See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/301695304

Laser Cladding of Railway Track to Promote

Wear/RCF, Bend Fatigue and IBJ Lipping
Resistance

Conference Paper · April 2016

DOI: 10.4203/ccp.110.149

CITATIONS READS

0 120

6 authors, including:

Stephen Lewis Paul Goodwin

The University of Sheffield The Welding Institute
36 PUBLICATIONS 113 CITATIONS 27 PUBLICATIONS 252 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Full-Scale Testting and Analysis of Laser Clad Railway Track View project

RSSB Sanding Project T797 View project

All content following this page was uploaded by Stephen Lewis on 06 July 2016.

The user has requested enhancement of the downloaded file. All in-text references underlined in blue are added to the original document
and are linked to publications on ResearchGate, letting you access and read them immediately.
Laser Cladding of Track to Promote Wear/RCF, Bend Fatigue and
IBJ Lipping Resistance

S. R. Lewis1, R. Lewis*1, P. S. Goodwin2, S. Fretwell-Smith3, D. I. Fletcher1, K.

Murray4
(1) Department of Mechanical Engineering, University of Sheffield, UK. (2) Laser Cladding Technology Ltd,
Worksop, UK.
(3) Tata Steel Rail Technologies, Rotherham, S60 3AR. (4) Sandvik Osprey Ltd, Neath, UK.

Corresponding author: Roger Lewis*, Professor, Department of Mechanical

Engineering, University of Sheffield, Sheffield, South Yorkshire, S1 3JD, UK, +44
(0)1142 227838, [email protected]

Abstract
This paper reports on three series of tests which were carried out on full-scale
sections of rail that had been treated by laser cladding aiming to create a layer (~1-
2mm) of high performance material on the rail head. In the first series of tests wear
and rolling contact fatigue (RCF) experiments were carried out on one metre lengths
of clad rail track using The University of Sheffield’s full-scale wheel-rail test facility
(FSTF). In a second series of tests on the FSTF, one metre long sections of rail were
tested including an insulating “end-post” inserted into the head of the rail halfway
along the length. This configuration simulated an insulated block joint (IBJ) loaded
under dry conditions and with plastic deformation at the insulating end-post
measured periodically. The third test series used a four-point bend fatigue
configuration to load clad rail specimens to investigate the integrity of the
clad/substrate bond.

Keywords: Full-Scale, Testing, Laser Cladding, Wear, Rolling Contact Fatigue,

Insulated Block Joint, Lipping, Bending Fatigue

1 Introduction
Replacement of worn rail, extension of rail life and prevention of rail failure
particularly from rolling contact fatigue (RCF) are significant cost contributors to
the running of a rail network. The same is true for high value track components such
as insulated block joints (IBJ’s) or switches and crossings. Network downtime
caused by replacement and maintenance is a significant disruption to travellers and
freight traffic and can have a follow on impact on the wider economy. The treatment
of rails and track components by laser cladding to improve rail durability has been
studied before [1 – 4] and shows good potential as a method for reducing wear and
increasing the RCF life of rails. Tests which simulated an IBJ [5] have shown that
laser cladding of the rail head around the area of the endpost can reduce the

1
tendency for the rail to lip across the endpost. These tests, however, were all carried
out using small scale twin-disc apparatus. Therefore, to validate the laser cladding
approach it was necessary to move to full-scale tests. In this study three series of
tests were carried out on full-scale sections of rail which had been treated by laser
cladding aiming to create a layer (~1-2mm) of high performance material on the rail
head. In the first series of tests wear and RCF experiments were carried out on one
metre lengths of clad rail track using The University of Sheffield’s full-scale wheel-
rail test rig (FSTF). The FSTF is capable of testing under controlled load, speed and
slip, and a test method as used in [6] was adopted for these tests with the intention of
generating RCF cracks. In the second full-scale test series, one metre long sections
of rail were tested including an insulating “end-post” inserted into the head of the
rail halfway along the length. This configuration simulated an insulated block joint
(IBJ) loaded under dry conditions and with plastic deformation at the end-post gap
measured periodically. In the field IBJs are used to electrically isolate two adjoining
sections of rail for signalling purposes. The primary failure mechanism of standard
un-clad IBJ’s in the field is by a mechanism referred to as lipping [7]. Lipping
occurs during repeated rolling over the IBJ by wheels and is a result of plastic flow
of the rail material across the insulating polymer endpost. Excessive lipping causes
failure of the signalling system as the two sections of rail become electrically
connected. The third test series used a four-point bend fatigue configuration to load
clad rail specimens to investigate the integrity of the interfacial clad/substrate bond
by subjecting the rail to the bending stresses that it would see in the field.

