.2) of rail; however, that the wheel load is taken 50 percent by a central
the variability of the interface compression means rail seat and 25 percent by adjoining rail seats.
considerable variation from this value would not be Because the calculations assume elastic response,
unrealistic. the support deformations are directly proportional
Raymond 87

DEPTH - INCHES DEPTH - INCHES

0 5 10 15 0 5 10 15
0.38 lb)
E P= 33,000 lb
E HEAVILY
COMPACTED
I I SUPPORT
~ ~
w w
0 0
~2 .50 MODULUS - v = 1/2 ~0. 25 0.010
MODULUS - V: 1/2
w 5,100 lb/ in 2 12" BALLAST LLI 51,000 lb/in
2
12" BALLAST
(/) 2,550 4" SUBBALLAST (/)
<l (/) 25,500 4'' SUBBALLAST (/)
2,550 4" FILTER SAND <l 4" FILTER SAND
m LLI m 25,500 LLI
w
1,275 4" QUARTZITE ::r: 12,BOO 4" QUARTZITE I
u LLI u
f-
l
z f- z
l
z z
Q 1. 25 0 .05 0
i== 0 .13 0.005
~ <t
:::!
a:: :::!
a::
•
e
w 6.2mm TRELLEBORG 8m• 0.030"
GEOTEXTILE(+4 lb/in2) =0.010"
0
u.
w 6.2mm TRELLEBORG 8m= 0.030"
0 0 GEOTEXTILE(+41b/i;,) =0 .010"
EVA PAD 8m = 0.005" EVA PAD 8m =0.005"
H H
~---,.-=-=---::+:=---~-=----!--,--~-10. 00
150 300 450 600 750
"""''--~-,!-,,..-~-.....1,_~~--'~~~---'---~~---l o.oo
300 450 600 750
DEPTH-mm DEPTH -mm
FIGURE 12 Example of simplified estimation of contribution to track deformation of support materials: (a) poorly compacted
and (h) heavily compacted.

to any variation in the central rail seat load of M(dw 2 /dt 2 ) + K (w + s) O (17)
73.5 kN (16,500 lb). The calculations should, of
course, be done for the actual distribution of rail where
seat loads produced by the different track moduli to
obtain the best estimate of modulus. w displacement of the unsprung mass from the
The 50 to 25 percent load distribution does, how- static equilibrium position,
ever, permit a quick rough estimate of track modulus s amplitude of the irregularity,
because, from Equation 8, M unsprung mass, and
K track stiffness.
(15)
Track stiffness is related to the track modulus by
where Qo is the maximum rail seat load for the
single wheel load P and 6 0 is the total summa- K • (64 E I u3)0.25 (18)
tion of rail seat deformation at Q0 • Thus a rough
estimate of U can be obtained as The solution to Equation 17 leads to a relationship
for dynamic increment of load of the form
U = P/(2S650) (16)
Pa - Po= cs (M K)o. 5 f(V) (19)
where 6 50 is the total deformation using 50 per-
cent of P as the maximum rail seat load (note Figure where
12 shows only the contribution to the total deforma-
tion due to the soil support of the shown 610 mm or Pa = dynamicwheel load,
24 in. directly below the tie base). P0 static wheel load,
C = a proportionality parameter, and
f(V) = a function of speed.
CALCULATION OF DYNAMIC INCREMENT FOR CONTINUOUS For a well-maintained track made of continuous
WELDED RAIL welded rail that is not corrugated the major track
or wheel irregularity is believed to be due to wheel
Kenney (_1) has shown that on perfect track trans- flats. Data on wheel flat impact obtained by the
versed by perfect wheels track forces rise extremely Association of American Railroads (5) on wheel flat
slowly with speed requiring a speed of about 1600 impacts along with a first trial solution of Equa-
km/hr (l,000 MPH) to reach track resonance. Track tion 19 are shown in Figure 13. According to Equa-
irregularities or wheel irregularities are therefore tion 19, for any given speed and weight of rail the
the principal cause of major dynamic track forces dynamic increment should increase linearly with the
given current (1985) speeds. Any complete theory irregularity depth. Because the wheel flats were
dealing with track and wheel irregularities is carefully made as square flats, the irregularity
clearly complex but because the track mass is much increases with flat length as shown. This is seen as
less than the unsprung mass of a wheel set it ap- in reasonable agreement with the data considering
pears reasonable to neglect the track mass as a the scatter in the results. Also from Equation 19,
first approximation and to take the effect of the for any given speed the axle load should make no
suspension spring as a steady force. The equation of difference to the dynamic increment obtained from a
motion then reduces to given wheel flat and this is reflected in the theo-

.. .
88 Transportation Research Record 1022

1001b(50kg)RAIL 1311b(65kgl RAIL that suggests an 8 percent increase in impact load-

