05 Britto's Line 2 3 Update To MSP 2013-05-06

Transportation Technology Center, Inc.
, a subsidiary of the Association of American Railroads
Update on TTCI/MSP
Vehicle Dynamics
Projects
Ruben Pea
Stan Gurule
Russ Walker
March 26, 2015
TTCI/AAR, 1/10/2012. filename, p1
Introduction
Line 3 both Cobrasma (MTE) and (Mafersa/Tokyu)

Objective: To learn how the age of the primary
suspension affects performance. This will allow MSP to
decide on maintenance intervals
Line 2 Millenium Fleet Cars (Mafersa/Tokyu)
Characterize and measure the performance of the Line
2 Millenium fleet and make recommendations for any
necessary changes to optimize suspension
performance.
Line 5 (CAF/Alstom)
Characterize and measure the performance of the Line
5 fleet and make recommendations for any necessary
changes to optimize suspension performance.
TTCI/AAR, 1/11/2012, Filename p2
Primary Suspension
Line 3 MTE/Cobrasma
The
Cobrasma/MTE
French Design
Primary
Suspension
The Cobrasma/MTE French Design

Complete Truck
Primary Suspension
Line 2 and 3 Tokyu/Mafersa
Exploded View of
Primary Chevrons of
Mafersa/Tokyu
Truck
Lines 2 and 3
Alstom/CAF
Line 5 Car
Bogie
Arrangement
View
Line 5 Car
Bogie Arrangement Line 5 Car
Method
Characterize the track

Characterize the cars
Truck load equalization (dQ/Q) testing of cars
Test the cars on Jabaquara curve and on perturbations
at Belem yard, or at Capao Redondo for Line 5
Use computer modeling of the car to estimate wheel
loads and L/V Ratio in a variety of conditions
What We Need to Prevent!
Performance Criteria
Parameter
The performance measures

selected for this computer
modeling study are based
on the United States (US)
Code of Federal Regulations
(CFR) Title 49 Part 213.333
Vehicle Track Interaction
safety limits.
The wheel L/V safety limit is
a function of flange angle.
The G-29 flange angle is
64.753 degrees
corresponding to a wheel
L/V safety limit of 0.79.
Vehicle/Track Interaction Safety Limits

Safety Limit
Filter/Window
Requirements
Wheel/Rail
1
Forces
0.1
5 ft
Single Wheel
Vertical Load
Ratio
Single Wheel
L/V Ratio
.5
1
.5
5 ft
No wheel of the equipment shall be permitted

to unload to less than 10% of the static
vertical wheel load. The static vertical wheel
load is defines as the load that the wheel
would carry when stationary on level track.
The vertical wheel load limit shall be
increased by the amount of measurement
error.
The ratio of the lateral force that any wheel
exerts on an individual rail to the vertical force
exerted by the same wheel on the rail shall be
less than the safety limit calculated for the
wheels flange angle ( ).
The net lateral force exerted by any axle on
the track shall not exceed 50% of the static
vertical load that the axle exerts on the track.
The ratio of the lateral forces that the wheels
on one side of any truck exert on an individual
rail to the vertical forces exerted by the same
wheels on that rail shall be less than 0.6
Net Axle L/V

Ratio
0.5
5 ft
Truck Side
L/V Ratio
0.6
5 ft
0.5 g peakto-peak
10 Hz 1 sec
window
The peak-to-peak acceleration, measured as

the algebraic difference betweeen the two
extreme values of measured acceleration in a
one second time period, shall not exceed 0.5
g.
Carbody
Vertical
0.6 g peakto-peak
10 Hz 1 sec
window
The peak-to-peak acceleration, measured as

the algebraic difference betweeen the two
extreme values of measured acceleration in a
one second time period, shall not exceed 0.6
g.
Truck
3
Lateral
0.4 g RMS
meanremoved
10 Hz 2 sec
window
Truck hunting shall not develop below the

maximum authorized speed.
Acceleration
Carbody
2
Lateral
1. The lateral and vertical wheel forces shall be measured with instrumented wheelsets
with the measurements processed through a low pass filter with a minimum cut-off
frequency of 25 Hz. The sample rate for wheel force data shall be at least 250
samples/sec.
2. Carbody lateral and vertical accelerations shall be measured near the car ends at the
floor level.
3. Truck accelerations in the lateral direction shall be measured on the truck frame. The
measurements shall be processed through a filter having a pass band of 0.5 to 10 Hz.
4. Truck hunting is defined as a sustained cyclic oscillation of the truck which is
evidenced be lateral accelerations in excess of 0.4 g root mean square (mean
removed) for 2 seconds.
Vehicle Measurements
Wayside Instrumentation
Bill, Mazza, and Mike placed strain gages on the rails

to measure vertical and lateral wheel forces.
Capao Redondo
Belem
Jabaquara
Curve
2 Sites on Line
2
1+ Site on
Line 5
Wayside Instrumentation
Wayside measurements
provide:
Vertical Wheel Forces
Example of Wayside Data
L/V Ratio
These can be compared

directly to the metrics
Only available at specific
points on the track
Vehicle Dynamics Model
Modeling vehicle using NUCARS

