Investigation of the risk for Rolling Contact Fatigue on ... - SimPack

<strong>Investigation</strong> <strong>of</strong> <strong>the</strong> <strong>risk</strong> <strong>for</strong> <strong>Rolling</strong> <strong>Contact</strong>
<strong>Fatigue</strong> on wheels <strong>of</strong> different passenger trains
Jakob Wingren, Centre <strong>of</strong> Competence <strong>for</strong> Vehicle Dynamics
prepared by
S. Stichel, H. Mohr, J. Ågren ,R. Enblom
Bombardier Transportation
SIMPACK UM 2007, Bonn

Background
During <strong>the</strong> last years RCF has become more common also
on passenger trains which resulted in wheel damages and
more frequent repr<strong>of</strong>iling and reduced <strong>the</strong> life time <strong>of</strong> <strong>the</strong>
wheels.
There<strong>for</strong>e in this study different trains have been
compared regarding <strong>the</strong>ir <strong>risk</strong> to develop RCF on <strong>the</strong>
wheels by calculating <strong>the</strong> contact conditions between
wheel and rail during curving with <strong>the</strong> multibody
simulation tool SIMPACK.
The overall aim <strong>of</strong> <strong>the</strong> study is to develop criteria that
indicate <strong>the</strong> <strong>risk</strong> <strong>for</strong> RCF already in <strong>the</strong> design stage <strong>of</strong> <strong>the</strong>
vehicle.
SIMPACK UM 2007, Bonn
2

The phenomenon <strong>of</strong> <strong>Rolling</strong> <strong>Contact</strong> fatigue
� Surface-initiated fatigue sometimes denoted
as spalling
� Subsurface-initiated fatigue sometimes
denoted as shelling
� <strong>Fatigue</strong> initiated at deep defects, sometimes
denoted as deep shelling or shattered rims
SIMPACK UM 2007, Bonn
3

Phases <strong>of</strong> surface initiated fatigue
� Crack initiation
� Crack propagation
� Crack branching towards tread surface
and wheel web
� Final fracture due to single overloads
SIMPACK UM 2007, Bonn
4

Visual appearance <strong>of</strong> spalling
SIMPACK UM 2007, Bonn
Cross section
Circumferential section
5

Effect <strong>of</strong> trapped fluid on crack propagation
Crack opens just be<strong>for</strong>e rail contact
SIMPACK UM 2007, Bonn
F t
Hydrostatic
pressure
Crack closes at rail contact
6

Two criteria tested
SIMPACK UM 2007, Bonn
� Shakedown map and fatigue index
- developed by Chalmers, Go<strong>the</strong>nburg
� RCF damage function (Tγ)
- developed by AEA Technology
7

Shake down map and fatigue index
SIMPACK UM 2007, Bonn
Limit line
<strong>Fatigue</strong> index developed by
Chalmers (Go<strong>the</strong>nburg)
FI surf
= µ −
8
k
p
0
2πabk
= µ −
3F
ζ

RCF damage function
Developed by AEA Technology
- four different areas
RCF Damage increases
No damage
SIMPACK UM 2007, Bonn
Damage (Nf x 1E-6)
15
10
5
0
-5
-10
-15
Tγ = T x ν x + T y ν y
RCF Damage function
0 50 100 150 200 250 300
Wear number, Ty (Nm/m)
Slower damage increase
due to increased wear
Cracks are worn away
due to severe wear
9

Shakedown evaluation
p0/k
p0/k
10,0
9,0
8,0
7,0
6,0
5,0
4,0
3,0
2,0
1,0
0,0
10,0
9,0
8,0
7,0
6,0
5,0
4,0
3,0
2,0
1,0
0,0
Vehicle 1, Shakedown map, inner wheel
0,0 0,1 0,2 0,3 0,4 0,5 0,6
SIMPACK UM 2007, Bonn
mue
Vehicle 3, Shakedown map, inner wheel
0,0 0,1 0,2 0,3 0,4 0,5 0,6
mue
1. R=739m
2. R=2922m
3. R=570m
4. R=736m
5. straight track
6. straight track
7. R=430m
boundary curve
1. R=739m
2. R=2922m
3. R=570m
4. R=736m
5. straight track
6. straight track
7. R=430m
boundary curve
p0/k
10,0
9,0
8,0
7,0
6,0
5,0
4,0
3,0
2,0
1,0
0,0
Vehicle 2, Shakedown map, inner wheel
0,0 0,1 0,2 0,3 0,4 0,5 0,6
mue
Vehicles 1 and 2 suffer from RCF
10
1. R=300m
2. R=600m
3. R=1000m
4. R=296m
5. R=761m
6. R=592m
7. R=800m
8. R=1000m
9. R=579m
boundary curve
Vehicle 3 does not suffer from RCF