0 64 89 114 0 (mm) 64 89 114 ing on wood cross ties. For concrete cross tie track
with stiff pads and stiff subgrades (e.g., dried out
clay) , the dynamic increment is estimated at an
additional 98 percent, whereas soft pads would re-
duce this increment to 55 percent. For 100-ton cars
on concrete crosstie track the increment would be
69
105 and 59 percent for the stiff and soft pad, re-
(/) spectively. However, due to the nonlinear compres-
"' 0 l - - - ---1--1---1+-- - - . l - - l -- -I-< 0 sion of the components of the track structure, the
(/) 30 1-- --+- +--l-I 207 difference in track moduli for wood and concrete
(/)
w crossties to be applied to the dynamic increment is
~ 20
(/)
1---- --- ·. . .'++-~~=~~+'"'"'-"-'"-' 138 probably not as numerically diverse as shown in Fig-
0
ure 11 because the track has been compressed by the
69 a.. dead load when the impact of the wheel flat occurs.
"' Field data in support of differences of dynamic
0 wheel increment associated with different track
_J 30 t-- ---l--+---lr+-- - - + -- + --H 207 stiffness, assembled from published data for passen-
~
a: ger train wheel loading on the u.s. Northeast Corri-
138 dor (NEC), are shown in Figure 14 (6,7). It may be
seen that more than 90 percent of all wheels give
similar loadings consistent with near-perfect wheels
on track of similar roughness. The few percent of
0 wheels having major imperfections are those, accord-
21/2 3 1/24V2 O
ing to theory, that would show major differences in
LENGTH OF FLAT WHEEL SPOT (INCHES) dynamic increment and this difference is seen (Fig-
DATA POINTS FROM AREA 1952 V.53
ure 14) to be largest between wood crosstie track
FIGURE 13 Compariaon of aimplified theory for and concrete crosstie track. Although the differences
flat wheel apot with experimental field data in measured wheel load on concrete crosstie track
obtained by Aaaociation of American Railways. are small but nevertheless measurable, the effect of
a softer track on the rail seat bending moments of
the differently padded concrete crossties is clearly
retical curves shown. Thus, from the test data ob- evident. Because of the better load-spreading capa-
tained, the dynamic increment for wheel flats from bility of the same rail section on softer track, the
other vehicles and other track constructions may he bending moments associated with softer pads are con-
obtained as siderably less than those associated with the
stiffer pads when the dynamic increment becomes
Pa - P 0 is proportional to M0.5 u0.375 (20) large (i.e., at small occurrence levels). It is un-
derstood that the experimental results have (rightly
The data of Figure 13 were obtained using a 70-ton according to theory) led to the adoption of softer
(net) freight car with an unsprung axle mass of 1705 pads for all further concrete crosstie installations
kg ( 3, 750 lb) on wood cross tie track. One hundred on the NEC and the use of flat wheel detectors as an
ton (net) cars with 20 percent heavier axle loads addition to the normal visual inspection that forms
are in more common use today (1985). These cars have part of normal railroad practice for restricting the
an unsprung axle mass of about 1977 kg (4,350 lb) access of wheels with flats to track. It is clear

(KIP-IN) (KIP)
-~2~ 0~
0 -....:0r----=2~0~0;.._~
4~0~0--=60=-:;0 0 30 45 60 75 90
..J 100..-- 10 0;:---.;,c;--- ;;,.::;.---'-;=---.;;=-- -'-'1"---"'-'i
w
>
w
..J
zLLJ 10 10
>
5
LLJ w
€; I
_J

m
<{ \ u"'
(/)

WOOD CROSSTIE LL
(/) FLEXIBLE LL
RESULTS NEC \ 0
~ PADS
~ PASSENGER ,
a.. 0.1 0.1 TRAF"FIC AHLB ECK
0 et al (1976l - \
<{ I
g 5mm I
6.5mm
~ 0.01 0 .01 CONCRETE CROSSTIE
RESULTS FROM NEC PASSANGER
1- CONCRETE TRAFFIC EXCEEDING llOkmlh (70mph)
z CROSSTI E DEAN et al. (1982-HH -41)
LLJ
u
a:: .0 01 U....- - ' - - - - ' - - - - ' - - - - '--' 0.001.__ _ _.L--_ _ _...__ _ _. . __ ____,
LLJ -20 0 20 40 60 0 100 200 300 400
a. (kN-m) (kN)
(a)RAIL SEAT BENDING MOMENT (b) VERTICAL WHEEL LOAD
FIGURE 14 Meaaured dynamic wheel loads for Northeaat Corridor passenger traffic (6, 7).
Raymond 89