Multibody vehicle dynamics model.
Inputs
Masses and mass moments of inertia
Stiffness
Damping
Friction
Outputs
Wheel/rail forces
Body motions
.
Includes Track Model
Nucars can model the rail and ties (or slab) as masses
with many degrees of freedom
Requires that we characterize the track in addition to
the vehicle
Track Characterization
Track characterization provides inputs for the computer

model
Done with 2 types of tests
Track Hammer Test
Allows us to determine the stiffness of rail connections to the

slab
Instrumentation on slab to measure displacements under traffic
Allows us to determine the stiffness of the slab to ground
Track Hammer Test
Accelerometers are placed on the rail and on the tie or

slab
The rail is struck with an instrumented hammer
The vibration of the track helps identify the stiffness of
the fasteners
Slab Track Instrumentation
Sections of track
were instrumented
with rail force
circuits and slab
displacement
DZ07
measurements
DZ05
124m slab in curve

suspended on coil
springs
DZ08
DY03
DZ06
62m
DZ01
Adjacent slab, not
suspended on coil
springs
DZ03
DZ04
DY02
Vertical
Displacement
Measurement
31m
DZ02
DY01
Lateral
Displacement
Measurement
Example of Slab Track Data
Vehicle Characterization Test
These tests are only to identify model inputs

Suspension Stiffness Test
Car body modal test Used to estimate the car center of
gravity height and car body mass moments of inertia

Bogie modal test Used to identify the bogie mass moments of
inertia
Stiffness Testing
Lateral Stiffness Test
Longitudinal Stiffness Test
3.5
3
y=5.7293x+0.1249
R=0.9817
Force(kN)
2.5
2
1.5
1
0.5
PrimaryLateral(Loaded)
Linear(PrimaryLateral(Loaded))
0.5
0.2
0.2
0.4
0.6
Displacement(mm)
Vertical Stiffness Test

Carbody Modal Testing
Car is instrumented with enough accelerometers to

identify the five modes of vibration
Car is excited by hand or by quickly releasing a jack to
let the car settle
Bogie Modal Test
Truck Load Equalization
The truck must

distribute the load to
the track equally on
all the wheels
This is key to
preventing flange
climb derailment
Flexible body
Stiff Body
Truck Load Equalization
Raise 1 wheel of a truck while measuring load on all

4 wheels
APTA SS-M-014-06
Class G
Used where track standards allow a difference in
crosslevel of no more than 76.2mm in 18.9m
35% wheel load at 63.5mm and >0% wheel load at
76.2mm
Class R Using this level MSP can advise if it is
appropriate
Used where track standards allow difference in
crosslevel of no more than 76.2mm in 18.9m and
no more than 57.2mm in 3.0m
35% wheel load at 50.8mm and >0% wheel load at
63.5mm
This standard needs

to be evaluated
considering MSP
track maintenance
practices.
Curve Testing
Curve tests were performed either at

Jabaquara on a 154m curve or at
Capao Redondo on a 300m curve
1
0.9
0.8
0.7
L/VRatio
0.6
0.5
0.4
0.3
LoadedSouth
0.2
LoadedNorth
0.1
0
0
Wayside gages recorded L/V ratios

and onboard measurements recorded
suspension displacements
10
15
20
25
30
Speed(kph)
1
0.9
0.8
L/VRatio
0.7
0.6
0.5
0.4
EmptySouth
0.3
EmptyNorth
0.2
0.1
0
0
10
15
20
25
30
35
Speed(kph)
35
Perturbed Track Tests
Track geometry (deviations in

the track) may excite the car in
some mode of rigid body
vibration
Perturbed Track Tests
This sketch shows how the perturbed track zones are

set up at Belem Yard
Similar zones are installed at Capao Redondo for line
5 tests
N
Yaw
Bounce
Car 3144
Yaw tests and bounce tests

were performed eastbound
with car 3139 leading
Car 3139
Roll
Roll tests were

performed westbound
with car 3144 leading
Outline of Line 3 Results
Review results for line 3 cars Cobrasma (MTE) and

