Shock and Vibration in Rail and other Transport Modes · Broad overview of shock & vibration issues in railroading zVibration issues zTypical mechanical shocks zVarious industry test

Shock and Vibration in Rail and other Transport Modes

William C. Shust

© 2007OBJECTIVE ENGINEERS, INC.

Acknowledgements

AAR / TTCIASF-Keystone, Union PacificFRAChicago Transit AuthorityNorfolk SouthernMonsanto CompanyInternational Truck & Engine


“No, No, No…

That regular rock.

Me need Phillips.”

Bill’s


Broad overview of shock & vibration issues in railroading

Vibration issues Typical mechanical shocksVarious industry test standardsMitigation techniquesBrief comparison of other transport modes to the rail shipping environment.

Technical conclusionsFuture trends for S&V in railroading


SHOCK & VIBRATION ISSUES

• Overall car stability• Component strength• Human comfort• Cargo damage


A closer look at the railcar environment: dynamic excitations

VehicleVehicle

Bogie and Unsprung Mass

Bogie and Unsprung Mass

Irregular running surfaces of wheel and rail[Wheelflats, out-of-round wheels, wheel corrugation,

rail corrugation, dipped welds and joints, shelling, etc]

Irregular running surfaces of wheel and rail[Wheelflats, out-of-round wheels, wheel corrugation,

rail corrugation, dipped welds and joints, shelling, etc]

Track components[Rail bending, railpads, tie bending,

ballast and subgrade]

Track components[Rail bending, railpads, tie bending,

ballast and subgrade]

Wheel/rail noise[Rolling and impact noises from irregularities on

wheel and rail, and squeal from stick-slip vibration]

Wheel/rail noise[Rolling and impact noises from irregularities on

wheel and rail, and squeal from stick-slip vibration]

0 - 20 Hz0 - 20 Hz

0- 500 Hz0- 500 Hz

0 - 1500 Hz0 - 1500 Hz

0 - 1500 Hz0 - 1500 Hz

0 - 5000 Hz0 - 5000 Hz


At the low frequency end: rigid carbody motions

Pitch & bounce, yaw & sway, twist & roll, lateral hunting… all under 4 Hz


60’ Wavelength

Right RailVerticalPosition

Left RailVerticalPosition

0.75” Amplitude

Roll Input for AAR Chapter XI spec.(Left and right track vertical profiles)


0

1

2

3

4

5

0 20 40 60 80Simulated operating speed (mph) over repeated 3/4-inch twist and

roll perturbations

Max

. Pea

k-to

-pea

k C

arbo

dy ro

ll (d

egre

es)

60ft D.C.32ft Flatcar

0.6 Hz resonance -- Carbody roll(constant amplitude cross-level deviations)


Typical Twist & Roll resonance (generic tank car, 18 mph, 39’ cusps)


More low freq. behavior ~ 2 Hz Wheelset Lateral Hunting


At the high frequency end -- potential fatigue of & on circuit boards

• Small mechanical fittings on a high frequency source (100-200 Hz)

• PC Board bending (70-150 Hz)

• Tiny cantilevered components (up to 1500 Hz if small enough)


A few peculiarities related to certain car types

Covered hoppers – high CG, car rock-offTank Cars – tors. stiff, track twist sensitiveConversion of 89’ flats to early car haulersHeavy duty span bolster cars (8+ axles)Lighter weight coal cars


Example project A: Infrequent top chord yielding of coal cars

Suspension bottoming, or component strength?


-5000

-4000

-3000

-2000

-1000

0

1000

0 10 20 30 40 50 60Train Speed (mph)

Min

imum

Str

ain

in

10-S

econ

d B

urst

6000 miles later; several dozen data bursts in 5 general groups

-3500ueDesignLimit


Vertical Bounce Detail ~ 1.9 Hz: Strain, neg(Accel.), Spring Deflection

Time (Secs)4 4.5 5 5.5 6 6.5 7

0.6

0.4

0.2

0

-0.2

-0.4

-0.6

-0.8

-1

-1.2

-1.4

Ver

t CB

Acc

el 1

7361

(Gs)

0.5

0

-0.5

-1

-1.5

-2

Vert

Mot

ion

at A

R 2

6 (In

ches

)

-800

-1000

-1200

-1400

-1600

-1800

-2000

-2200

-2400

-2600

Str

ain

AR

Top

Cho

rd (

u)


Vibration issue: Over time, a little bounce leads to more bounce


Quasi-static event: Rotary Car Dumper Coal Unloading ~ highest strains of test


Example project B: Development of a test spec. for aftermarket hardware

ABCD

TOR nozzle location at end of sand bracketTruck frameElectrical control cabinetCompressor room floor

