Wear 265 (2008) 1283–1291

Contents lists available at ScienceDirect

Wear
journal homepage: www.elsevier.com/locate/wear

The detrimental effects of hollow wear––field

experiences and numerical simulations
Robert Fröhling a,∗ , Anders Ekberg b , Elena Kabo b
a
Transnet Freight Rail, South Africa
b
Chalmers University of Technology, Sweden

a r t i c l e i n f o a b s t r a c t

Article history: The current paper deals with the influence of hollow wheels on the risk for surface initiated rolling contact
Accepted 30 January 2008 fatigue. To this end, the problem is introduced and field experiences are outlined. The problem is then
Available online 21 May 2008 analyzed more in detail through numerical simulations of bogie curving. Quasi-static simulations featuring
several hollowed profiles are carried out to give a first rough estimation of the severity. Simulations of
Keywords: the dynamic curving behaviour are carried out for a selected case to provide a more detailed image of the
Hollow wear
detrimental nature of wear hollowing. These are complemented by a FE-simulation for a detailed study of
Rolling contact fatigue
the contact configuration. Finally, the results are discussed and conclusions are drawn. In particular it is
Fatigue assessment
shown how high rolling contact fatigue impact is associated with high lateral forces and a narrow contact
patch. It is also seen to be likely that quasi-static simulations underestimate fatigue impact and that the
current practice of characterizing hollow wear by the wear depth is flawed.
© 2008 Robert Fröhling. Published by Elsevier B.V. All rights reserved.

1. Introduction vex contact conditions between the false flanges and the rail. The
high contact pressures in this region together with surface trac-
1.1. Background tion results in surface initiated rolling contact fatigue and plastic
deformations of the wheel and rail material.
As heavy haul railroads all over the world increase axle loads
and improve their bogie designs, there is a steady transition from 1.2. Effects of hollow wear
the wear to the stress regime [1]. The wear regime was typically
characterized by operational conditions of curved track and non- From heavy haul operations it is well known that hollow wheel
steering bogies, which resulted in unacceptably high wheel flange tread wear is detrimental with respect to surface initiated rolling
wear. The most effective remedy under such operational conditions contact fatigue.
was gauge face lubrication. However, as most heavy haul railroads Limiting and controlling the shape of hollow wear is thus impor-
increased the relative proportion of tangent track and introduced tant. From a track engineer perspective it is necessary in preventing
steering and other premium bogies, wheel tread wear started to rolling contact fatigue damage of rails. This is especially the case on
occur in a concentrated contact band. This resulted in hollow worn the gauge corner of the high rail and the field side of the low rail in
wheels with so-called “false flanges” next to the wheel flange and a curve, but also on stock rails in turnouts.
on the field side of the wheel tread. Actual examples of such profiles From a vehicle perspective, the high contact pressures and the
are given in Figs. 6–11. lateral loading when a hollow worn wheel negotiates the lower side
As hollow wear develops, the contact between the wheel tread of a curve tend to promote plastic flow towards the field side of the
and the railhead becomes more conformal. Hence, high conicities wheel. This will cause so-called lip formation. In addition, the plas-
are generated, which can result in a sudden high lateral input to tic flow at the surface of the wheel promotes plastic ratcheting and
the wheel or even in vehicle instability. Furthermore, the benefits the subsequent formation of surface initiated rolling contact fatigue
obtained by the nominal decrease in contact pressure between the cracks. The repeated contact close to the field side of the wheel
wheel tread and the rail head are exceeded by the detrimental con- where the resistance of the wheel material to plastic deformation
is reduced, due to reduced constraint, may also lead to the develop-
ment of subsurface cracking. This phenomenon is promoted by the
∗ Corresponding author. poor contact geometry due to the hollow wear as discussed above.
E-mail address: Robert.Frohling@Transnet.net (R. Fröhling). With continued crack propagation chunks of wheel material start to