2 Specimens
R260 rail grade was used as a baseline (unclad) and as a substrate for the clad
samples in these tests. Two cladding materials were applied: (i) Stellite 6, and (ii) a
grade of martensitic stainless steel, hereon referred to as MSS. These cladding
materials were selected from a larger list of potential candidate materials based on
their favourable wear and RCF properties relative to standard 260 grade rail material
during twin-disc tests performed previously [2]. For test series 1 both types of clad
specimen were prepared for FSTF testing starting from a standard R260 grade one
metre long section with 60-E2 profile. The crown radius and the gauge corners of
the rails were machined away to a depth of 2mm. The profiles of the rails were then
restored by cladding the deposit materials in two layers which were then ground to
restore the original profile and achieve a good surface finish leaving a 2mm thick
layer of clad at the surface as shown in Figure 1. Even though only part of the rail
head comes into contact with the wheel it was easier in these early trials to clad the
full profile of the rail. Note that if cladding were to be developed commercially it
would be expected to roll a rail profile specifically to be clad. This machining
process is only needed for the small batch specimen production used here.

2
Figure 1: A section image of a R260 60-E2 rail clad to a depth of 2mm used in series
1 tests

The bend fatigue samples were only 500mm in length, but otherwise used the same
cladding method. The rail web and foot were removed for these tests so the work
could focus just on the surface and cladding performance. The prohibitively high
cost of grinding the clad rails meant that the bend fatigue specimens were not
ground to the original 60-E2 profile. As the bend specimens did not need to come
into contact with a wheel the specimens were instead machined to an approximate
profile as shown in Figure 2.

Figure 2: Image of a bend fatigue specimen

3
For test series 2 the IBJ specimens consisted of a clad one metre long rail with a 6.4
mm wide groove machined into the railhead approximately halfway along its length.
A nylon endpost was glued into the groove which had the same profile as the rail
and sat flush with the rail crown as shown in Figure 3. This solution did not fully
represent an in-track IBJ, but was chosen as it did not require a fully bolted rail joint
which would not have fitted into the FSTF without heavy modifications to the
machine. For this test the rail surface contact and plastic flow was of interest, and as
the rail could not bend within the FSTF the lack of a full rail joint was not
significant.

Figure 3:Image of IBJ sample

3 Apparatus
3.1 Wear, RCF and IBJ Tests
For the Wear/RCF and IBJ tests (series 1 and 2), clad one metre long rail sections
were tested in The University of Sheffield’s Full-Scale Test Facility (FSTF). A
schematic of the FSTF is shown in Figure 4. It consists of a locomotive wheel with
P8 profile which sits within a pivoted loading frame. The wheel rests on top of the
rail specimen under test. Figure 4 shows the operating motion of the FSTF. At the
start position the specified vertical load (1) is applied representing a half axle load.
The rail is then pulled in the direction of travel with the wheel free to rotate as the
rail passes underneath it (2). A third actuator connected to the wheel rim via a chain
causes creep by pulling the wheel and increasing its surface velocity above that of
the rail (3). One cycle is defined as the machine traversing from the start position to
the end position and back to the start again.

4
 Figure 4: A working schematic of the University of Sheffield Full-Scale Rig.

3.2 Four Point Bend Fatigue Tests

The series 3 four point bend fatigue tests were performed using a Schenck single
axis top loading servo hydraulic machine of 250kN capacity. A two point loading
frame was mounted to the top crosshead of the machine. The bottom two supports
were attached to a welded steel frame which sat on the bed of the machine. The top
two pivots were 25 mm diameter and the bottom pair were of 60mm diameter. The
web and foot of the rails were removed to reduce the load required to achieve a
specific maximum stress at the rail head. The samples were 500mm in length. The
bottom pivots were set at a distance of 350mm apart equidistant of the force centre
giving a sample overhang of 75 mm. The top two pivots were 100mm either side of
the force centre as shown in Figure 5.