Maferasa (Tokyu)
Model Validation
Test result from recently re-profiled wheel and the same
wheel after re-profiling marks had worn away.
Simulations
Turnout simulations
Worn wheels and rails

Gage clearance variations
Turnout defects
Primary suspension stiffness variation
Model Validation
Line 3
Truck load equalization

predictions match test
results showing that the
model correctly represents
the primary suspension
Cobrasma
R35%
200%
G35% G0%
R0%
Test1L
Test1R
Test2L
Test2R
Model1L
Model1R
Model2L
Model2R
150%
100%
50%
35%WheelLoad
0%
PercentVerticalLoad
PercentVerticalLoad
200%
Mafersa
Test1L
Test1R
Test2L
Test2R
Model1L
Model1R
Model2L
Model2R
150%
100%
R35%
G35% G0%
R0%
50%
35%WheelLoad
0%
0
20
40
60
Wheel1LHeight(mm)
80
20
40
60
Wheel1LHeight(mm)
80
Model Validation Line 3
Models reproduce dynamic test results (Example: Em T&R)

Cobrasma
Mafersa
60
TestLLVert
40
TestLRVert
30
TestTLVert
TestTRVert
20
ModelLLVert
10
ModelLRVert
ModelTLVert
10
30
50
Speed(kph)
70
ModelTRVert
Dsiplacement(mm)
Displacement(mm)
50
50
TestLLVert
40
TestLRVert
30
TestTLVert
TestTRVert
20
ModelLLVert
10
ModelLRVert
ModelTLVert
10
30
50
Speed(km/h)
70
ModelTLVert
Key Test Result Cobrasma Car
The wheels on the lead truck of the Cobrasma car were

re-profiled just before the first empty curving test. The
measured L/V ratios were very high.
The test was repeated after some wear in. The L/V ratio
had dropped significantly.
1.2
L/VRatio
1
0.8
0.6
0.4
ImmediatelyAfterReprofilingOuterRail
ImmediatelyAfterReprofilingInnerRail
RepeatAfterWearInOuterRail
RepeatAfterWearInInnerRail
0.2
0
0
10
20
Speed(kph)
30
40
Line 3 Turnout Simulations Line 3

Input
Covers conditions in RFP

New and worn wheel profiles
New and curve worn rail
profiles
Diverging route simulation of
1:8, 1:9, 1:11.4, and 1:14
turnouts
Switch embedding for 1:8,
1:9, 1:11.4, and 1:14 turnouts
Speeds of 20, 40, and
maximum civil speed for the
turnout. Maximum civil
speeds are 44kph for 1:8 and
1:9, 63kph for 1:11.4, and
73kph for 1:14.
Track and wheelset gauge
variations
1:8
Wheel
Profile
New and
Worn
New and
Worn
New and
Worn
New and
Worn
New
Rail Profile
New or Measured
and Worn
New or Measured
and Worn
New or Measured
and Worn
New or Measured
and Worn
New or Measured
Speed
System File
20, 40,
Empy Cab Leading
44kph
and Trailing
1:9
20, 40,
Empy Cab Leading
44kph
and Trailing
1:11.4
20, 40,
Empy Cab Leading
63kph
and Trailing
1:14
20, 40,
Empy Cab Leading
73kph
and Trailing
1:8
20, 40,
Loaded Cab Leading
44kph
and Trailing
1:9
New
New or Measured
20, 40,
Loaded Cab Leading
44kph
and Trailing
1:11.4
New
New or Measured
20, 40,
Loaded Cab Leading
63kph
and Trailing
1:14
New
New or Measured
20, 40,
Loaded Cab Leading
73kph
and Trailing
1:8 narrow
New Wide
New or Measured
20, 40,
Empy Cab Leading
gage
B-B
44kph
and Trailing
1:9 narrow
New Wide
New or Measured
20, 40,
Empy Cab Leading
gage
B-B
44kph
and Trailing
1:11.4 narrow New Wide
New or Measured
20, 40,
Empy Cab Leading
gage
B-B
63kph
and Trailing
1:14 narrow
New Wide
New or Measured
20, 40,
Empy Cab Leading
gage
B-B
73kph
and Trailing
1:8 wide
Worn
Worn
20, 40,
Empy Cab Leading
gage
Narrow B-B
44kph
and Trailing
1:9 wide
Worn
Worn
20, 40,
Empy Cab Leading
gage
Narrow B-B
44kph
and Trailing
1:11.4 wide
Worn
Worn
20, 40,
Empy Cab Leading
gage
Narrow B-B
63kph
and Trailing
1:14 wide
Worn
Worn
20, 40,
Empy Cab Leading
gage
Narrow B-B
73kph
and Trailing
1:8 with
New
New or Measured
20, 40,
Empy Cab Leading
embedding
44kph
and Trailing
1:9 with
New
New or Measured
20, 40,
Empy Cab Leading
embedding
44kph
and Trailing
1:11.4 with
New
New or Measured
20, 40,
Empy Cab Leading
embedding
63kph
and Trailing
1:14 with
New
New or Measured
20, 40,
Empy Cab Leading
embedding
73kph
and Trailing
Total Number of simulations for Cobrasma (and the same number for Mafersa)
Num.
of Sim.
24
24
24
24
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
192
Line 3 Turnout Simulations
Embedding Switchpoint defect cased
Defect
Switch Rail
Mainline Stock Rail
Defect Length Varies 0 to 2.1m
Ramp Length = 1m
Line 3 Primary Stiffness Variations
Simulating vehicle
performance with a wide
range of primary
suspension stiffness.
Measured stiffness of worn
suspension is normally
higher that the nominal
stiffness given on drawings
or procurement standards
Higher levels of suspension
stiffness will be simulated
as well
Primary suspension stiffness values being