0043-1648/$ – see front matter © 2008 Robert Fröhling. Published by Elsevier B.V. All rights reserved.
doi:10.1016/j.wear.2008.01.032
1284 R. Fröhling et al. / Wear 265 (2008) 1283–1291

fall out of the tread and rim. This poses a risk for secondary damages
and derailments.
From a management perspective, the question is how damaging
hollow wheels are [2]? This question also involves the maintenance
costs to remove hollow worn wheels as compared to the bene-
fits of preventing, or at least limiting, any negative impact on the
wheel/rail system [3,4]. To enable such a comparison, various tests
and simulations have been conducted in an attempt to quantify
damage caused by hollow wheel profiles and define wheel removal
criteria [5,6].
In operations, the negative impact of excessively worn hollow
wheels can be visible in terms of:

• An increased rolling resistance and thus fuel consumption to haul

wagons.
• Increased rail wear.
• Increased wheel flange wear. Fig. 1. Rolling contact fatigue cracks on the gauge corner of the high leg.
• “False flange” damage to the rail surface and special track work.
• Increased lateral forces in curved track.
[7]. The depth of hollow wear is here defined by the difference in
• An increased risk of hunting or other unsound running behaviour
tread wear at the taping line, which is 82.5 mm from the back-of-
on tangent track. flange, and the wear 15 mm from the field side of the wheel tread. It
should here be noted that substantial field side wear thus reduces
The “false flanges” that may be introduced if the hollow wear is the tendency for hollow wear of the wheel tread.
sufficiently large will, in addition to the detrimental effects men- Some of the most important consequences of hollow worn
tioned above, also cause poor steering of the wheelset. In the wheels on railway dynamic/rolling contact fatigue characteristics
extreme case the radius differential on a wheelset could become during curving are:
negative. Under these conditions the longitudinal creep forces will
steer the wheelset towards the high rail, resulting in accelerated • Negative radius differentials and effective conicities result in
wheel flange and rail gauge face wear [7]. steering forces towards the outside of a curve.
Remedial actions include the introduction of variable track • Rail surface fatigue on the gauge corner of the high rail (see Fig. 1)
gauge to minimise the hollow wear rate. Of more importance is and on the field side of the low rail (see Fig. 2).
that a systemic wheel–rail maintenance strategy is followed. This • Rolling contact fatigue damage in the swing nose and blade area of
often requires a compromise between the rolling stock operator turnout rails, caused by false flange contact of excessively hollow
and the infrastructure maintainer. worn wheels. Fig. 3 shows the rolling contact fatigue damage on
the stock rail next to the blade of a 1:20 turnout.
1.3. Outline of the paper • Subsurface initiated longitudinal cracks below the gauge corner,
which branch off as a transverse crack eventually covering a large
The current paper further investigates and analyses the detri- percentage of the crown width. These cracks can be the cause of
mental characteristics of hollow wear. In particular, it aims at a derailments due to fractured rails. See Fig. 4 for an example.
detailed investigation on the fatigue impact caused by hollowed • Increased rolling contact fatigue on the wheel tread, as exempli-
wheel profiles during curving. This knowledge is vital in defining fied in Fig. 5.
suitable measures to characterise hollow wear and in prescribing
quantifying limits for acceptable levels of hollow wear. Regarding running characteristics of a hollow wheel on a
Field experience of the occurrence and impact of hollow wear is straight track, it is difficult to draw specific conclusions from sim-
initially presented. The aim is to present the nature of hollow wear ulations. As mentioned above, hollow worn profiles may cause
and under which circumstances it is detrimental. Numerical simu- severe hunting. However, if the hollow wear is not symmetrical
lations of train–track interaction of wheels with different degrees of
hollowing are then carried out. In these simulations, the dynamic
behaviour of the running gear is studied and the risk of surface
and subsurface initiated fatigue quantified by use of shakedown
diagrams and the FIERCE model [8]. The aim is to relate opera-
tional knowledge of detrimental hollow wear to physical quantities.
In addition to the train–track interaction/fatigue impact analysis,
a detailed study is carried out on the contact pressure of a hol-
low wheel in contact with a rail. This is crucial in estimating how
the simplifications in the shakedown based rolling contact fatigue
models are affecting predictions, and also in understanding the
interaction between plastic deformation of the surface material and
hollow wear.