Figure 5: Schematic of the four point bend fatigue test set-up

4 Methodology
4.1 Wear/RCF Tests (Series 1)
To limit the number of test samples and test time required it was decided to test for
wear and RCF on a single sample rather than having separate samples for both tests.

5
Burstow et al. [6] developed a test method to promote crack growth using alternating
periods of dry and wet cycling with the FSTF. In that work it was found that small
crack like features were seen in the rail after testing under this regime, but it was
concluded that these cracks were not large enough to constitute full RCF cracks. In
the current work the intention was to run the tests until small, nucleated, cracks were
visable so Burstow’s test technique was deemed sufficient for these tests. This
technique also allowed the measurement of wear and RCF simultaneously. In the
wear/RCF tests the rail samples were tested in alternating series of 10,000 dry cycles
followed by 10,000 wet cycles as developed in [6]. A gravity-fed water system
delivered water at a rate of 0.5ml/s to a point just before the contact between the
wheel and rail. To avoid excessive test time a limit was set of a cumulative total of
80,000 cycles per rail sample. Wear was initially intended to be measured every
10,000 cycles by measuring the rail profile using a Greenwood Engineering Mini-
Prof rail profiler. However, it was discovered during the tests that the wear rate of
the rail was too low for this technique to capture any wear data. RCF was monitored
periodically using magnetic particle inspection, however, no cracks were seen using
this method during the entire test regime on any of the specimens. Wear was
eventually quantified by measuring the width of the running band on each rail and
making a comparison of how much the band tapered along the length of the rail in
the direction of rolling as shown in Figure 6. Tests were performed using a vertical
load of 110 kN which gave a contact pressure of approximately 1500 MPa when the
wheel and rail profiles were new and unworn. This was measured using ultrasonic
measurement of the contact patch [8]. The rail travel distance was 300 mm and a
wheel slip of 1% was used.

Figure 6: Running band of MSS sample after 50,000 test cycles

6
4.2 IBJ Lipping Tests (Series 2)
Four IBJ samples were tested including 2 reference un-clad R260 samples and 2
clad R260 samples clad with Stellite 6 and MSS respectively. Lipping of the endpost
gap was monitored by measuring the gap every 5,000 cycles. Tests were run in
periods of 5,000 cycles until the gap reduction had stopped or stabilised. The IBJ
tests were run at 110 kN vertical load, a relatively high slip of 3% and a rail travel of
300 mm. The high slip level was chosen to ensure that lipping of the endpost was
seen within a reasonable time frame. The rail traverse was set so that the wheel fully
rolled over the endpost gap. A second IBJ reference specimen was run with a
reduced slip level of 1% to observe the effect that slip magnitude was having on the
lipping rate. All IBJ tests were performed in dry conditions.

4.3 Four Point Bend Fatigue Tests (Series 3)

When a wheel rolls over a section of track which is supported either side of the
wheel contact by two equally spaced sleepers the rail head will be subject to a
maximum compressive bending stress directly under the wheel. However, either side
of the wheel the rail head will be subject to a maximum tensile stress. The tensile
region ahead of the wheel will move with the rolling wheel much like “bow wave”
[9]. The test samples were subject to a higher stress range of 350 MPa with the
railhead in tension. If the samples did not fail at this stress range then they could be
safely assumed to withstand conditions in the field. In the case of a sample without
fatigue failure a run out limit was set to 670,000 cycles. If a sample failed before this
limit then a new sample of the same material was tested at a reduced stress range
until run-out occurred. Table 1 lists the samples tested. As the MSS sample ran out
at the highest stress range a new MSS sample was tested again at the higher stress
range to validate the MSS result.

