‘ r. 'I‘M V’.‘
‘5‘ A 5%: ;$"“'"“’§’1'::' .
I... 3% ..

I .11}. . . . .:'

I)"
um ét‘,_
“‘4"-

Q
. 7,
if??? -
Ida's
Gfiifi
Ifi‘r ..

"‘2‘“ u

45‘
~31: .1135 if;

4
I v 1-..:.“

:c" “3”!"

¢ . ‘_-

)IJ’E‘S’ :1 7
Izam
" 2,.

'I -,I %er~
.. ‘va

JEJ‘.”I

”€751?

{:5

i’:

' if"; a.” I '

".312”, ‘4'}, 1”,}???

”’I‘Ili “1:” I923?"

fill," I}???
4%,“? I:

{2%. gr! {96:1}:

‘-. . I
{555:}; { {TV/"I
‘. I

. I

’ng'" - ‘f’ lIv'I
. IquIul fij": 3:". ",“1h/
. 4/11.,33.’ .
,‘flqffff’ ..
f“.
I'

'0

‘ I
1",).4)?
5,31

II,

1);, '7'IJ.{',"/J r u) ”/12"!

'.'1 HIV}
“'0', I fulfill; .,
1!. If. -_

i5 47;; 3"

I
[I

Jim/t U I

['I. I'll:‘ I'I

WW

3.. ..a ‘g‘h'n'V‘
ya‘lgwan‘ ,‘
. *I’Q;a£wfiwn
' » ”Mn.- -
1}»

.
-
l."

.91."; - 1.,
. W .
113" ‘1‘; .,:Nfié)i§¥"‘:h TI
mi, 4;.- :2?sz

‘4
R‘ulg’i‘.
'V 1‘)

J.’
a; -
,
3‘:.:“

.-..

#4:.

[-I ‘13:"! ff”? 3-] ~

£ 97“ 519F335“);
I W“

.~ .\ .’
35 If?

3"
,5,

rm, f 3%,: ‘
' f] 'T I" II"
(II/,Jir I’M/.5132,“

'l
’5- if; I)"; 2.1,: ‘14.;

‘4 n.
J
"I.

40/. :{’r'-.

t‘
1“.

\'

(111.32%: 4V3,”

5:

e ‘1:
I §~'m\ :¢_.:\..

'-/.;.’
”:25;
3.I1"-¢"Ir"""‘

. ‘o/
I” ”fluid:

 

2:,
5‘ . ’lh’
.fi‘uPIE' 'J’f'IJ.)I'I

r
r‘I‘J‘ :l/,( (‘I/II’gffJT’t

uII‘ijIAkfI 2,3,3:

"If, '[r

3.
1,5“
91.41

: #1

,L.

17:.

I)”
a

' I
I 'h'
l/l;"’. I“ . I' . u I‘;/: ’fl/i-J'
’21..” . ML ”0 ,‘Ifits‘t‘. V" ’4 ‘54:;
1'.

1.01:.

. film/I'm .
40:...

:3
"13»?

(2,-

£15,} ' r. . I;-

”$59”, ’ I’Iplr /

:Im:

fffh
A” .'

,‘I’ I.
'27]. I'I'W'f’
U I

-' L‘Ilu’r‘ll t
W41? ‘ I
I. 1-7: .ul'ncl.’ :’
.,;r; , L
‘3 .o/ . é,”
r w w.» III/2,.
9' griffin. waivf'

In
" / Vv‘ ,-;
II'. ‘I 4/ '05"!

9’7"” (1' “/3? "a? E .
I’Jfi 77: I' J;
I ’ 1:35 MI- («(3.5 J

”I,” ”{1}”;
.._ 113-};/'
m1! 7" ‘

Yb,

I l-thvilv’31AQé/I

34“}

6‘; 'I
I'll 41/}

“1&1, ”IQ/,7, hf

)/)JJ It?

trill

It.”

Hat”, .. _
a” fun; WK: “3*“
‘5”; 353;;54 .
' - ”1"3‘33‘4'415335
154$ L-w...:‘.x
~36 :'I$'='-~ '...‘¢.'
4 i=3};

w
In“?

I .'
.. :5I.'II'-,_‘
jlln,‘If '1

‘-

3;

j» v'l';
‘3 ’J‘v‘z‘rv
Two-33¢. Egg-5
J15~

MR.

u

t J-{L ,—_A
“*r-In‘
1-,“: 7~)I-
”-4.1,

$33?“

‘1.
V‘-

‘- ’IJ'fl

. g.‘,.,'lv~ I I‘,
' .