2. Field experience

Field experience has shown that wheels with more than 2 mm

hollow wear have a detrimental impact on the wheel/rail interface Fig. 2. Rolling contact fatigue damage on the field side of the low leg.
 R. Fröhling et al. / Wear 265 (2008) 1283–1291 1285

Fig. 6. Wheel with 1.8 mm hollow wear and double false flange (1.8d) as compared
to a nominal profile (WP21).

with respect to the centre of the track the wheel may move to
one side of the track and not show conventional hunting move-
ments. This in turn may promote increased flange wear and a rapid
deterioration of the wheel.

Fig. 3. Rolling contact fatigue damage on the stock rail and the blade.
3. Quasi-static numerical simulations of hollow wear
curving

To obtain a first estimation of the severity of hollow wear of

wheel profiles, quasi-static simulations were conducted using the
commercial code MEDYNA Version 8.2.1. Within MEDYNA Kalker’s
FASTSIM stripe theory was used to compute the wheel/rail contact
conditions. The simulations featured a wagon negotiating a curve of
500 m radius without track irregularities. The wagon characteristics
were chosen as representative of a South African coal export wagon
equipped with self-steering bogies as developed by H Scheffel in the
early 1970’s [9]. Rail profiles used are representative worn high and
low rail profiles as measured in a 500 m curve on the coal export
line.
Main parameters in the simulations were:

• Axle load: 26 ton.

• Curve radius: 500 m.
• Cant of the curve: 20 mm.
• Vehicle speed: 50 km/h.
• Nominal wheel radius: 458 mm.
• Track gauge: 1065 mm + 10 mm (wide track).
• Wheelset back-to-back distance: 987 mm.
• Wheel taping line distance: 82.5 mm.
• Distance between rail centres: 1152 mm.
• Maximum coefficient of friction between wheel and rail: 0.4.

Fig. 4. Fractured rail with final crack initiated from a longitudinal crack below the In operational service, all wheel profile shapes have their own
gauge corner. peculiarities. After considering many measured wheel profiles, the
following cases were selected to be included in the simulations:

• The nominal wheel profile, in the following denoted WP21.

• Four wheel profiles (1.8d, 2.3d, 2.5d and 2.7d) with double false
flange.
• Two wheel profiles (1.8 s and 2.4 s) with a single false flange on
the field side of the wheel.

In the denotation, the number indicates the depth of the hollow

wear. For example 2.5 indicates 2.5 mm hollow wear. Comparisons
between the nominal and the various worn wheel profiles are
shown in Figs. 6–11. In these comparisons, the profiles have been
placed so that the tips of the flanges coincide. The tip of the flange
is the only part of the wheel tread that does not wear.

3.1. Forces and contact geometries

The coordinate system employed in the following orientates is

shown in Fig. 12. The contact semi-axes a and b are related to the x
Fig. 5. Rolling contact fatigue cracks initiated close to the field side of the wheel
(so-called “spread rim”). and y directions, respectively.
1286 R. Fröhling et al. / Wear 265 (2008) 1283–1291

Table 1
Results from quasi-static simulations of curve negotiations

WP21 1.8d 2.3d 2.5d 2.7d 1.8 s 2.4 s 2.4 s (dyn)

uy [mm] −13.85 −9.52 −12.02 −12.78 −14.37 −17.09 −17.44 −21.67

␣ [mrad] −0.37 −0.72 −0.75 −0.62 −0.52 −1.04 −1.20 −3.48

Wheel on low rail

Fz [N] 115519 115527 111105 109519 108035 109159 107270 117098
Fx [N] 6062 6457 2921 1053 531 −131 −1913 3843
Fy [N] −939 −5295 −9176 −8173 −7758 −9060 −10598 −25595
a [mm] 7.86 7.02 8.29 7.72 8.09 9.10 9.27 9.10
b [mm] 5.18 7.67 5.54 5.78 4.77 4.04 3.54 2.98
yw [mm] 9.26 12.50 23.67 27.27 32.17 25.34 26.16 27.9
yr [mm] −2.01 5.68 14.57 17.49 20.88 11.28 11.99 10.85
R [mm] 457.69 457.82 458.71 459.02 459.46 458.68 459.22 459.33