Test No Clad Stress Range, MPa

MSS-350 MSS 350
MSS-350B MSS 350
ST6-350 Stellite 6 350
ST6-275 Stellite 6 275
ST6-200 Stellite 6 200
Table 1: Bend fatigue tests performed

It has been shown in [10] that the cladding process can generate inclusions in the
clad layer and at the clad/parent material interface. These inclusions can potentially
be sites for fatigue crack initiation. The large heat input during the cladding process
changes the microstructure of the parent material directly below the clad/parent
interface. This area known as the heat affected zone, HAZ, is also another potential
area of weakness in clad samples. Equation 1 [11] was used to calculate the load
required to subject the sample to the required stress range. A stress ratio of R = 0.1
was used i.e. the sample was always under load, to avoid it moving during the test.

7
 𝑅𝑒𝑞𝑢𝑖𝑟𝑒𝑑 𝑆𝑡𝑟𝑒𝑠𝑠.4.2𝑛𝑑 𝑀𝑜𝑚𝑒𝑛𝑡 𝑜𝑓 𝐴𝑟𝑒𝑎
𝐿𝑜𝑎𝑑 = (1)
𝐸𝑓𝑓𝑒𝑐𝑡𝑖𝑣𝑒 𝑆𝑝𝑎𝑛.𝑦

where the 2nd Moment of Area, M, is a property of the sample and is calculated using
Equation 2. The Effective Span is equal to the bottom span minus the upper span and
y is the distance from the samples neutral axis to the outer fibre (in tension).

𝑊𝑖𝑑𝑡ℎ.𝑇ℎ𝑖𝑐𝑘𝑛𝑒𝑠𝑠 3
𝑀= (2)
12

5 Results
5.1 Wear/RCF Tests (Series 1)
The amount of wear damage seen on all of the samples tested was too low to be
measured using a profilometer. Rail wear was therefore assessed by measuring the
width of the contact band. Typically the contact band would taper from start to
finish and it was the amount of taper which was used to gauge the amount of wear. It
should be noted that changes in the width of the contact band will be a combined
effect of both wear (material loss) and plastic flow leading to displacement of
material to either side of the contact band.

Magnetic particle inspection did not reveal any RCF cracks in any of the samples. It
may have been the case that the tests were not run for long enough for RCF cracks
to be generated. However, evidence of plastic deformation, shown in Figure 7, on
the R260 grade reference rail as also seen in [6] does show that these test conditions
were sufficient for ratchetting strain accumulation to take place which is a usual
precursor of RCF crack formation.

8
Figure 7: Evidence of plastic deformation on the R260 reference sample after 80,000
alternating dry/wet cycles

The reference R260 grade and the Stellite sample were each tested to 80,000 cycles.
At the time of writing the MSS sample had only completed 50,000. Table 2 shows
the measurements of the contact bands on the R260 and Stellite rails after 80,000
cycles.

Sample Width of Wear Band, mm Taper

Start (0 mm) End (300 mm) mm As % of 260-2
R260 13 22 9 -
Stellite 6 13 15 2 -78%
Table 2: Wear band taper results at 80,000 cycles

The MSS sample is omitted from Table 2 as it had only completed 50,000 cycles
and therefore a direct comparison with the R260 and Stellite samples is not possible.
It can be seen that the width of the contact band at the start is identical for both
specimens in Table 2. However, by the end of the contact travel the R260 sample
tapers to a greater extent than the Stellite (9mm compared to 2mm). An illustration
of how the taper is measured is shown in Figure 8.

9
Figure 8: Illustration of how the wear band taper is measured

5.2 IBJ Lipping Tests (Series 2)

In the IBJ lipping tests the endpost gap was measured every 5,000 cycles up to a
total of 40,000 cycles. A machine failure on the FSTF while testing the Stellite
sample between 30 – 35,000 cycles meant that only data up to 30,000 cycles could
be reported. Figure 9 shows the evolution of the endpost gap for each sample. Three
of the samples were tested at 3% slip and a second R260 reference sample was also
tested at 1% slip.

Figure 9: Evolution of the endpost gap with the number of test cycles (R260 3% test
stopped early due to large scale deformation)

Figure 10 and Figure 11 show plan view images of the rail head of the R260 and
Stellite 6 IBJ specimens tested at 3% slip. The untested IBJ can be seen in each
Figure as being straight and undeformed. After 5,000 cycles for both samples it can
be seen how rail material flows across the endpost gap and a lip is formed. The

10
Nylon endpost appears to get crushed by this flowing material. Comparison of
Figure 10 and Figure 11 clearly shows that the clad rail gives much greater
resistance to lipping than the unclad rail.