1,“. ')";‘?'l".'" :‘III. .

I M")

i .~' .

a ,

""1! 1"

11/13: :‘I 11

{'IU/

"lg/"4‘0"

.1; =11:
5337?“;

. 9
65"; z: “W
‘wr, 7.

’I

“lg-

1-”

f u.
vT‘i‘fi‘m 'A
:3 ~;

If], '31, .J 'I
{’2'hr5ffi/v!
if“??? W1

'9".

~321~.3»=f*1.
Ln~_ .3“. _ _
‘71-.

v
Q
~

n

. KIA"; ~

 

 

 

~ llllllll'llll(lllllllllllllll Ill/Mill L/

LIBRARY ‘
Michigan Statel
University

 

 

 

 

 

This is to certify that the

dissertation entitled

SHOCK AND VIBRATION IN THE RAIL ENVIRONMENT:
TRAILERr—ON—FLAT CAR AND CONTAINER—IN—WELL CAR

presented by

Daniel Lee Goodwin

has been accepted towards fulfillment
of the requirements for

DOCTOR OF PHI LOSOPHY degree in A9 ricu ltura 1
Engineering
Technology

ajor prof t

Date 21/25/58

”(III-n. A!" n’ A ‘ :— 1m, - . . . 0.12771

 

 

MSU

LIBRARIES
.—;—.

 

 

 

RETURNING MATERIALS:
Place in book drop to
remove this checkout from
your record. Elfl§§_will
be charged if book is
returned after the date
stamped below.

 

 

mm 0 9 “1m

 

sflalfifio

 

 

 

SHOCK AND VIBRATION IN THE RAIL ENVIRONMENT:

TRAILER-ON-FLAT CAR AND CONTAINER-IN-WELL CAR

BY

Daniel Lee Goodwin

A DISSERTATION
. Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Agricultural Engineering Technology
Department of Agricultural Engineering

1988

ABSTRACT

SHOCK AND VIBRATION IN THE RAIL ENVIRONMENT:
TRAILER-ON-FLAT CAR AND CONTAINER-IN—WELL CAR

BY

Daniel Lee Goodwin

The purpose of this study was to measure the shock and vibra-
tion characteristics of the intermodal rail shipping environ-
ment. This included trailer—on—flatcar and container-in-well
car. Data obtained was to form the basis for a laboratory

simulation which could be used as a preshipment test for pro-

duct and package systems.

A test method was developed to measure the dynamics of the
two rail environments. Shocks and vibrations were recorded,
in all three axes, for an intermodal distribution cycle from
Rochester, NY to Los Angeles, CA. Unattended instrumentation
gathered the rail data. Vibration input was processed in a
spectrum analyzer and represented as Power Spectral Density
(PSD) profiles. The recorded shocks were classified by range

and distribution for each axis.

Daniel Lee Goodwin
Power Density maximums were found to be 1.5 x 10-4 gz/Hz,
for TOFC, and occurred in the lateral axis. For CIWC the
maximum was found in the vertical axis and had a value of

S gZ/Hz. Since transient shocks increased the vibra-

7x10-
tion responses of the vehicles, the test was to include both

dynamic inputs.

For TOFC measurements, 76% of the shocks were in the 2 to 5 9
range, in the vertical axis forward location. The vertical
rear experienced 50% of the inputs between 4 and 6 9'3. In
the longitudinal direction 75% of the impacts were in the O

to 2 9 range. Fifty percent of the lateral shocks were seen
from 11 to 12 9'8. In the CIWC format, 83% of the inputs were
from 2 to 3 g's at the vertical forward location, and 75% were
in the 2 to 7 9 range for the vertical rear. Thirteen percent
of the longitudinal shocks fell between 1 and 4 g's, and 20%
from 7 to 10 9'5. Forty-four percent were in the 15 to 20 g

range and 79% of all lateral shocks were measured from 6 to 13

g's.

A digital control system was used to combine shocks with vibra-
tion on an electrohydraulic shaker in the laboratory. The test

was used as a means of evaluating pallet load stability.