(Note: Locationsvary slightly withlocomotive type)


1

10

100

1000

1 2 3 4 5 6 7 8 9 10Correlation with Field Environment

Cos

t / C

ompl

exity

-1 .5

-1

-0.5

0

0.5

1

1 .5

0 1 2 3

Lab tests are always a compromise between reality and cost

Over-the-Road Simulation, Whole Vehicle

Sine Sweep or Dwell, Single Axis

Multiple Axes, Shaped Random

Single Axis, Shaped Random Single Axis, Shaped Random

0 500 1000 1500 2000

10-4

10-3

10-2

10-1

100

810F Min. Integrity = 7.6 grms

MIL-STD-810F and IEC61373 Vibration Spectra

Frequency (Hz)

Vibr

atio

n (g

*g/H

z)

810F Rail 0.49 grms

IEC bogie long. 2.0 grms

IEC bogie lat. 3.8 grms

IEC bogie vert. 4.3 grms

Single Shaker, Random Flat PSD


Mid-cost & moderately real: Random Shaker Input With Same Frequency Content As Field

time (secs)232 233 234 235 236 237

10

5

0

-5

-10

-15

Side

Fram

e Ac

cel.

(g)

CD12_072 Y

time (secs)70 71 72 73 74 75

10

5

0

-5

-10

-15

Side

Fram

e Ac

cel.

(g)

CD3_072 Y

time (secs)215 216 217 218 219 220

10

5

0

-5

-10

-15

Side

Fram

e Ac

cel.

(g)

CD6_121 Y

97th Percentile Data Records

Random shaker drive file with similar energy distribution; similar

but more frequent peaks

66

72.080.1250

=

=

fieldRMS

labRMS

gg

5 second segment0

15

10

5

0

-5

-10

-15

Shak

er A

ccel

. (g)

Random Vibration with similar Energy Distribution as SideFrame

(e.g. Time Compression 250X)


Data collection on locomotive type A: Sander Bracket

Tri-axial access. on sander bracket:

at side frame (upper)

at free end (lower)


Data collection on locomotive type B: Sander Bracket


Revenue Service Routesfor Locomotive Vibration Tests


Overall Comparison – PeakAccelerations versus Location

1

10

100

1000

Contro

ller

Front P

lowLu

be R

es.

Side Fram

e

Sande

r Res

ps.

Journ

al Box

Max

. Acc

el. (

g pe

ak-to

-pea

k)


Field Data to Lab Test Conversion Process

Reduce field dataUsing all valid files, compute aggregate (damage-equivalent) side frame RMS accelerationsEnvelope resulting shape of severe PSDs (power spectral densities)

Simplify PSD envelope, amplify to compress shake table time

Select 10-20 breakpoints for PSD shapeScale for 250:1 time compression vs. field

Augment 8-hour shaker period (per axis) with brief segments achieving similar g level extremes as found in field data


Locomotive Bearing Box Raw Vibration

time (secs)327.4 327.5 327.6 327.7 327.8

120

80

40

0

-40

-80

-120

time (secs)327.4 327.5 327.6 327.7 327.8

120

80

40

0

-40

-80

-120

• 251. g peak-to-peak, vertical direction

• Almost all energy around 500 Hz

• Estimated 0.01”double amplitude displacement


Locomotive Bearing Box Same data in freq. domain (PSD)

Block length = ½ second

Bandwidth = 2 Hz

N=144 blocksFrequency (Hz)

0 100 200 300 400 500 600 700 800 900 1000

101

100

10-1

10 -2

10 -3

10 -4

((g)

)

Frequency (Hz)0 100 200 300 400 500 600 700 800 900 1000

101

100

10-1

10 -2

10 -3

10 -4

((g)

)


Most Severe Spectra from 2 loco. types and two service routes

Frequency (Hz)25 50 75 100 125 150 175 200 225 250

10 0

10 -1

10 -2

10 -3

10 -4

10 -5

Acce

l. ((

g^2)

/Hz)