simulated for the Cobrasma/MTE car.
Body
Nominal Stiffness
(Long and Lat Estimated)
Measured Stiffness
2 x Increase
3 x Increase
Longitudinal
Primary Stiffness
(kN/mm)
7.0
Lateral Primary
Stiffness
(kN/mm)
4.0
Vertical Primary
Stiffness
(kN/mm)
1.735
15.0
23.0
31.0
9.0
14.0
19.0
3.6
5.5
7.4
Primary suspension stiffness values being

simulated for the Mafersa/Tokyu car.
Body
Nominal Stiffness
(except longitudinal)
Measured Stiffness
2 x Increase
3 x Increase
(longitudinal is nominal)
Longitudinal
Primary Stiffness
(kN/mm)
17.0
Lateral Primary
Stiffness
(kN/mm)
9.5
Vertical Primary
Stiffness
(kN/mm)
5.9
30.4
57.1
83.8
28.9
48.3
67.7
9.9
13.9
17.9
Simulation cases:
(Jacking Test) with

airbags inflated
with airbags deflated
Nominal 1:9 turnout
1:9 turnout with truck
twist track perturbation
1:9 turnout with truck
twist perturbation with
double amplitude
Specific cases of asymmetric suspension conditions

One case investigates performance of a suspension if 1
element is stiffer than the other 3.

Three of the primary springs of the lead truck will use the
measured stiffness, and the lead inside wheel will use 2x the
measured stiffness.
Two cases investigate performance of an improperly shimmed
truck.
All 4 primary springs of the lead truck will use the Measured
Stiffness values.
Lead outside spring will be offset 5mm or 10mm below the
other 3 springs of that truck (to take less weight).
These conditions will be simulated for the 1:9 turnout with the
track twist perturbation.
Wheel/Rail Profile Simulation Predictions

Line 3
Simulations using new and worn wheel profiles show

that worn profiles are worse, but both cars meet the
FRA 213 safety criteria for these combinations of
wheel/rail profiles on nominal inputs
Cobrasma
NewWheel,NewRail
NewWheel,WornRail
WornWheel,NewRail
WornWheel,WornRail
Limit
1
0.8
0.6
0.4
0.8
0.6
0.4
0.2
0.2
0
15
35
NewWheel,NewRail
NewWheel,WornRail
WornWheel,NewRail
WornWheel,WornRail
Limit
1.2
WheelL/V
1.2
WheelL/V
Mafersa
55
75
15
Speed(kph)
1:9 Turnout
35
55
75
Speed(kph)
1:9 Turnout
Gage Clearance Simulation Predictions

Line 3
Simulations using narrow and wide gage clearance

show that nominal gage clearance perform better than
either narrow or wide gage clearance.
Cobrasma is close to but meets FRA 213 wheel L/V safety
criteria
Mafersa is close to but does not meet FRA 213 wheel L/V
safety criteria
Cobrasma
NewWhandRaStGaandBB
NewWhandRaNarGaWideBB
WornWhandRaStGaandBB
WornWhandRaWideGaNarBB
Limit
1
0.8
0.6
0.4
0.8
0.6
0.4
0.2
0.2
0
15
35
55
NewWhandRaStGaandBB
Limit
1.2
WheelL/V
1.2
WheelL/V
Mafersa
75
15
Speed(kph)
1:9 Turnout
35
55
75
Speed(kph)
1:9 Turnout
Turnout Defect Simulation Predictions

Line 3
The FRA 213 wheel L/V criteria with a 5-foot window

does not properly indicate the tendency for flange climb
derailment for turnout defect simulations.
The high angle of attack and the large lateral displacements
cause derailments in an extremely short distance.