Copyright by
DANI EL LEE GOODWI N

1988

iv

To Latty, Correy and Eric

ACKNOWLEDGEMENTS

Sincere thanks is extended to Dr. Larry Segerlind for
serving as Major Professor throughout this long endeavor.
Dr. Gary Burgess has offered much constructive criticism in
the development and organization of the research activity.
His assistance is greatly appreciated. Thanks, also, to
Dr. Tee Downes and Dr. Ajit Srivastava for their patience

and support._

Mr. Ronald Holland has been an invaluable friend and ally
as the various obstacles and set-backs of data collection
were met and overcome. Gratitude is also expressed to Dr.

Russel Cain for his insight and suggestions for dealing

with the frustrations of original research.

vi

TABLE OF CONTENTS

LIST OF TABLES...........................................iX
LIST OF FIGURES.......................................... X
CHAPTER ONE INTRODUCTION...............................
1.1 OBJECTIVE...........................................
CHAPTER TWO LITERATURE REVIEW........................... 7

2.1 VIBRATION IN COMMON CARRIER VEHICLES. 19

2.1.1 Over-The-Road Truck............................... 19
2.1.2 Rail Cars......................................... 24
2.2 Shock Inputs........................................ 28
CHAPTER THREE TEST METHODOLOGY......................... 35
3.1 VIBRATION EQUIPMENT................................. 38
3.2 SHOCK INSTRUMENTATION............................... 41
3.3 TEST PLAN........................................... 44
CHAPTER FOUR DATA ANALYSIS............................. 51
4.1 RANDOM VIBRATION DATA............................... 51
4.2 TRANSIENT SHOCK DATA................................ 51
4.3 VIBRATION RESULTS................................... 53
4.4 SHOCK RESULTS....................................... 57
CHAPTER FIVE LABORATORY TEST PROFILE................... 70
CHAPTER SIX OBSERVATIONS............................... 77
CHAPTER SEVEN SUMMARY AND CONCLUSIONS.................. 82

CHAPTER EIGHT RECOMMENDATIONS FOR FURTHER STUDY....... 87
vii

APPENDIXA FRAGILITYOIOOOOOOOOOOOO0.0.0.000...

0....

APPENDIX B DETERMINATION OF PEAK ACCELERATIONS.....

APPENDIX C MODAL ANALYSIS OF A TRUCK TRAILER FLOOR.

C.1 Test Set-up................
C.1.1 Instrumentation Hardware.
0.1.2 Instrumentation Software.
C.2 Measurement Methodology....
C.3 Estimate Modal Parameters..
0.4 Sort Modal Data............
0.5 Display Mode Shapes........
APPENDIX D PSD PROFILES:
APPENDIX E SHOCK DISTRIBUTION:

BIBLIOGRAPHY.......

TOFC AND

TOFC AND CIWC.

CIWC.

GENERAL REFERENCES.O00...OOOOOOOOOOOOOOOOOOOOO.

viii

89

. 92

.103
.103
.105
.107
.109

.112

LIST OF TABLES

3-1 a Train Status 0 o o o o o o o o o o o o o a o o
5-1 0 Peak coordinates o o o o o o o o o o o o 0
C-1. Calibration Settings.........

0-20 ReSOnanceS......o............

C-3. Modal Frequency and Residue (Sample)

E-lO Test ReportOOOOOOOOOOO00.00.000.00...

E-ZO ShOCk Data SheetOOOOOO0000.00.00.00.

ix

Page
...50
...75
..101
..101
..104
..107
..108

Figure

2.10.
2-11 0
2-120

2-130

LIST OF FIGURES

Page

Straddle crane.................................... 8
Container stacking................................ 9
Articulated coupling..............................11
Five platform format for well cars................12
Side bearing minimizes lateral movement...........l4
Comparison of well car and flatcar............ ..15
Frequency spectra for truck (vertical)............20
Truck frequency spectra - summary of PSD data.....22
Vibration envelope curve - truck summary..........23
Track comparisons................................25

Railcar frequency spectra - summary of PSD datan.29

Vertical acceleration envelope - railroad.......30

Frequency spectra for railroad (vertical)........31

Over-the-road truck..............................36
Pinwheel pallet loading configuration............39
Instrumentation placement and protection.........39
Tape recorder timing module......................AO
Vibration recorder power circuit.................42
Instrumentation schematic........................43
Trailer-on-flat car..............................45
Container-in-well car....................... ....46