Figure 10: Images of the R260 IBJ specimen before the test and after 5,000 cycles at
3% slip. The rolling direction of the wheel is from left to right

Figure 11: Images of the Stellite 6 IBJ specimen before the test and after 5,000
cycles and 25,000 cycles at 3% slip The rolling direction of the wheel is from left to
right. Reflection from the nylon endpost indicates transfer of metal

5.3 Four Point Bend Fatigue Tests (Series 3)

Table 3 shows the results of the 4-point bend fatigue tests. The MSS clad was the
first bend specimen to be tested at the maximum stress range of 350 MPa (MSS-350,
see Table 1). MSS-350 did not show any signs of failure and the test was stopped at
670,000 cycles defined as a run-out. The ST6-350 Stellite sample saw complete
fracture after only 61,000 cycles, less than 10% of the MSS-350 sample which was
tested under the same conditions. Sample ST6-275 tested at 275 MPa saw complete
fracture after 259,000 cycles. Lowering the stress range to 200 MPa on a new
Stellite 6 clad sample produced run-out to 670,000 cycles on sample ST6-200. To
check the seemingly better integrity of the MSS samples over that of the Stellite

11
ones a second MSS sample, MSS-350b, was run at a stress range of 350 MPa. The
sample also ran out to 670,000 cycles.

Sample Stress Range, Number of Cycles to

Clad
Designation MPa Failure
MSS-350 350 MSS 670,000
ST6-350 350 Stellite 6 61,000
ST6-275 275 Stellite 6 259,000
ST6-200 200 Stellite 6 670,000
Table 3. Results of 4-point bend fatigue testing on clad R260 grade rail samples

An unclad R260 reference sample will be tested at a stress range of 350 MPa to
gauge the effect of the cladding on the fatigue performance of the base rail.

Figure 12 shows the broken section of the ST6-350 sample which failed at 61,000
cycles.

Figure 12: Section view of ST6 – 350 sample after failure at 61,000 cycles

Figure 13 and Figure 14 show close up images of the broken ST6 – 350 sample and
show the same type of spherical inclusions which were also seen in the Stellite 6
Wear/RCF sample.

12
 Figure 13: Close up image of broken ST6 – 350 sample section after failure at 61,000
cycles

Figure 14: Additional area of close up image of broken ST6 – 350 sample section after
failure at 61,000 cycles

It can be seen from Figure 13 and Figure 14 that some of the cracks in the sample
appear to originate from the edges of inclusions, some appear to originate from the
clad interface and some pass from the clad into the substrate. However, sectioning a
sample prior to failure would be required to verify the earlier stays of crack growth.

13
 Figure 15: Close up image of broken ST6 – 275 sample section after failure at 259,000
cycles

Figure 15 shows a close up image of the failed section of the ST6 – 275 sample.
This sample also had inclusions in the clad layer, however, they were concentrated
towards the sample’s edges. Cracks were also seen to be growing from the edges of
the inclusions downward through the clad/substrate interface. Figure 15 shows
cracks growing through the interface between the deposit and the HAZ. It is not
clear whether these cracks originate in the clad layer or at the interface. An
important purpose of the bend fatigue tests was to test the integrity of the
clad/substrate bond and these results indicate it is of sufficient integrity. There are
no signs of delamination between the clad and substrate were seen in any of the
failed or run-out samples and failure is instead being controlled by the presence of
inclusions in the deposit.

6 Discussion
6.1 Wear/RCF Tests (Series 1)
The MSS sample had only completed 50,000 cycles at the time of writing and hence
a direct comparison with the R260 and Stellite 6 samples is not possible which both
completed 80,000 cycles. The amount of material loss from the rails was too low for
wear to be measured by measuring the rail profiles. It can be seen that the width of
the contact band at the start is identical for both R260 and Stellite 6 specimens
(Table 2). As the wheel starts from rest at the beginning of each cycle the slip is
zero. The slip increases to the desired level as the wheel starts to roll along the rail.
Any other effects such as lateral and spin creep will similarly increase as the wheel
begins moving and reaches the desired slip. Therefore is it hypothesised that the
width of the wear band at its beginning is the same for all samples as the mechanical