Lateral Spectra (Auto Parts Service)10 0

10 -1

10 -2

10 -3

10 -4

10 -5

10 0

10 -1

10 -2

10 -3

10 -4

10 -5

10 0

10 -1

10 -2

10 -3

10 -4

10 -5

10 0

10 -1

10 -2

10 -3

10 -4

10 -5

10 0

10 -1

10 -2

10 -3

10 -4

10 -5

10 0

10 -1

10 -2

10 -3

10 -4

10 -5

10 0

10 -1

10 -2

10 -3

10 -4

10 -5

10 0

10 -1

10 -2

10 -3

10 -4

10 -5

10 0

10 -1

10 -2

10 -3

10 -4

10 -5

10 0

10 -1

10 -2

10 -3

10 -4

10 -5

10 0

10 -1

10 -2

10 -3

10 -4

10 -5

Frequency (Hz)25 50 75 100 125 150 175 200 225 250

10 0

10 -1

10 -2

10 -3

10 -4

10 -5

Acce

l. ((

g^2)

/Hz)

Lateral Spectra (Coal Service)10 0

10 -1

10 -2

10 -3

10 -4

10 -5

10 0

10 -1

10 -2

10 -3

10 -4

10 -5

10 0

10 -1

10 -2

10 -3

10 -4

10 -5

10 0

10 -1

10 -2

10 -3

10 -4

10 -5

10 0

10 -1

10 -2

10 -3

10 -4

10 -5

10 0

10 -1

10 -2

10 -3

10 -4

10 -5

10 0

10 -1

10 -2

10 -3

10 -4

10 -5

10 0

10 -1

10 -2

10 -3

10 -4

10 -5

10 0

10 -1

10 -2

10 -3

10 -4

10 -5

10 0

10 -1

10 -2

10 -3

10 -4

10 -5

10 0

10 -1

10 -2

10 -3

10 -4

10 -5


PSD Envelope of Many Field Spectraand simplified shake table Breakpoints

200 400 600 800 1000

10-4

10-3

10-2

10-1

100

g*g/

Hz

Freq. (Hz)

Overall lateral ampl. using breakpoints: 1.8 g RMS

200 400 600 800 1000

10-4

10-3

10-2

10-1

100

g*g/

Hz

Freq. (Hz)


200 400 600 800 1000

10-4

10-3

10-2

10-1

100

g*g/

Hz

Freq. (Hz)



Summary of Shake Table Amplitudes (compared to MIL & Euro. specs)

Location 120-day Simulations Brief Field ExtremesCab Vertical 0.26 1.32 0.81Compressor Vertical 0.52 1.31 0.81Plow Longitudinal 0.26 1.44 0.40Sideframe Lateral 1.80 4.32 3.77Unsprung Vertical 10.5 44.5 30.6

MIL-STD-810F Rail Transport

0.49

MIL-STD-810F "Minimum Integrity" 7.7

Proposed Shake Table Vibration (g RMS) IEC 61373

"Long Life"


Finally Add Representative Thermal Cycling (-40 to 130 ºF)

012345

Vibr

atio

n (g

RM

S)

0306090

120150

Time

Tem

pera

ture

LateralDay 1

Long.Day 2

Vert.Day 3

Lat.Day4

Record Vibr. andSample Spray 2-Minute Field Extremes


Proof of test: freezing, shaking & baking a locomotive rail friction-modifier nozzle.

Electro-magnetic shaker with 15000 pounds of force

Heating/cooling source


Collection Grid Pans to check Spray Pattern and Application Amount


Final product: AAR spec. of minimum performance, before fleet installations

Clean nozzle After build-up


0 5 10 15 20 25 3010

-3

10-2

10-1

100

Human Comfort Issues ISO-2631 (empirical weighting of spectra)

Original Signal 0.83 g rms

Energy Spectrum

Weighting Values

Final Area 0.63 g weighted

0 30 Hz

Ver

t. A

ccel

. (g)

0

100

%

g*g/

Hz

Time (sec)PSD

Similar process for noise dBA


TYPICAL SHOCKS

• Car coupling• In-train forces (slack action)• Wheel imperfections• Suspension bottoming out• Rattling pieces


Coal Car Project A -- revisitedZero Speed Yard Data, Unattended (unexplained)

0.2

0

-0.2-1

-1.2

-400-1000-1600-2200

Time (Secs)993600 10 6 1.02*10 6

10

5

0

Vertical Car Body Accel.

Vert. Spring Defl.

Top Chord Strain

Train Speed

1 4 1 5 1 6

-1 6 0 0

-2 0 0 0

-2 4 0 0

-2 8 0 0

-3 2 0 0

-3 6 0 0

-4328 ue for 1/15th sec, then 2 hours at –2000 ue before relaxing


0 1 2 3 4 5 6 7

x 104

-1

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8

1

Accel

Shock events produce accelerations with energy into higher frequencies


0 1 2 3 4 5 6 7

x 104

-0.1

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Strain Gage

But unlike car bounce, the shock accelerations are not scaled duplicates of strains


And thus, a common 4g Design spec. willbe exceeded by high-freq. accel. data


One coupling shock: accel. filtered at 4 freqs., and coupler load cell trace


Similar sampling/filtering questions for various impacts related to passing wheels.