that, although no optimization of track modulus has have been measured and thus the calculation illus-
been established or is evident for dynamically trates the negative effect of having extremely stiff
loaded track, there is a penalty to be paid by over- track associated with the high unsprung mass of
stiffening track that is subject to wheel and rail freight vehicles.
irregularities, which presently (1985) is occurring
on North American track. Optimum track stiffness for
dynamic loading is probably less than that for BALLAST DEPTH DESIGN
static loading.
Clearly evident from Figure 15 is the dramatic ef-
fect that a higher track modulus has on the dynamic
TYPICAL STATIC PLUS DYNAMIC SOLUTION increment. Fortunately concrete crosstie track and
its higher track moduli are generally associated
Figure 15 shows results from a typical calculation with firmer, and thus generally stronger, subgrades.
based on a 51-mm (2-in.) square flat on both wheels In addition, wheels do not generally have carefully
of one inside axle of two fully loaded coupled G-75 made square flats associated with "worst condi-
trucks (i.e., 30-tonne or 66-kip axle load) for tions.• Thus stresses as high as the upper limit
three track conditions. The 51-mm (2-in.) flat is shown on Figure 15 are the exception not the rule.
the maximum permitted by Association of American Typical foundation design is often based on dead
Railroad Cll.l recommendations. The track conditions load plus some percentage (often 50 percent) of the
were track made from wood crossties 229 mm (9 in.) live loadi the reasoning is that safe bearing
wide with 458-mm (18-in.) spacing giving an assumed stresses include some measure of safety factor and
track modulus of 14 MN/m 2 (2 kip/in. 2 ) of raili may thus be exceeded on a limited basis. As an ex-
track made from concrete crossties with soft pads ample, the Manual for Railway Engineering of the
279 mm (11 in.) wide with 610-mm (24-in.) spacing American Railway Engineering Association (2) sug-
giving an assumed track modulus of 84 MN/m 2 (12 gests a dynamic increment of approximately 1-percent
kip/in. 2 ) of raili and track made from similar of static loading for each 1. 6 km/hr (1. 0 MPH) of
crossties with stiff pads giving an assumed track the maximum track speed although the writer would
modulus of 224 MN/m2 (32 kip/in. 2 ) of rail. question the sufficiency of the recommendation for
Figure 15 shows the static loading conditions for the more recently introduced stiffer concrete cross-
the maximum vertical support stress along with a tie track. On the basis of percentage live load
combination of both static and dynamic conditions. chosen, the limits, after proportioning, may be re-
The total dynamic axle (static + increment) loads lated to the safe bearing stress established for the
for the three assumed moduli were calculated as 294 subgrade soil in question to obtain the required
kN (66 kips), 447 kN (100 kips), and 926 kN (208 granular (subballast plus ballast) design depth.
kips) , respectively. These values are likely to be Figure 16 shows the typical safe bearing stress val-
higher than generally measured in track because ues or related empirical test values developed for
wheel flats are never square in practice and are use in highway and airport design C2l, both of which
normally taken out of service before becoming the require much smaller surface deflection than do
maximum allowable. The concrete cross tie moduli are railways for good performance. Because railroads are
also somewhat higher than estimated in Figure 11. more flexible than major highways it is this
These higher values were used because such values writer's opinion that the limits shown in Figure 16

2
lb/in

Cf)

~ 0.2
1-
w 10
:::;,
I
279mm (llinl WIDE TIE 610mm (24in) C.C.
w
Cf) 2/914 mm (36in) BEARING LENGTHS Cf)
<l w
en I
w 0.4 u
z
1-
3:
0
_J 20
w
en STRESSES BELOW INTERIOR AXLE OF COUPLED
I
I- G75 TRUCK - 294 kN (66Kip) AXLES ON TWO
68 kg/m ( 36 lb/yd) RAIL 1.5 m (60.:.i:.:.n,_).:::C.:.;.C:.:·-----
~ 0 .6
0
- - - - STATIC LOADING ii.7Sm ~
l.99m +l.7~m ~
DYNAMIC LOADING 51mm (2inl
SQUARE WHEEL FLAT