14
work which is being put into the rail at this location is the same for all samples
tested i.e. pure rolling. (The contact band width at the start of the MSS sample was
12 mm and it is assumed that this will grow to 13mm as 80,000 cycles are
approached). It can be seen from Table 2 that by the end of the contact travel the
R260 sample has tapered to a greater extent than the Stellite 6 one (9mm compared
to 2mm). It is suggested that as the starting contact width is the same for both
samples the amount of taper can be directly compared as an indicator of wear of the
rail samples. This is because the amount of taper seen on a particular sample will be
a reflection of that particular materials ability to resist the shear forces generated
within the rolling sliding contact. As the wheel starts to roll along the rail and the
creep builds up to its maximum level the rail will wear at an increased rate. A
material which is more prone to plastic flow (a precursor to wear and RCF) will
experience more material loss or displacement leading to a larger contact area. This
larger area is visible in the increased contact width at the end of the rail travel. By
this hypothesis it can be said that the Stellite sample wears at a rate of 22% of that of
the reference sample. During small scale tests in [2] the Stellite 6 samples wore at an
average rate which was 51% of the reference. The MSS sample showed no tapering
after 50,000 cycles indicating that it is wearing at a lower rate than the Stellite
sample. This difference in relative wear is to be expected due to the different contact
conditions between twin-disc and full-scale. For example side flow displacement of
material is possible in a full size 3D contact but not in the 2D line contact of a twin-
disc test. It also needs to be noted that in these full-scale tests alternating dry and
wet cycles were used where as the twin-disc tests in [2] were fully dry. Burstow [6]
performed fully dry and alternating wet/dry tests and states that “the tests run with
intermittent lubrication exhibited a wear rate of approximately half that of the dry
material”. Although a direct comparison is difficult between the two clad samples
the full-scale results do reflect those of the small-scale [2] where the MSS was seen
to wear at a lower rate than the Stellite. In [2] the MSS samples were seen to wear at
an average rate of 19% of the reference and the Stellite 6 samples wore at a rate of
51% of the reference.

These tests were intended to test the specimens for their RCF resistance. MPI of all
the samples showed no evidence of RCF cracks. Non-destructive and destructive
testing of the samples will be carried out at a later date to investigate the sub-surface
state of the samples. This stage of investigation may reveal early signs of RCF in the
R260 sample. Evidence from small-scale testing in [2] however, shows that the
cladding is highly resistant to sub-surface deformation and it is expected that the
clad specimens will show little to no signs of ratcheting. It may be that the test
regime of alternating wet and dry cycles is not sufficient to lead to the early
development of RCF and that the test regime needs to be adapted accordingly.

6.2 IBJ Lipping Tests (Series 2)

It can be seen from Figure 9 that the gap on the R260 sample tested at 3% slip more
than halves in 5,000 cycles reducing to 3.4mm. For the same number of cycles the
MSS sample only sees a reduction of 0.8mm and the Stellite only 0.7mm. A second

15
R260 sample tested at a reduced slip of only 1% showed a reduction of 0.6mm. This
is quite a significant result as it shows that an endpost between two clad rails can
resist almost 3 times the energy input into the contact (i.e. 3% slip compared to 1%
slip) as a standard unclad IBJ. This is reinforced by the fact that the R260 sample
tested at 1% tends to follow the same trend as the two clad specimens tested at 3%
slip throughout the whole 40,000 cycle duration as seen in Figure 9.

Beyond 5,000 cycles of testing the endpost gaps in the Stellite and MSS samples
continue to reduce until they reach approximately 5.3 and 5.1mm at 10,000 and
15,000 cycles respectively.

It can be noted from Figure 9 that the R260-1% and the Stellite-3% samples exhibit
fluctuating behaviour where the gap reduces, increases and is then reduced again.
This phenomenon is caused when metal flows across the endpost gap initially
reducing it. Then a part of that deformed metal breaks away increasing the gap
again. With more test cycles the material flows across the gap reducing it once more.
This is quite clearly seen for the Stellite sample between 10,000 and 20,000 cycles.
It is also quite possible that this is what is happening when the gap appears to
increase for this sample at 30,000 cycles. A similar variation is seen in rail wear rate
during twin – disc testing [12]. It is attributed to the accumulation of strain varying
with depth, (driven by varying contact stress with depth) and the strain history of the
material wearing away and reaching the rail surface. It is possible a similar
mechanism is behind the fluctuation in gap size observed here.