Wheel on high rail

Fz [N] 139590 139270 142911 137150 144739 127237 128712 74438
Fx [N] 6149 6536 3116 −1502 809 −9261 −10523 −17424
Fy [N] 9122 11648 16762 18569 25976 15848 15687 18826
a [mm] 7.47 8.07 8.52 8.88 9.44 9.24 7.56 6.25
b [mm] 8.94 6.14 6.82 7.64 8.59 5.81 7.53 6.45
yw [mm] −27.38 −25.00 −30.71 −31.92 −36.45 −29.83 −33.54 −32.7
yr [mm] −10.11 −12.18 −15.61 −16.14 −19.16 −9.77 −13.37 −12.36
R [mm] 459.42 459.60 460.08 460.07 460.59 459.26 459.55 459.44

In Table 1 results of quasi-static simulations for the different yw contact position on the wheel
wheel profiles are shown. The presented variables are: yr contact position on rail
R rolling radius
uy lateral wheelset displacement
˛ angle of attack The simulation 2.4 s (dyn) represents the results from a full sim-
Fz normal force ulation of dynamic interaction with the 2.4 s profile using Medyna.
Fx longitudinal creep force The tabulated results represent the values at the highest dynamic
Fy lateral creep force contact pressure between the wheel and the rail.
a contact semi-axes in the longitudinal rail direction
b contact semi-axes in the transverse rail direction
3.2. Rolling contact fatigue assessment

To assess the severity of the different wheel profiles with respect

to the risk of surface initiated rolling contact fatigue, the out-
put from the quasi-static simulations are plotted in a shakedown
diagram, [10], in Figs. 13 and 14. In this context, the material
yield limit in shear was taken as k = 300 MPa. All work points
(denoted by an “x”) that are to the top-right of the curved solid line
denoted “Surface fatigue” in the shakedown diagram are deemed as
Fig. 7. Wheel with 2.3 mm hollow wear and double false flange (2.3d) as compared
prone to surface plasticity and thus surface initiated rolling contact
to the nominal profile (WP21). fatigue.

Fig. 8. Wheel with 2.5 mm hollow wear, with double false flange (2.5d) as compared Fig. 10. Wheel with 1.8 mm hollow wear and single false flange (1.8 s) as compared
to a nominal profile. to a nominal profile.

Fig. 9. Wheel with 2.7 mm hollow wear and double false flange (2.7d) as compared Fig. 11. Wheel with 2.4 mm hollow wear and single false flange (2.4 s) as compared
to a nominal profile. to a nominal profile.
 R. Fröhling et al. / Wear 265 (2008) 1283–1291 1287

Fig. 14. Shakedown map for the low rail.

4. Simulations of dynamic train–track interaction in

curving

4.1. Simulations of train–track dynamics

As mentioned above, the simulations of dynamic train–track

interaction were carried out using Medyna for the selected case
of a profile with 2.4 mm hollow wear and a single false flange.
Fig. 12. Coordinate system of wheel and low rail. The simulations comprised a curve negotiation at 50 km/h. The
track layout consisted of 20 m of straight track followed by a 60 m
long transition curve leading into a 500 m radius curve with a cant
When bearing in mind that the closer the work points are to of 20 mm. Typical track geometry irregularities were included in
the “Surface fatigue limit”, the higher the rolling contact fatigue the simulations. Due to unstable initial conditions, the first 55.5 m
impact is, it is obvious from Fig. 14 that the current practice of relat- (corresponding to the first 80 time increments) were neglected in
ing the severity of hollow wear to the wear depth can be seriously the following evaluation of the results.
questioned.
As seen in the shakedown maps, the 2.4 s profile on the low rail 4.2. Dynamic contact forces
showed the highest disposition to surface initiated rolling contact
fatigue. This case was therefore selected for a more thorough study. In Fig. 15 the force history during the negotiation of a curve
(R = 500 m) is presented. It is seen that, in particular, Fz and Fy com-