Because of this an output that indicates the position of flange
contact was substituted.

Line 3
Simulations show that long defect lengths create potential for

flange climb derailment on 1:8 and 1:9 turnouts.
Cobrasma: Defect lengths should be limited to 1m or less for 1:8 and
1:9 turnouts
Mafersa: Defect lengths should be limited to 1.4m and 0.7m for 1:8
and 1:9 turnouts respectively.
Because both cars use the same lines the most restrictive case should
be used.
1:9
Mafersa
0.03
New
0.025
0.7mDefect
0.02
1.0mDefect
0.015
1.4mDefect
0.01
1.6mDefect
1.8mDefect
0.005
2.0mdefect
0
15
25
35
Speed(kph)
45
2.1mDefect
RollingRadius
(toindicateflangecontactposition)
RollingRadius
Cobrasma
Limit
1:9
0.03
New
0.025
0.7mDefect
0.02
1.0mDefect
0.015
1.4mDefect
0.01
1.6mDefect
1.8mDefect
0.005
2.0mdefect
0
15
25
35
Speed(kph)
45
2.1mDefect
Limit
Primary Stiffness Simulation Predictions

Line 3
Primary stiffness has a significant effect on truck load

equalization simulation predictions.
The truck load equalization test uses very large
displacements (up to 75mm) compared to MSPs track
twist limit (12mm).
R35%
100%
Mafersa
G35% G0%
R0%
Nominal
Measured
2xIncrease
3xIncrease
35%Load
50%
35%WheelLoad
0%
PercentVerticalLoad
PercentVerticalLoad
Cobrasma
R35%
100%
G35% G0%
R0%
Nominal
Measured
2xIncrease
3xIncrease
35%Load
50%
35%WheelLoad
0%
0.0
EmptyAirbagDeflated
20.0 40.0 60.0 80.0

Wheel1LHeight(mm)
0.0
EmptyAirbagDeflated
20.0 40.0 60.0 80.0

Wheel1LHeight(mm)

Line 3
While the Cobrasma car is below the limit for all primary
suspension stiffness values checked, the Mafersa car is
very close to the limit in all cases.
The Nominal stiffness for the Mafersa car is greater than
the 2 x Increase case for the Cobrasma car.
Cobrasma
1.0
0.8
0.6
0.4
1.2
1.0
WheelL/V
Nominal
Measured
2xIncrease
3xIncrease
Limit
1.2
WheelL/V
Mafersa
0.8
0.6
0.4
0.2
0.2
0.0
0.0
10
Perturbation
20
30
Speed(kph)
40
Nominal
Measured
2xIncrease
3xIncrease
Limit
10
20
30
Speed(kph)
40
Perturbation
Special Cases Simulation Predictions

Line 3
The softer primary stiffness of the Cobrasma car makes

it more tolerant of conditions that cause asymetrical
loading of the truck, such as improper shimming or
variations in primary suspension stiffness.
Mafersa
0.8
0.8
WheelL/V
WheelL/V
Cobrasma
0.6
MeasuredStiffness
HigherStiffnessononeaxlebox
ImproperShimononeaxlebox(5mm)
ImproperShimononeaxlebox(10mm)
0.4
0.2
0.6
MeasuredStiffness
Higherstiffnessononeaxlebox
Impropershimononeaxlebox(5mm)
Impropershimononeaxlebox(10mm)
0.4
0.2
0
0
0
10
20
30
Speed(kph)
40
50
10
20
30
Speed(kph)
40
50
Summary Final Report Recommendations

for Line 3 Cobrasma
The line 3 Cobrasma car performed well with the highest

stiffness simulated, 7.4 kN/mm, which is twice the the
stiffness measured on the worn suspension. MSP
should evaluate the risks of extending the current
interval of suspension replacement.
Summary Final Report Recommendations

for Line 3 Mafersa
The maximum primary stiffness of the Mafersa car

should be maintained at 10kN/mm at each axle box.
This is consistent with the stiffness measured at MSPs
current maintenance interval.
MSP should continue to a maintain track twist of 12mm in 4 m
on all of their tracks.