6.3 Four Point Bend Fatigue Tests (Series 3)

At a stress range of 350 MPa the Stellite 6 rail fails at 60,717 cycles. Based on a 3
car multiple unit train with a total of 12 axles this equates to 5,059 train passes. At
the lower stress range of 200 MPa the rail did not fail before the 670,000 cycle limit.
At this stress range the rail can be said to last 55,833 train passes. In service rail is
typically subject to a maximum stress range of 105 MPa. Therefore it is reasonable
to assume that if Stellite 6 clad rail were to be put into service it would far surpass
55,833 train passes before failure. The same can also be said of the MSS sample, but
with MSS being able to take much higher stress without failure. Therefore a MSS
sample clad rail subject to field conditions should be able to far outlast a Stellite 6
one.

One of the aims of the bend fatigue tests was to investigate the integrity of the
clad/substrate interface. Previous work in a number of locations around the world
had resulted in delamination of the clad layers under service conditions and this has
resulted in a significant degree of scepticism in the rail industry regarding clad
surface layers. It has been found that careful control of the microstructure developed
in the heat affected zone (HAZ) could avoid substrate cracking and clad layer
delamination [10]. Investigation of the failed Stellite 6 samples showed that the
clad/parent interface was still in-tact and there was no bulk delamination present see
Figures 11, 12, 13 and 14.

16
However, inclusions remain a problem in the samples. From the samples which
failed (ST6 – 350, ST6 – 275) it can be seen that cracking appears to start in the clad
layer or at the deposit/HAZ boundary and protrude into the substrate material. As
neither of the MSS clad samples failed at the higher stress range it is hypothesised
that the stress range in the upper most layer of the substrate material is either
insufficient to nucleate cracks or is sufficiently tolerant of small defects which may
exist in the deposit and hence the sample does not fail within the 670,000 cycle
limit. Calculations show that for a sample tested under the 350 MPa range the outer
fibre of the sample (clad layer) is under a stress of around 389 MPa. Assuming that
the clad layer is 2mm deep as designed then the maximum stress seen at the
clad/substrate interface is approximately 339 MPa. The outcomes will depend on the
presence of fatigue initiators i.e. inclusions and will also depend on the properties of
the HAZ. A brittle HAZ will not have a high fatigue strength and this depends on
the heat input and subsequent cooling rate during the cladding process. For the
Stellite 6 samples it is a combination of inclusions in the clad deposit and a
weakness in the interface/HAZ which leads to sample failure.

7 Conclusions
7.1 Wear/RCF Tests (Series 1)
Full-scale rolling sliding tests have been carried out on sections of rail which were
treated with laser cladding. The following conclusions can be made:

 No visible RCF was generated in any of the full-scale test specimens

 Some plastic deformation was seen in the baseline R260 grade sample this
being a precursor to RCF formation
 Bulk plastic flow through the depth if the rail was not seen in either of the clad
full-scale specimens
 No delamination of the clad layer was seen for either of the clad full-scale
specimens
 The Stellite 6 full-scale specimen wears at a rate approximately 51% lower
than the wear rate of the R260 Grade baseline specimen
 The MSS sample had only completed 50,000 cycles at the time of writing and
thus cannot be compared to R260 and Stellite 6 samples and thus no definite
conclusion can be made regarding the wear of this sample at this time

7.2 IBJ Lipping Tests (Series 2)

Full-scale rolling sliding tests have been carried out on sections of rail which were
treated with laser cladding and had a nylon endpost inserted halfway along their
length. It was found that:
 Cladding either side of an IBJ greatly improves its lipping resistance compared
to a non-clad standard R260 IBJ

17
  Cladding of an IBJ allows it to resist almost 3 times the energy input into the
contact as a standard unclad IBJ
 MSS and Stellite clads have approximately equal performance in terms of
resisting plastic flow at the simulated IBJ

