Project A6:Threedimensional Microstructural Modelling Of Crack Initiation In Rail Steel

1
Project A6 Three-dimensional Microstructural
Modelling Of Crack Initiation In Rail Steel
John Garnham, Claire Davis - Metallurgy and
Materials, Birmingham Francis Franklin, David
Fletcher - Mechanical Engineering, Newcastle
Objectives

• To characterise the 3-dimensional nature of crack
  initiation, early stage crack growth and
  microstructural deformation.
• To develop the mechanics model to include
  3-dimensional microstructural information.
• To provide maintenance planning guidance for
  infrastructure managers.

Now Mechanical Engineering, Sheffield.
2
Background
Compressive, directional, work-hardening of
pearlitic rail microstructures
Material response to cyclic loading. Above the
plastic shakedown limit there will be permanent,
incremental plastic deformation.
Top image R.Care EIS / ARUP / TTCI / Network
Rail, 2003 Bottom image After Bower Johnson,
WEAR 144 / CM1990
3
Background
Shakedown diagrams for rail microstructures Whe
el-rail contact loads are such that stresses are
in a zone near the plastic shakedown threshold
(top image). Different contact geometries
(wheel transverse location, wheel rail
profiles, etc.) plus dynamic forces can push
contact above the threshold, as shown by the
measured data (bottom image). The right
shakedown limit calculation is needed! Improved
calculations for strain hardening and layered
materials have been made since the theory was
introduced. Images R.Care EIS / ARUP / TTCI
/ Network Rail, 2003
4
Background (Project A2)
Microstructural observations, mechanisms and data

• Fatigue crack initiation in rail principally
  occurs along highly strained pro-eutectoid
  ferrite bands, located at prior austenite grain
  boundaries, or along highly strained pearlite
  nodule boundaries.
• Microstructures containing less pro-eutectoid
  ferrite exhibit long lives until fatigue crack
  initiation, however a larger degree of scatter in
  life is observed.
• Preferential straining of pro-eutectoid ferrite
  appears to be dependent on size and morphology.

Surface of 220 grade rail SUROS test disc.
Fatigue crack along strain flattened
pro-eutectoid ferrite with crack branch to
flattened ductile inclusion. (Section transverse
to rail axis.)
5
Modelling
Microstructure simulation for input into wear and
crack initiation model Initially a simple
arrangement of hexagons was used to represent
pearlite grains and pro-eutectoid (PE) ferrite.
Further developments to use more realistic grain
size variations and extensions to 3D capability
have occurred.
More realistic microstructure generated by
cross-section through the 3D microstructure
3D microstructure generated by the Voronoi 1000
cell method
Hexagonal representation.
6
Modelling
The wear - crack initiation model has been used
to investigate the effect of changing
microstructural parameters on the development of
damage and hence cracking. The model
predictions mirror the experimental results in
that microstructures with increased pro-eutectoid
ferrite show increased damage and shorter lives
until crack formation.
Model results showing increased strain in thicker
pro-eutectoid ferrite regions
7
3D RCF development - very small cracks
Observations of flaking on rail and twin disc
surfaces, coupled to sections through the flakes,
indicate that cracks form relatively quickly,
grow to one prior austenite grain length then
there is an incubation period before further
growth.
Surface flaking during twin disc rolling contact
fatigue testing of rail steel, at 25 RCF life.
8
3D RCF development - very small cracks
Very small crack shape is linked to the prior
austenite grain shape. Modelling indicates
multiple single grain cracks that are similar to
the multiple flakes and sectioned cracks observed
during experimental studies.
One grain length crack during twin disc rolling
contact fatigue testing of rail steel, at 25 RCF
life.
Crack path map, which starts with
surface-connected facets near-normal to the
x-axis, growing to near-parallel neighbouring
facets while shear deformation accumulates
9
3D RCF development - small/medium cracks
Serial sectioning of a 7 mm surface length RCF
crack, taken from a 260 grade trafficked rail,
has indicated that the crack has an approximate
semi-circular shape, as predicted by contact
mechanics.
Superimposed images of one crack at eight
polishing stages
3D crack shape data is being used to verify, and
adjust, modelling for fatigue crack growth in
rails.
10
3D RCF modelling of small / medium cracks
Crack data ( thick blue lines, right) has been
used to re-construct the crack for modelling The
data fitted a truncated semi-circle, i.e. a
semi-circular crack with its top (open) edge worn
away Most models have considered semi-circular
or semi-elliptical cracks. Using this new (real)
crack shape will improve crack growth rate
predictions The models are not FE based, so
accounting for non-idealised cracks is much more
than simply re-meshing
Some wheel contact pressure and crack location
combinations considered in modelling crack
growth rate
11
3D RCF development - medium/large cracks
Alternative 3D imaging of large cracks
Longer cracks have been examined in 3-dimensions
by comparative serial microscopy and X-ray
micro-tomography Larger cracks become complex in
shape (varying angles of propagation and
non-symmetric nature). Information is required
on the conditions for change from semi-elliptical
to complex for modelling.
X-ray computed micro-tomography 3D images of a
section cut around a small RCF crack in a rail
complete section to left and section with corner
slice digitally removed to right. Acknowledgeme
nt assistance of Mayorkinos Papaelias (Univ.
Birmingham) and TWI.
12
Additional observations - inclusions
In both worn rail and rail test discs,
near-contact surface, some of the distribution of
strain-flattened, ductile inclusions had
facilitated crack initiation and propagation.
Flattened MnS based, ductile inclusions near test
surface (view transverse to rail axis).
Disc
Rail
13
New and proposed research project
New (since 2008) rail research projects

• SAFERAIL EU project, 4.5M Euro overall budget
  with 3M Euro EU Contribution, (Bham)
• INTERAIL EU project, 4.8M Euro overall budget
  with 3.28M Euro EU contribution, (Bham)
• Network rail acoustic emission testing of
  cracking, 80k, (Bham)
• Network Rail 5 year research collaboration
  agreement with Engineering Faculty at Sheffield
• Korean Rail Research Institute (KRRI) to base
  researcher at Sheffield to developed links with
  UK rail research

Proposed rail research projects

• EPSRC Bridging the gap between academic and
  industrial degradation models (MMU / Bham / Newc
  / Sheff)
• EPSRC - Examining the role of ductile inclusions
  on crack initiation