_ _ _

Collection Time(secs)20 40 60 80 100 120

-300

-200

-100

0

100

200

300

Filter_010.sif - [email protected]_1.f.f

Collection Time(secs)20 40 60 80 100 120

-300

-200

-100

0

100

200

300

Original Data Lowpassed at 333Hz=fc

After instead applying 10 Hz Lowpass filter

Revenue Train Sampled at 1000 Hz


Zoom in on only 6 seconds

Original Data 333Hz=fc

Filter_010.sif - [email protected]_1.f.f

Collection Time(secs)60 61 62 63 64 65 66

-300

-200

-100

0

100

200

300

After applying 10 Hz Lowpass filter

_ _ _

Collection Time(secs)60 61 62 63 64 65 66

-300

-200

-100

0

100

200

300


Rail Strain Gage with various lowpass filtering

0%10%20%30%40%50%60%70%80%90%

100%

0 50 100 150 200 250 300 350

Filter cutoff Frequency (Hz)

Perc

ent o

f wid

eban

d va

lue

MaxMinRMS

Max is 95%

Min is 41%, What’s going on?


Answer found by again zooming in on the original data…

Collection Time(secs)60.00 60.25 60.50 60.75 61.00

RailSG4(microstrain)

-300

-200

-100

0

100

200

300

Different frequency content !


A SAMPLING OF INDUSTRY STANDARDS (S&V and related issues)

Car body stability & healthTrack health

Car component health

Cargo stability & healthLongitudinal shock

Human stability & comfort

Locomotive specific issues

FRA 213 Track GeometryAAR Chapter XI, and M-976 specsWheel Impact DetectorsAAR S-4200 ECP brake specpending AAR brake beam workIEC-61373AAR Dam. Prev. Stds.MIL-STD-810ISO-2631 ride comfortAAR on-board rail lubricator vibration spec


Another recent example, AAR S-4200 ECP brake spec

Design to withstand:Vibrations 0.4 g rms 1-150 Hz (with +/-3 g peaks)Half-sine shocks of 10g peak (20 – 50 msec)If on car strength members, local resonances can raise levels to:

15g (100-150 Hz)50g (200-500 Hz)


S & V MITIGATION TECHNIQUES

Limit input energy availableTrack geometry standardsWheel flat limits, wheel impact detectors

Interrupt transmission pathsSpring/elastomer isolators

between car componentsbetween loco. structure & crewtrack to ground

Add damping, stiffness, mass


Limit the dynamic excitations to carsvia geometry or speed limits


Title 49 CFR, Part 213Track Safety Stds. (track geometry specs.)


WILD Detectors: maintain wheels based on revenue track impacts


Transmission and reflections of rail vibration are highly site dependent

from “Transit Noise and Vibration Impact Assessment,” Hanson, Towers, and

Meister, Fed.Transit.Admin. 2006


Earth vibration due to freight trains

50556065707580859095

0 10 20 30 40 50Distance from Track (m)

Gnd

. Vib

. (fr

eigh

t, Ve

l. dB

re 1

e-6

ips)


Noise, Squeal and Corrugation(A human perception problem)

Potential SolutionsGrinding to remove corrugationImprove curving via W/R profileLubricationReduce surface roughness of wheel turning and rail grinding

Structural or surface changes to wheel (block or cut sound transmission path)


Chicago Transit Authority: Wheel screech damper


More costly transit N&V mitigation


Example project C: (rarely) we come across a beneficial use of vibration

Civil engineers brought us “dynamic track stabilization” for after tamping operationsEssentially a combination of

adding stiffness/strength via aggregate material changedelaying the time when small bounces will beget bigger bounces


Ballast Stabilization (Use of vibration to control track settlement)


Track stabilization worth about 10-20 trains of otherwise slow ordered traffic


Thus, the desired vibration lessens chances of this: Lateral Track (Panel) Shift


Another potential beneficial use…

From this campus, Dr. Weaver’s research on rail neutral temperature…Wavelength of high frequency rail vibrations changes with longitudinal stress on the railGoal: Non-destructive determination of rail neutral temperature.


RAIL VIBRATION COMPARED TO

OTHER TRANSIT MODES

(Example projects C & D)


First some

antique video

footage


Moving Cargo, Containers and Operators


Evolved from two projects involving weld failures on ISO tank containers


Unattended Container Shock Record(Does/does not exceed 4g Design Std?)