Figure Page
3-9. Rail route map....................................47
3-10. TOFC/CIWC travel times...........................48
4-1. Sample data showing noise floor...................52
4-2. Bump recorder data tape...........................S4
4-3. TOFC, Rochester to Chicago - vibration............55
4-4. CIWC. Chicago to Los Angeles - vibration..........56
4-5. TOFC -- shock distribution, vertical (front)......59
4-6. TOFC -- shock distribution, vertical (rear).......60
4-7. TOFC -- shock distribution, longitudinal..........61
4-8. TOFC -- shock distribution, lateral...............62
4-9. TOFC -- shock distribution, composite.............63
4-10. CIWC -- shock distribution, vertical (front).....64
4-11. CIWC -- shock distribution, vertical (rear)......65
4-12. CIWC -- shock distribution, longitudinal.........66
4-13. CIWC -- shock distribution, 1ateral..............67
4-14. CIWC -- shock distribution, composite 0-9 g’s....68
4-15. CIWC -- shock distribution, composite 9-25 g’s...69
5-1. Random vibration input............................71
5-2. Electrohydraulic vibration system.................72
5-3. Peak selection....................................74
5-4. TOFC half-sine shock..............................75
A-l. Fragility curve.................................. 39
A-2. Input shock and cushioned response............... 90
A-3. Spring-mass system -- product and component...... 90
A-4. Amplification of input shock..................... 91

8-1. Normal distribution of acceleration levels....... 92

Xi

Figure Page
8-2. Power Spectral Density plot...................... 93
8-3. Peak acceleration range.......................... 94
C-1. Trailer floor mapping grid....................... 96
C-2. Grid orientation................................. 97
C-3. Instrumentation schematic........................ 98
C-4. Sample measurement...............................100
C-5. Combining SDOF and MDOF methods..................102
D-l. PSD plots-TOFC...................................105

D-20 PSDoverlay.0.....000.....00.00000000000000000000106

xii

CHAPTER ONE

INTRODUCTION

Deregulation of the railroads in the United States, brought
about by the Staggers Act, 1980, has led to an increase in
their share of the common carrier market. Grimm and Smith
(1986) have noted increased service quality and management
performance, since passage of the act. Two modes of rail
shipment have shown significant profitability. These modes
are trailer-on-flat car (TOFC), and container-in-well car

(CIWC).

Advantages to the firms shipping via these modes include:
the elimination of car humping, the most severe load
handling hazard in the rail environment; products shipped
in trailer-load or container-load quantities will not go
through interim loading and unloading operations; and
line-haul trains give express service which can be traced

through the distribution channel on a mile-by-mile basis.

Flint (1986) observed that rail companies have utilized
market pricing, decreased the number of employees, and
computerized their operations in their move to recapture

market share from the trucking firms. Currently railroads

1

2

carry 12% of the nation’s goods, which is down from the
1962 level of 20%. However, market share for rail is
increasing in some areas and is already strong in some
others. Trains carry a large share of some product
categories: 67% of coal, grain and automobiles; 50% of

household appliances; and 50% of food.

Immediately after World War II, 1.4 million people operated
the railroads in the United States. In 1986, the same
industry carried 35% more tonnage, with only 300,000
workers. Through huge capital investments, more sensible
work rules, mergers, and abandonment of unprofitable track,

one worker today does the work that once took six to do.

Wesselman (1986) states that productivity has also
increased through the introduction of new rail technology.
Double-stacking containers in well cars offers a
significant economy of scale. Twice the cargo is carried
at only 60% of the fuel cost of equivalent TOFC shipments.
Even factoring in non-line-haul cost components, such as
door-to-door service, leaves a 20 to 25% savings to apply

to profits.

Flint (1986) demonstrates the profitability for rail
carriers through the comparison of the Dow Jones Industrial
Average, which has gone up 80% over the last decade, to the

Standard and Poor’s Index of railroad stock which is up

3
200% for the same period. In the same study, Mr. Drew
Lewis, CEO for the Union Pacific Railroad states, "The
break-even point for CIWC is very high, and the
profitability after break-even is so inelastic, it’s almost
straight up. ..., a 25% increase in volume would more than

double our profits."

Computerization has allowed trailer/container volume to be
used more efficiently. Flint (1986) cites the example of
the Chesapeake Seaboard Express (CSX) which began operating
the Orange Blossom Special in 1982. This is a fast run
from Florida to New England to try and win northbound fruit
and vegetable traffic from the trucking lines. The Special
now holds 8 to 9% of the market and is near break-even.
Effective volume control as well as competitive pricing has
started to increase railroad market share across the
country. The Union Pacific Railroad may change freight

rates for fruits and vegetables three or four times a day.