This high level of primary stiffness makes the vehicle more
susceptible to deterioration in performance caused by
asymmetric primary suspension setup. Use of a truck twist
index computed from wayside force measurements can identify
trucks with abnormal primary suspension setups. Such a
system could be used on Line 3 to identify cars with this
problem.
Additional Recommendations
MSP should refine surface finish requirements for wheel

reprofiling to minimize the effects seen during the line 3
Cobrasma empty curving test.
MSP should limit the size of turnout defects to 1.0m and
0.7m on 1:8 and 1:9 turnouts respectively.
MSP should consider pursuing a Wheel/Rail profile
study. A wheel profile with a steeper flange angle over a
longer distance on the flange would increase the L/V
ratio and distance to climb required for flange climb
derailment. Using a profile design based on the worn
shape may also reduce wheel wear, although care must
be taken to avoid rolling contact fatigue of wheels and
rails.
Outline of Line 2 Results
Review results for line 2 Millennium Cars

Model Validation
Simulations
Turnout simulations
Worn wheels and rails

Gage clearance variations
Turnout defects
Primary suspension stiffness variation
Possible suspension improvements Reducing primary
suspension vertical stiffness
Analysis of wayside data
Line 2 Model Validation

simulation predictions match
test results very closely.
Empty
R35% G35% G0%
R0%
Axle1L
150%
Axle1R
100%
Axle2R
Axle2L
Model1L
Model1R
50%
Model2L
35%WheelLoad
0%
Model2R
20
40
60
Wheel2LHeight(mm)
80
35%Load
Limit
200%
PercentVerticalLoad
PercentVerticalLoad
200%
Loaded
R35% G35% G0%
R0%
Axle1L
150%
Axle1R
100%
Axle2R
Axle2L
Model1L
Model1R
50%
Model2L
35%WheelLoad
0%
Model2R
20
40
60
Wheel2LHeight(mm)
80
35%Load
Limit
Line 2 Model Validation
Pitch and Bounce Model vs Test

Empty
Loaded
25
ModelLLVert
20
ModelLRVert
15
ModelTLVert
ModelTRVert
10
TestLLVert
TestLRVert
TestTLVert
10
30
50
Speed(kph)
70
TestTRVert
Displacement(mm)
Displacement(mm)
25
20
ModelLLVert
ModelLRVert
15
ModelTLVert
10
ModelTRVert
TestLLVert
TestLRVert
TestTLVert
0
10
30
50
Speed(kph)
70
TestTRVert
Wheel/Rail Profile Simulation Predictions

Line 2
Simulations using new and worn wheel and rail profiles

show that worn profiles are worse, but all combinations
of wheel/rail profiles meet the FRA 213 safety criteria for
nominal inputs.
NewRail,NewWheel
NewRail,WornWheel
WornRail,NewWheel
WornRail,WornWheel
Limit
1.2
WheelL/V
1
0.8
0.6
0.4
0.2
0
15
35
55
Speed(km/h)
75
1:9 Turnout
Gage Clearance Simulation Predictions

Line 2
Simulations using narrow and wide gage clearance

show that nominal gage clearance performs better than
either narrow or wide gage clearance. All combinations
of gage clearance meet the FRA 213 safety criteria for
nominal inputs.
NewWhandRaStGaandBB
Limit
1.2
WheelL/V
1
0.8
0.6
0.4
0.2
0
15
35
55
Speed(km/h)
75
1:9 Turnout

Line 2
RollingRadius
Simulations of turnout defects show 1:8 and 1:9

turnouts with defects longer than 1.6m and 1.0m
respectively have a tendency for flange climb
derailment. For 1:11.4 and 1:14 turnouts defect lengths
up to 2.1m (the longest simulated) do not show a
tendency for flange climb derailment.
1:9
0.03
New
0.025
0.7mDefect
0.02
1.0mDefect
0.015
1.4mDefect
0.01
1.6mDefect
1.8mDefect
0.005
2.0mdefect
0
15
25
35
Speed(km/h)
45
2.1mDefect
Limit

Line 2
R35% G35% G0%

R0%
100%
1.2
Nominal
Measured
2xIncrease
3xIncrease
35%Load
50%
35%WheelLoad
0%
1.0
WheelL/V
PercentVerticalLoad
Simulation predictions show that while primary stiffness

has a significant effect on truck load equalization
results, the effects are smaller for simulations using
track twist perturbations based on the normal MSP track
twist maintenance limits.
0.8
0.6
Nominal
Measured
2xIncrease
3xIncrease
Limit
0.4
0.2
0.0
0.0
EmptyAirbagDeflated
20.0 40.0 60.0 80.0

Wheel1LHeight(mm)
10
20
30
Speed(km/h)
40
Perturbation
Special Case Simulation Predictions