Fig. 13. Shakedown map for the high rail. Fig. 15. Force history at the low rail during curve negotiation.
1288 R. Fröhling et al. / Wear 265 (2008) 1283–1291

Fig. 17. Evolution of FIsurf and contact semi-axes a and b on the low rail during curve
Fig. 16. History of normal forces and contact point location at the low rail during negotiation.
curve negotiation.

that a high fatigue impact is mainly correlated to the Fy component,

ponents show an oscillatory trend. A comparison of average force i.e. the frictional force in the wheel axle direction.
magnitudes to the quasi-static magnitudes of Table 1 (Fz = 107 kN, In the same manner a study of the influence of the contact semi-
Fx = −1.9 kN and Fy = −10.6 kN) shows a reasonable agreement. axes a and b as shown in Fig. 22 reveals that it is mainly the semi-
In Fig. 16 the lateral displacement of the contact point on wheel axis in the wheel axle direction (b) that correlates to a high fatigue
and rail is shown. Coordinates of the contact point are according to impact.
the coordinate system presented in Fig. 12. It is seen that the wheel It can thus be concluded that, for the case studied, a high rolling
is relatively stable in the lateral position during curve negotiation. contact fatigue impact with respect to surface initiated cracks is
The disruptions after 55–65 m correspond to a disturbance in the related to a high lateral force combined with a narrow contact patch
vertical force. The cause is a geometrical track irregularity. due to the poor contact geometry between the hollow worn wheel
profile and the rail.
4.3. Fatigue assessment
5. Finite element simulations
To quantify the fatigue impact with respect to surface initiated
cracks, a previously derived fatigue index was employed [8]. This To study the contact stress distribution between wheel and rail
fatigue index may be expressed as in detail, a finite element simulation was carried out. In particular,
2abk shakedown analysis rests on the presumption of Hertzian con-
FIsurf =  − (1) tact (i.e. an elliptic contact patch), full slip and elastic conditions.
3Fz
The simulation was intended to show whether these presumptions
where were reasonably fulfilled.

Fx2 + Fy2
= (2)
|Fz |
A magnitude of FIsurf > 0 indicates that surface plasticity and
consequently surface initiated rolling contact fatigue is likely.
In Fig. 17, the evolution of FIsurf is plotted with the evolution of
the contact semi-axes a and b. It is seen that the oscillatory steady-
state behaviour is also reflected in the size of the contact patch,
mainly in the b semi-axes. Further, it is seen that the threshold
FIsurf > 0 is frequently exceeded. These high magnitudes are con-
firmed in the corresponding shakedown map in Fig. 18.
As seen from Eq. (1), a high fatigue impact can stem from either
a high frictional force (), or a high normal load applied to a small
contact patch size (ab/Fz ). In Fig. 19 combinations of normal and
frictional load for the different instances in time are plotted and
shaded so that the higher the fatigue impact, the whiter the spot.
It is seen that occasions of high fatigue impact correlate strongly
with high frictional forces, and less with high normal forces.
A study on the impact of high friction versus a small contact
patch reveals a strong correlation of both of these parameters with
high fatigue impact. See Fig. 20.
If the magnitude of FIsurf is related to the two components of the
frictional load at different instances in time, as in Fig. 21, it is seen Fig. 18. Shakedown diagram for contact with the low rail during curve negotiation.
 R. Fröhling et al. / Wear 265 (2008) 1283–1291 1289

Fig. 19. Combinations of frictional and normal load on the low rail at different Fig. 21. Absolute magnitudes of frictional load components on the low rail at dif-
instances in time during curve negotiation. A whiter spot indicates a higher fatigue ferent instances in time during curve negotiation. A whiter spot indicates a higher
impact. fatigue impact. Note the different scales on the axes.