Wesselman (1986) compares the various rail and trucking
modes and ranks the price advantage of each.
Containerization, which includes CIWC, offers the best rate
of return for shipments over 1,000 miles. TOFC is most
cost effective for distances of 500 to 1,000 miles.

Over-the-road truck fares best for distances less than 500

miles.

4

Manufacturers have found the value of goods contained in a
single trailer/container has increased significantly in
recent years. Preshipment testing of products, packages
and their components, and product/package systems has
become common practice throughout most of the manufacturing
sector. Through these tests, companies can identify design
flaws in products or packages and take appropriate steps to
correct them. Test levels and durations are often based on
characteristics of the shipping environment. It is also

important to know the basic dynamics of product and/or

package systems.

Kane (1986) states that intermodal rail shipping methods
are relatively new. TOFC has been in common use since the
1960’s. CIWC was pioneered by Sea-Land Services in the
1970’s. Since there is limited data on these environments,
manufacturers have been interested in learning the nature
and level of the shock and vibration dynamics. As will be
discussed in Chapter Two, few studies have been published
on TOFC or CIWC formats. The data which has been published
is primarily for short runs of 200 to 500 miles, and was
gathered on experimental trains that do not reflect actual
use conditions. Profitability of the aforementioned rail

modes is based on shipping distances of a; least 500 miles.

Kane (1986) finds double-stack trains (CIWC) offer more

versatility for load carrying formats. They can carry 20

5
and 40-foot containers, each 8 ft. 6 in. high, as well as
the newer 45-foot units, which are 9 ft. 6 in. high. Labor
inputs are significantly reduced through double-stacking
practices. If a double stack train hauls 200 containers
across the country in five days, only four or five workers
are needed. To move the same volume by truck, in the same

time period, would require 400 drivers.

Considering the fact that there have been technological
innovations in the physical characteristics of the
railways, railcars, and handling systems, TOFC and CIWC
shipping modes should be evaluated. Advances have also
been made in data gathering and analyzing equipment. Such
advances have broadened the capabilities of laboratory test

equipment to simulate shipping environment dynamics.

1.1 OBJECTIVE
The objective of this study was to measure the shock events
and vibration inputs to a 40-foot container filled with
43,000 lbs. of chemicals and photographic materials. This
container traveled from Rochester, NY to Chicago, IL on a
trailer chassis, as part of a TOFC shipment. In Chicago,
the container was transferred to a double-stack well car
and transported to Los Angeles, CA. The data collection
procedures were to utilize unattended instrumentation
techniques. Both shock and vibration events were to be

measured in all three axes. The shock and vibration data

6

was to be used to create a laboratory test procedure which

would simulate the intermodal rail shipping environment.

CHAPTER TWO

LITERATURE REVIEW

There are a number of factors that must be considered with

respect to the need for studying the TOFC and CIWC

distribution environments. Review of published
environmental studies revealed little data for the
double-stacked container in a well car format. Based on

information discussed previously on the profitability of
this mode of shipment, it is clear its usage is likely to
increase. Trains utilizing either TOFC or CIWC modes are
not coupled by ramming one car into another. Some TOFC
shipments are made up in the rail yard by straddle cranes,
which lift the trailers and place them on flatcars that are
already coupled. Others utilize ramps to back trailers
onto flatcar platforms. CIWC units are also already
coupled when devices such as Santa Fe’s port packers,

double-stack the containers. See Figures 2-1 and 2-2.

Shippers find there will be no interim handling of cargo en
route. In line-haul transport, cars will not be added or
removed from the train, nor will their lading be unloaded
and placed in other cars. Since the routes are determined

in advance, it is possible to track the cargo through the
7

 

 

Figure 2-1. Straddle crane.

 

 

Figure 2—2. Container stacking.

10

entire distribution system. This is an advantage for
determining the locations of various dynamic inputs which

are experienced during the journey.

Published studies on the TOFC environment are, for the most
part, measurements of experimental trains. The study by
Kenworthy (1979), for the Department of Transportion, is an
example. Such tests involve a locomotive, test cars,
buffer cars and an instrumentation car. In addition, the

runs are relatively short distances of 200 to 500 miles.