Line 2
Line simulations to investigate asymmetrical primary

suspension configuration were based on the condition
of a car involved in a yard derailment in May 2011.
One axle of the derailed truck had new primary
suspension elements while the other axle had worn
ones.
The wheels of the derailed truck had recently been reprofiled
Static loads on the derailed truck showed asymmetrical
wheel loads
Location
Side 1
Side 2
First Axle
33.5 kN
60.5 kN
Second Axle
60.1 kN
30.7kN
Special Cases Simulation Predictions

Line 2
Simulation predictions show that asymmetry in the

primary suspension can cause performance to degrade.
Simulations designed to replicate the static wheel loads
measured on a Line 2 car that derailed in May 2011
showed that the asymmetrical loading caused lower
vertical loads and higher L/V ratios
VerticalLoad
100%
80%
60%
40%
20%
0%
1.0
MaximumWheelL/V
NominalCondition
DerailedCondition
Limit
120%
0.8
0.6
NominalCondition
0.4
DerailedCondition
0.2
Limit
0.0
10
20
30
Speed(km/h)
40
10
20
30
Speed(km/h)
40
Suspension Improvements Reducing

Primary Stiffness, Line 2
Reducing primary stiffness 30% below nominal

improves truck load equalization performance but has
little effect on simulations that use track twist
perturbations similar to the 12mm maintenance limit
followed by MSP.
1
OriginalModel1L
OriginalModel1R
150%
0.8
OriginalModel2L
R35% G35% G0%
R0%
100%
OriginalModel2R
RevisedModel1L
RevisedModel1R
50%
RevisedModel2L
35%WheelLoad
0%
0.0
50.0
Wheel2LHeight(mm)
RevisedModel2R
35%Load
Limit
WheelL/V
PercentVerticalLoad
200%
0.6
ReducedPrimaryStiffness
0.4
Nominal
0.2
0
0
12mmPerturbationCL
10
20
30
Speed(km/h)
40
50
Data from Line 2 Track Characterization
Use wheel-rail force data for a vehicle performance

survey.
Wheel L/V ratio indicates curving performance
Criteria = 0.79
Defined a car body twist index to indicate secondary
suspension performance.
(L1-R1+L2-R2-L3+R3-L4+R4)/Car Weight
No established criteria. Use Quartile analysis to find outliers.
Defined a truck twist index to indicate primary suspension
performance.
(L1-R1-L2+R2)/Truck weight.
No established criteria. Use Quartile analysis to find outliers.
Analysis of Wayside Data Line 2
Wayside force measurements can

be used to identify trucks with
poor primary suspension
condition. If not already doing so,
MSP should consider using a
truck twist index to identify poorly
performing primary suspensions.
For comparison, the truck twist
index for car that derailed in May
2011 was 0.31
MaxL/VRatiobyTruck
0.08
0.06
0.04
CarTwistIndex
OuterFence0.104
14:00
16:00
18:00
Time May24,2011
TruckTwistIndex
CarTwistIndex
0.1
0
12:00
0.8
L/VRatio
OuterFence0.49
WheelL/VSafetyLimit
0.6
0.4
0.2
0
12:00
14:00
16:00
Time May24,2011
18:00
0.2
0.12
0.02
0.15
0.1
0.05
0
12:00
TruckTwistIndex
OuterFence0.13
14:00
16:00
18:00
Time May24,2011
Track Modeling and Characterization
What is the NUCARS Track model and how does it work

Multi-body representation of track structure
Flexible rails and ties (euler-bernoulle beams)

Up to 89 modes each: vertical, lateral, twist
Standard NUCARS connections for track fasteners and
ballast
Multiple layers possible
W/R Penetration model
Rail profiles can vary along track
Single Layer Model
Two Layer Model
Flexible rail modes
Line 2 Coil Spring Slab

Simulating 123m slab with 20 modes

25
20
15
Acceleration(g)
Hammer testing results.

Frequency match is good.
Indicates correct stiffness
of rail-slab fastener.
About 80kN/mm
Line2CoilSprings OverFastener
Model
Test
10
5
0
5
0.49
0.51
0.57
0.59
15
Time
25
Line2CoilSprings 1/2WayBetweenFasteners
20
Line2CoilSprings 1/4WayBetweenFasteners
20
15
10
5
0
10
5
0
10
10
15
15
0.49
0.51
0.53
0.55
Time(s)
0.57
Model
Test
15
Model
Test
Acceleration(g)
Acceleration(g)
0.55
10
20
25
0.53
0.59
0.49
0.51
0.53
0.55
Acceleration(g)
0.57
0.59
Slab displacement shows good agreement between

model and test. About 10 kN/mm.
0.0005
Line2CoilSprings
Displacement(m)
0.0005
0.001
0.0015
0.002
0.0025
DZ07
model
0.003
0.0035
0.004
Time(s)
Line 2 Conclusions
The maximum primary stiffness in the worn condition

should be maintained at 10 kN/mm at each axle box or
less for Line 2 Millennium Fleet cars. This is consistent
with the worn stiffness measured at MSPs current
primary suspension maintenance interval. To ensure
safety at this stiffness level further recommendations
are offered:
MSP should continue to maintain a maximum track twist of 12
mm in 4 m on all their tracks.