5.1. Numerical model 5.2. Results

The FE-model featured an elastic wheel and rail material (Elas- For a case of pure normal loading, the contact patch was found to
ticity modulus 209 GPa, Poisson’s ratio 0.29). The mesh, composed be elliptic in accordance to the Hertzian presumptions. The size of
of 20-node quadratic brick elements, was refined towards the con- the contact patch was evaluated as 9 mm × 21 mm, which in reason-
tact patch. The model consisted of 10 248 elements in the wheel able accordance to the size predicted by the Medyna simulations.
part, 16 500 elements in the rail part and a total of 380 457 degrees
of freedom (including Lagrangian multiplier variables). A coeffi-
cient of friction of 0.5 in the wheel–rail contact was prescribed.
The commercial package ABAQUS Version 6.5.1 was employed for
the simulations.
Boundary conditions were applied according to Fig. 23. It can be
noted that these boundary conditions are too rough to give a good
global distribution of the stresses, but were deemed acceptable for
the current aim of examining the contact stress distribution.
Geometry, load magnitudes and contact positions were chosen
as the load case 2.4 s at the lower rail according to Table 1.

Fig. 22. Contact patch semi-axes (a and b) on the low rail at different instances in
time during curve negotiation.

Fig. 20. Combinations of frictional load and patch size on the low rail at different
instances in time during curve negotiation. A whiter spot indicates a higher fatigue
impact. Fig. 23. Meshed FE-model of the wheel–rail interface.
1290 R. Fröhling et al. / Wear 265 (2008) 1283–1291

of 800 MPa. It is likely that the plastic zone size is smaller in real-
ity due to the component hardening imposed by residual stresses
at the contacting surfaces, which are not included in the present
simulations. Still, there will be significant plastic flow, which will
promote the plastic deformation and wear of the wheel surface.
The plastic flow and wear will promote an increased hollowing and
may also cause a plastic “lip” to be formed at the field side of the
wheel, see Fig. 25.
In conclusion, the finite element simulations show that several
assumptions implied in the shakedown analysis were reasonable.
In particular contact patch size and contact stress distributions
showed reasonable agreement to the presumption of Hertzian con-
tact. The main disagreement is that stress magnitudes exceed the
material yield limit, which is in contrast to the shakedown analysis
presumptions. This is of less interest in a comparative assessment
of the tendency to rolling contact fatigue initiation between dif-
ferent wheel profiles and in a qualitative assessment of risk fatigue
initiation. However, if the aim is to predict fatigue lives and the evo-
lution of hollow wear, more sophisticated material models that can
account for plastic deformation and model plastic ratcheting are
needed, cf [11] In addition, the presumption of full slip is doubtful.
Also in this case, a more detailed prediction of fatigue life would
call for a better modelling of interfacial friction.