30

MAXIMUM VERTICAL STESS IN TRACK SUPPORT - kPa

FIGURE 15 Static and dynamic loading from 51-mm (2-in.) square wheel flat effect on calculated vertical track
support stresses.
90 Transportation Research Record 1022

of stresses. These stresses were used to calculate

CALIFORNIA BEARING RATIO- CBR
strains on the basis of pseudoelastic properties for
2 3 4 5 678910 15 20 30 50 709 the different soil layers. The resulting deforma-
UNIFIED SOIL GP GW tions that were calculated to increase with an in-
CLASSIFICATION GM crease in track modulus were compared with the de-
GC
SW formations from beam on elastic support theory whose
deformations decrease with an increase in track mod-
p ulus. It is suggested that the best design track
SC modulus occurs where both solutions result in the
same deformations (assuming correctness of estimated
component track properties) • An optimum track modu-
lus based on static loading is suggested to occur at
about 35 to 70 MN/m 2 (5 to 10 kip/in.2) of rail,
RESISTANCE VALUE- R
20 30 40 50 55 60 70 8085 and this is regarded as possibly too high on the
basis of dynamic loading for present (1985) North
American track and vehicle standards.
MODULUS OF SOIL REACTION- k It is shown theoretically that the variation of
100 150 200 250 400 600 BOO track modulus has a major effect on the track compo-
lb/in3 nent stresses and that limited field data support
MN/m the theoretical conclusions.

ACKNOWLEDGMENT

The work reported forms part of general studies

kPo funded partly by Canadian Pacific Rail, Canadian
National Rail, Via Rail, and Transport Canada Re-
2 3 4 search and Development Centre through the Canadian
CALIFORNIA BEARING RATIO-CBR Institute of Guided Ground Transport and partly by
the Natural Science and Engineering Research Council
APPROXIMATE INTERRELATIONSHIPS of Canada.
AT MODIFIED MAXIMUM DENSITY
FIGURE 16 Safe hearing stresses developed for REFERENCES
highway and airport design (9).
1. Stresses in Railroad Track--The Talbot Reports.
American Railway Engineering Association, Wash-
are conservative limits for use in track support ington, D.C., 1980 (reprint of Report of the
subgrade design and could probably be increased 50 Special Committee on Stress in Railroad Track,
percent. In addition, where climatic conditions re- 1918-1940, A.N. Talbot, chairman).
sult in freezing temperatures, minimum granular 2. Manual for Railway Engineering, American Railway
cover equal to at least half the depth of frost pen- Engineering Association, Washington, D.C., Vols
etration is generally specified. 1 and 2, 1984.
3. A,E.H, Love. The Stress Produced in a Semi-
BALLAST STRESSES Infinite Body by Pressure on Part of the Bound-
ary. Philosophical Transactions of the Royal
It is quite obvious from the data presented that as Society, Series A, London, England, Vol. 228,
the track modulus increases, so do the stresses in 1928, pp. 377-420.
the ballast, particularly those close to the cross- 4. L.T. Kenney. Steady State Vibrations of Beam on
tie bearing area. Because a higher modulus is asso- Elastic Foundation for Moving Load. Journal of
ciated with concrete crossties than with wood cross- Applied Mechanics, Vol, 76, 1954, pp. 359-364.
ties, it is not surprising that higher quality 5. Effect of Flat Wheels on Track and Equipment:
ballast is generally specified when concrete cross- Abstract of Report of Mechanical and Engineering
ties are used. In addition, most concrete crossties Divisions of the Association of American Rail-
are made with silica sands composed of minerals hav- roads. Bulletin 53, American Railway Engineering
ing a Mohs hardness greater than 6. To prevent Association, 1952, pp. 423-448.
mineral-to-mineral abrasion, or ballast powdering at 6. F,E. Dean, D.R. Ahlbeck, H.D. Harrison, and J.M.
the cross tie surface, the ballast must be composed Tuten. Effect of Tie Pad Stiffness on the Impact
of minerals of similar hardness. Thus ballast for Loading on Concrete Ties. Proc., 2nd Interna-
concrete crossties must be obtained from hard tional Heavy Haul Conference, Pueblo, Colo.,
mineral tough aggregate. This is best assured by 1982.
conducting "thin section analysis" on proposed ag- 7. D.R. Ahlbeck, H.D. Harrison, R.H. Prause, and
gregate sources. Thin section analysis also enables M.R. Johnson. Evaluation of Analytical and Ex-
the rejection of aggregates that have extensive perimental Methodologies for Characterization of
micro-fractures and former fractures bonded with Wheel/Rail Loads. Interim Report PB 272 063.
weak secondary minerals that weather quickly. u.s. Department of Transportation, 1976.
8. Manual of Standards and Recommended Practice,
Section G, Part II, Wheel and Axle Manual. Asso-
SUMMARY AND CONCLUSIONS ciation of American Railroads, Washington, D.C.,
1984.
A design methodology for calculating stresses below 9. Soil Primer Handbook. Portland Cement Associa-
conventional railway crosstie tracks has been pre- tion, Chicago, Ill., 1956.
sented. Rail seat loads were calculated from the
beam on elastic support. The rail seat loads were
assumed to cause a rectangular pressure distribution
on the ballast from which stresses were calculated Publication of this paper sponsored by Committee on
using isotropic elastic theory for the distribution Mechanics of Earth Masses and Layered Systems.