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1-30

-20

-10

0

10

Logg

ed E

vent

Val

ues

(g)

SN 13392 Event 39, Long.

Original extremum value: 20.4

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1-10

-5

0

5

Filte

red

Val

ues

(g)

Time (sec)

8.6 = extremum after 90 Hz lowpass3.5 = " " 40 Hz "2.5 = " " 30 Hz "1.3 = " " 16 Hz "

-20.4 g raw data?

-2.5g after 30 Hz lowpass filter


“Well then, this is a much lower frequency event, surely it is abusive”

-30.0

-22.5

-15.0

-7.5

0.0

+7.5

+15.0

+22.5

+30.0

0.0 0.1 0.199 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

g

Seconds

g ggX Y Z

Integrating the accel. signal yields a velocity change of 102 mph in 1/3 second. Invalid data!


Vibration for many vehicles

compared

(Bob Fries, N.Cooper-rider 1993)


TECHNICAL OBSERVATIONSThe highway truck environment is generally more harsh on cargo & humans than rail. Shipboard environment is generally less harsh than rail. Ship environment is more like a factory floor.Low frequencies (<15 Hz) strains ~ accels.

Less true as frequency content increasesNot true for raw shock data

Field data is incomparable to any other criteria or test without knowing the sampling & filteringGround borne vibrations—rules of thumb

Twice the train speed ~ twice the vibrationTwice the distance from track ~ half the vibr.


Observed & periodic opportunity for engineering confusion

CarModel

+VA

Load Measuring Axle

+Y

+LB+LA +T

+VB

+Z+X

+T


Another test risk: small populations

Four strain gages at “same” locations on cantilever beam –next to fixed end.

(22” long, 3 ½” wide,

1/16” thick)


Strain variations, even with careful attention to similar gage placement

Time (secs)25.3 25.4 25.5 25.6 25.7

300250200150100500

-50-100-150-200-250

gage

3 (m

icro

stra

in)

4 gages at 1st bending freq. ~ 3 Hz

300250200150100500-50-100-150-200-250

gage

4 (m

icro

stra

in)

13% Variation in

Peak-to-peak

g3 = 543ue

g4 = 502ue

g7 = 477ue

g8 = 521ue


What about variations on a railcar?Structural member strains, Left vs. Right

time(secs)245 250 255

-200

-100

0

100

200

300

50% difference in amplitude.


Final reminder: accelerations vs. strains

CB_VT_SQBURST.SIF Channel 4

-8

-6

-4

-2

0

2

4

CB_VT_SQBURST.SIF - Channel 5

-200

-100

0

100

200

300

CB_VT_SQBURST.SIF - Channel 6

time(secs)245 250 255 260 265

-400

-200

0

200

400

600

Low freq. correlation, but not at high freqs.

Lt.S

train

R

t. St

rain

A

ccel

.


Wrap-up: industry trends relating to S&V

“Perfect engineering specification”would fail 100.0% of latent bad designs or processes,while passing 100.0% of the good.

Actual specifications tended to promote certain dimensional tolerances, minimum design strengths, or perfunctory initial performance… not perfect.For about a decade, advances in technology have gradually begun to provide operational feedback, largely due to monitoring shock and/or vibration (or close cousins).


Implementation risksThis monitoring of S&V (or close cousins) can be

infinitely more complex than the old visual inspection. Thus, there is much concern about the false positives or negatives.

Fortunately, the industry is gradually moving ahead:Wheel Impact MonitorsAcoustic Bearing SignaturesExcessive Railcar BouncePeak Wheel ForcesYard Coupling Shocks, or desired lack thereofetc.


Two opinions about S&V work in railroading

PLUS: In an otherwise mature industry, further exploiting shock and vibration to assist business decisions holds great opportunity for rail engineers.MINUS: Fostering adoption of these technologies requires great patience and persistence. (Railroads, suppliers, shippers, car owners, and the FRA have a complicated and symbiotic relationship. In some cases, these newer and more effective specifications redistribute costs. This may be unpopular.)


Example project E: 1994 brake beam tests. Adoption of standards is still in-process


Give an engineer just a simple request…

© 2007OBJECTIVE ENGINEERS, INC.And the engineer will aim to please…



Engineering Information is not the same thing as test (or analysis) data

Like Dilbert, we need to provide information, not just data, “Here is $7.14, you owe

me five & a quarter.”

Shock and Vibration in Rail and other Transport Modes · Broad overview of shock & vibration issues in railroading zVibration issues zTypical mechanical shocks zVarious industry test

Documents