6. Concluding remarks

Causes for, and consequences of, hollow worn wheels have been
discussed. From field experience it is pointed out that hollow worn
wheels will inflict rolling contact fatigue damage and increased
wear on wheels, rails and switches. Thus, the adoption of wheel
profile management that accounts both for the detrimental effect
of hollow worn wheels and the costs of maintaining low magni-
tudes of hollow wear must consider both damages on wheels, rails
and switches. Further, to establish appropriate limits, it is vital to
be able to quantify the severity of different amounts and types of
Fig. 24. Von Mises stresses at wheel–rail interface evaluated using elastic FE- hollowing.
simulations. To this end, numerical simulations were carried out to quantify
the fatigue impact with respect to surface initiated rolling contact
The application of frictional loading distorts the contact patch. This fatigue.
is seen in Fig. 24 where von Mises stresses at the contacting sur- A comparison between quasi-static simulations and full sim-
face of the wheel (upper) and rail (lower), respectively as evaluated ulations of the dynamic train–track interaction during curve
from the FE-analysis are presented. It is seen that a large portion negotiation revealed that the former underestimates the fatigue
of the contacting surface experiences stress magnitudes in excess impact (as quantified by shakedown theory). Furthermore, the
results showed the probability that the depth of hollow wear is not
a conclusive measure of the severity with respect to rolling contact
fatigue.
From numerical simulations of dynamic train–track interaction,
it was concluded that a high fatigue impact was mainly related to
a high lateral force combined with a narrow (in the wheel axle
direction) contact patch. The high lateral force is due to the poor
steering caused by the hollow worn profile. The narrow contact
patch is due to the poor contact geometry that is caused by the
mismatch between the hollow worn wheel profile and the rail head
profile.
Finally, finite element simulations showed that several pre-
sumptions implied by shakedown analysis were reasonable. The
main incompatibility is that evaluated stress magnitudes exceed
the material yield limit. The consequences of this were discussed.
From a practical point of view, it is hoped that the current paper
will contribute to a deeper understanding of the influence of hollow
wheel wear. In particular the paper points out the need for further
studies to establish proper hollow wear limits since the current
practice of employing the wear depth can be questioned. In addi-
Fig. 25. Plastic “lip” on the field side of a wheel profile. tion to a deeper theoretical knowledge, there are also the practical
 R. Fröhling et al. / Wear 265 (2008) 1283–1291 1291

issues on how to measure and quantify hollow wear that needs to be the International Heavy Haul Association Conference on Wheel/Rail Interface,
resolved if a more refined criterion is to be put into practice. For this Moscow, 1999, pp. 299–305.
[4] K. Sawley, S.L. Clark, The Economics of Removing Hollow Wheels from Service,
reason we believe that a 2 mm hollow wear limit is for the moment TTCI Technology Digest 99-034, December 1999.
a good guideline for effective wheel/rail interface management. [5] K.J. Sawley, D. Oliva-Maal, J. LoPresti, Effect of hollow worn railroad wheels on
fuel use and track damage, RTD-Vol. 15, Rail Transportation, ASME 1998, pp.
71–76.
Acknowledgement [6] K. Sawley, D. Oliva, J. LoPresti, Effect of Hollow-Worn Wheels on Wheel/Rail
Rolling Resistance, TTCI Technology Digest 98-004, February 1998.
[7] R. Fröhling, Analysis of asymmetric wheel profile wear and its consequences,
The work at Chalmers is part of the on-going activities within in: Proceedings of the 19th IAVSD Symposium, Politecnico Milano, August
the National Centre of Excellence in Railway Mechanics CHARMEC 29–September 2, 2005, Supplement to the International Journal of Vehicle
(www.charmec.chalmers.se). System Dynamics, Volume 44, 2006, pp. 590–600.
[8] A. Ekberg, E. Kabo, H. Andersson, An engineering model for prediction of rolling
contact fatigue of railway wheels, Fatigue Fract. Eng. Mater. Struct. 25 (2002)
References 899–909.
[9] H. Scheffel, The hunting stability and curving ability of railway vehicles, Rail
[1] H.M. Tournay, J.M. Mulder, The transition from the wear to the stress regime, International (February) (1974) 154–176.
Wear 191 (1996) 107–112. [10] K.L. Johnson, The strength of surfaces in rolling contact, in: Proceedings of the
[2] K. Sawley, S. Clark, J. LoPresti, How damaging are hollow-worn wheels? Railway Institution of Mechanical Engineers, vol 203, 1989, pp. 151–163.
Age (February) (1999) 41–45. [11] M. Ekh, A. Johansson, H. Thorberntsson, L. Josefson, Models for cyclic ratchetting
[3] K.J. Sawley, S.L. Clark, Engineering and economic implications of hollow worn plasticity––integration and calibration, ASME J. Eng. Mater. Technol. 122 (2000)
wheels on wheel and rail asset life and fuel consumption, in: Proceedings of 49–55.