This high level of primary stiffness makes the vehicle more
susceptible to deterioration in performance caused by
asymmetric primary suspension setup. Use of a truck twist
index computed from wayside force measurements can identify
trucks with abnormal primary suspension setups.
Additional Recommendations
Simulations of turnout defects show that 1:8 and 1:9

turnouts with defects longer than 1.6m and 1.0m
respectively have a tendency for flange climb
derailment.
MSP should refine surface finish requirements for wheel
reprofiling to minimize the effects seen during the line 3
Cobrasma empty curving test.
MSP should consider pursuing a Wheel/Rail profile
study. A wheel profile with a steeper flange angle over a
longer distance on the flange would increase the L/V
ratio and distance to climb required for flange climb
derailment. Using a profile design based on the worn
shape may also reduce wheel wear, although care must
be taken to avoid rolling contact fatigue of wheels and
rails.

05 Britto's Line 2 3 Update To MSP 2013-05-06

Încărcat de

Informații document

Titlu original

Drepturi de autor

Formate disponibile

Partajați acest document

Partajați sau inserați document

Opțiuni de partajare

Vi se pare util acest document?

Este necorespunzător acest conținut?

Drepturi de autor:

Formate disponibile

05 Britto's Line 2 3 Update To MSP 2013-05-06

Încărcat de

Drepturi de autor:

Formate disponibile

Transportation Technology Center, Inc.

, a subsidiary of the Association of American Railroads

TTCI/AAR, 1/10/2012. filename, p1

Line 3 both Cobrasma (MTE) and (Mafersa/Tokyu)

TTCI/AAR, 1/11/2012, Filename p2

TTCI/AAR, 1/11/2012, Filename p3

TTCI/AAR, 1/11/2012, Filename p4

The Cobrasma/MTE French Design

TTCI/AAR, 1/11/2012, Filename p5

TTCI/AAR, 1/11/2012, Filename p6

TTCI/AAR, 1/11/2012, Filename p7

TTCI/AAR, 1/11/2012, Filename p8

TTCI/AAR, 1/11/2012, Filename p9

Bogie Arrangement Line 5 Car

TTCI/AAR, 1/11/2012, Filename p10

Characterize the track

TTCI/AAR, 1/11/2012, Filename p11

What We Need to Prevent!

TTCI/AAR, 1/11/2012, Filename p12

The performance measures

Vehicle/Track Interaction Safety Limits

No wheel of the equipment shall be permitted

Net Axle L/V

The peak-to-peak acceleration, measured as

The peak-to-peak acceleration, measured as

Truck hunting shall not develop below the

TTCI/AAR, 1/11/2012, Filename p14

Bill, Mazza, and Mike placed strain gages on the rails

TTCI/AAR, 1/11/2012, Filename p15

Example of Wayside Data

These can be compared

TTCI/AAR, 1/11/2012, Filename p16

Vehicle Dynamics Model

Modeling vehicle using NUCARS

TTCI/AAR, 1/11/2012, Filename p17

Includes Track Model

TTCI/AAR, 1/11/2012, Filename p18

Track characterization provides inputs for the computer

Allows us to determine the stiffness of rail connections to the

TTCI/AAR, 1/11/2012, Filename p19

Track Hammer Test

Accelerometers are placed on the rail and on the tie or

TTCI/AAR, 1/11/2012, Filename p20

Slab Track Instrumentation

124m slab in curve

Example of Slab Track Data

TTCI/AAR, 1/11/2012, Filename p22

Vehicle Characterization Test

These tests are only to identify model inputs

gravity height and car body mass moments of inertia

TTCI/AAR, 1/11/2012, Filename p23

Lateral Stiffness Test

Longitudinal Stiffness Test

Vertical Stiffness Test

Carbody Modal Testing

Car is instrumented with enough accelerometers to

TTCI/AAR, 1/11/2012, Filename p25

Bogie Modal Test

TTCI/AAR, 1/11/2012, Filename p26

Truck Load Equalization

The truck must

TTCI/AAR, 1/11/2012, Filename p27

Truck Load Equalization