7.3 Four Point Bend Fatigue Tests (Series 3)

Four point-bend fatigue tests have been carried out on clad sections of rail. Results
show that:

 The MSS bend samples did not fail within the 670,000 cycle limit while tested
at the highest stress range of 350 MPa
 The Stellite 6 bend samples did not fail at a stress range of 200 MPa, but did
fail at a higher stress ranges
 Failure of the Stellite 6 bend was linked to inclusions in the deposit. Some
cracks in the Stellite 6 bend sample also appear to originate at the
clad/substrate interface
 The deposition of a good quality clad with no internal inclusions thus seems to
be important to maintaining rail integrity
 MSS therefore shows greater fatigue resistance compared to Stellite 6 when
used as a cladding material on R260 grade rail
 No delamination of the clad deposited layers was seen in any of the broken
Stellite 6 fatigue samples. None of the MSS samples showed delamination
within the test time

The MSS clad sample wore at a rate of at least half that of R260 grade rail tested
under the same conditions assessed based on running band width development. MSS
also shows favourable bend fatigue resistance compared to Stellite 6, and good
lipping resistance when used to treat an IBJ.

References
[1] S. R. Lewis, R. Lewis, D. I. Fletcher, 2015. Assessment of laser cladding as an
option for repairing/enhancing rails. Wear, Vol. 330-331, pp 581 - 591
[2] S. R. Lewis, R. Lewis, P. S. Goodwin, S. Fretwell-Smith, D. I. Fletcher, K.
Murray, 2015. Improving Rail Wear and RCF Performance using Laser
Cladding, Proceedings of the 10th International Conference on Contact
Mechanics and Wear of Wheel/Rail Systems (CM2015), Colorado Springs,
USA
[3] Z. K. Fu, H. H. Ding, W. J. Wang, Q. Y. Liu, J. Guo, M. H. Zhu, 2015.
Investigation on microstructure and wear characteristic of laser cladding Fe-
based alloy on wheel/rail materials. Wear, Vol. 330-331, pp 592 – 599
[4] W. J. Wang, J. Hu, J. Guo, Q. Y. Liu, M. H. Zhu, 2014. Effect of laser cladding
on wear and damage behaviors of heavy-haul wheel/rail materials, Wear, 311,
pp 130–136

18
 [5] Beaty, P., Temple, B., Marshall, M.B., Lewis, R., 2015. Experimental
Modelling of Lipping in Insulated Rail Joints and Investigation of Rail Head
Material Improvements, in press, Journal of Rail and Rapid Transit,
Proceedings of the IMechE Part F.
[6] Whole Life Rail Model Application and Development for RSSB (T115) –
Rolling Contact Fatigue Testing, Rail Safety and Standards Board Research
Program Report, M. C. Burstow, February 2006.
[7] Network Rail. Identified failure modes of IBJs (based on Trust output). Report
no., 2010. London, UK: Network Rail.
[8] Zhou, L., Brunskill, H.P., Lewis, R., Marshall, M.B., Dwyer-Joyce, R.S., 2014.
Dynamic Characterisation of the Wheel/Rail Contact using Ultrasonic
Reflectometry, Proceedings of Railways 2014, The Second International
Conference on Railway Technology: Research, Development and Maintenance,
8-11 April, Corsica, France.
[9] Dukkipati, R.V., Dong, R., 1999. Idealized steady state interaction between
railway vehicle and track, Proceedings of the Institution of Mechanical
Engineers, Part F: Journal of Rail and Rapid Transit Vol. 213 (2), pp 15 – 29
[10] RailClad Project Deliverable D4, Metallographic Characterisation of Laser
Clad Rail Samples, Sandra Fretwel-Smith, Tata Steel Rail Technologies., 18th
February 2014.
[11] Hearn, E. J., 1977. Mechanics of Materials: an introduction to the mechanics of
elastic and plastic deformation of solids and structural components. Vol.2.
Oxford: Pergamon
[12] Tyfour, W.R., Beynon, J.H., Kapoor, A., The steady state wear behaviour of
pearlitic rail steel under dry rolling-sliding contact conditions. Wear, Vol. 180,
1995, pp 79-89

19

View publication stats