Evaluation with damage parameter
1
0,8
0,6
0,4
0,2
0
Vehicle 1, D and FI versus R, nominal gauge
0 500 1000 1500 2000 2500 3000 3500
-0,2
0,8
0,6
0,4
0,2
0,0
-0,2
-0,4
SIMPACK UM 2007, Bonn
R in m
Vehicle 3, D and FI versus R, nominal gauge
0 500 1000 151500 2000 2500 3000 3500
R in m
D, P8
FI, P8
D, S1002
FI, S1002
D
FI
1
0,8
0,6
0,4
0,2
0
Vehicle 2, D and FI versus R, nominal gauge
0 200 400 600 800 1000 1200
-0,2
R in m
- Also <strong>the</strong> damage function produces
results that correlate to <strong>the</strong> real
behaviour <strong>of</strong> <strong>the</strong> wheels.
- The damage function seems to predict
too much damage <strong>for</strong> vehicle 2 or too
little damage <strong>for</strong> vehicle 1.
11
D
FI

KTH wear measure
1,5
1
0,5
0
-0,5
-1
1
0,8
0,6
0,4
0,2
0
-0,2
SIMPACK UM 2007, Bonn
Vehicle 2, nominal gauge
0 200 400 600 800 1000 1200
R in m
Vehicle 1, nominal gauge
0 500 1000 1500 2000 2500 3000 3500
R in m
D
FI
FI/KTHwr
D
FI
FI/KTHwr
Wear measure:
FI / Wear volume
Wear model developed by
Enblom, Jendel
KTH Rail Vehicles
12

Guideline <strong>for</strong> assessment
FI
0,25
0,20
0,15
0,10
0,05
-0,05
-0,10
-0,15
SIMPACK UM 2007, Bonn
Suggested limits <strong>for</strong> FI and D
0,00
0,00 0,20 0,40 0,60 0,80 1,00
Area 3
D
Area 1
Area 2
Area 1 High <strong>risk</strong> <strong>for</strong> RCF
A minor portion <strong>of</strong> <strong>the</strong> operation in
this area will lead to RCF damages.
Area 2 Risk <strong>for</strong> RCF
A larger portion <strong>of</strong> <strong>the</strong> operation in
this area will lead to RCF damages.
Area 3 Low <strong>risk</strong> <strong>for</strong> RCF
A significant portion <strong>of</strong> <strong>the</strong>
operation in this area may lead to
RCF damages.
13

Conclusions
� Both <strong>the</strong> shakedown <strong>the</strong>ory and <strong>the</strong> damage function can be used
<strong>for</strong> indicative predictions using quasistatic simulations.
� It seems that <strong>the</strong> damage function is having a slightly better
correlation to <strong>the</strong> experiences <strong>of</strong> RCF.
� The study indicates that <strong>the</strong> contact pressure may not be decisive
<strong>for</strong> <strong>the</strong> onset <strong>of</strong> RCF considering <strong>the</strong> results <strong>for</strong> vehicle 1/S1002,
i.e. high damage values and low FI. This is in line with <strong>the</strong>
experiences from <strong>the</strong> UK rail study.
� The evaluation <strong>of</strong> <strong>the</strong> FI/KTHwear measure shows that it can be
very sensitive to small variations in contact pressure. It is likely
that this sensitivity will decrease if applying it to transient
simulations with track irregularities due to average effects.
SIMPACK UM 2007, Bonn
14

Fur<strong>the</strong>r work
� Analyze more vehicles in order to improve <strong>the</strong> validation.
� Develop an evaluation method able to accumulate damage at
different positions on <strong>the</strong> tread during transient simulations with
track irregularities.
� Define a methodology to accumulate damage considering both <strong>the</strong>
curve and gauge distribution.
� Implement braking <strong>for</strong>ces in <strong>the</strong> simulations.
SIMPACK UM 2007, Bonn
15