Well car design offers a means to reduce some of the
dynamics experienced in conventional flatcar formats.
Longitudinal slack action is reduced through use of an
articulated coupling. This is shown in Figure 2-3. Figure
2-4 shows the five platform configuration for these cars.
Each end of the unit has a standard coupling, but the
intermediate couplings are fashioned with pins and provide
a fixed, rigid connection. This technique minimizes the
"run-out" and "run-in" as the train speeds up, slows down
or goes up and down hills. The benefits of the new well
car designs were described in the July 14, 1986 issue of
Pacific Shipper, in an article entitled, "APC Dramatizes

Stack Train’s Smooth Ride by Carrying Table-for-Two With

China."

The above article also found that horizontal stability is

enhanced through the use of side bearings in the truck

11

.mcwamcoo coumHsowuud .muw ousmwm

 

I a J

 

 

-mcwamaoo

to

:E

 

HMO Ham3 HMO HH03

 

 

 

12

.mumo Haoz mom umEHOM EuomumHmlm>ah .cnm ousmwm

mcwamsoo xomam cumccmum

 

 

magi“. 40H. 04. area». Po Po 00 Q?
W

mmcfiamsoo nonmaaufluu<

13

assemblies. Vertical vibrations are minimized by reducing

the distance between truck centers on each platform, as

shown in Figure 2-5.

In addition to the above, the well car design adds
stability and provides a lower center of gravity, as can be
seen in Figure 2-6. This reduces side-sway. APC has
experienced significantly reduced damage levels in the well
cars. Standard flatcars regularly encounter vertical
accelerations of 1.7 g’s and longitudinal forces of 2 g’s.
The well car design keeps both levels below 0.2 g’s. These

facts are reported in the American President Company study

of July, 1987.

Shippers who rely on preshipment testing of products and
product/package systems attempt to anticipate or recreate
damage seen in the field. In such cases, the worst
possible hazards are used in determining test levels. The
basis for many of the common test procedures is found
through measurement of field parameters. Since the early
1970’s, several procedures for random vibration testing
have relied on Power Spectral Density profiles to identify
the energy at various frequencies. Hanes (1983) measured
the environmental dynamics for both over-the-road truck and
padded van, and developed random vibration tests for IBM,
that replaced existing sine-dwell methods. However, there

is still some inconsistency regarding the methods for

reporting vibration data.

14

CENTERPLATE

  
   

SIDE BEARING

MAIN
SUSPENSION
SPRINGS

Figure 2-5. Side bearing minimizes lateral movement.

15

 

.Hmoumam can umo Hao3 mo camfiummaoo

umuumam Hmcowuco>cou umu Hao3

.m-~ musmam

 

 

 

NJ) N <>J

 

 

 

 

 

\, /

 

 

 

 

 

macawmucou

16

Random vibration data is converted to Power Spectral
Density (PSD) profiles through the use of spectrum
analyzers, which perform Fast Fourier Transform operations

on the raw information.

The equation for determining Power Density is:

H
ll
2

(RMS)?L /N

H
II
0

PD gz/Hz (2.1)

 

BW

PD = Power Density (gZ/Hz)

RMS = Root Mean Square acceleration (g's)

N = Number of data samples averaged

BW = Bandwidth of the filter (Hz)
(usually normalized to 1 Hz).

See Appendix B for a complete discussion. When this method

of data processing is completed, the PSD profiles become

the test profile. However, transient shocks occurring in
the environment are averaged out through the data
manipulation. It will be shown that transients,
superimposed on the random vibration profile, add

significantly to the dynamics of the system response.

The use of digital control vibration systems has made it
possible to program these shocks back into the random
vibration input. Also, the advent of electrohydraulic

shaker systems, with multiple degrees of freedom, makes it

17

possible to input forces from all three axes,
simultaneously.
Accuracy of data collection has improved with the

capability of measuring vibration input down to nearly DC
level, or 0 Hz. Low-end frequencies are important to
consider in developing test profiles, due to the large
displacements that occur in these ranges, as a result of
input from vehicle suspension systems. This was reported

by Schlue and Phelps (1968).

The common carrier shipping environment can be defined as
including all conditions likely to be encountered by
products moving from point of manufacture, or harvest, to
the final consumer. Various vehicular transportation
modes, handling, and storage activities are associated with

this process.

Previous studies have reported data and descriptions of the
environments in a variety of different formats. This
variation is due to the types of instrumentation available,
data reduction systems and the type of end-use concerned.
The following discussion will address the nature of test
procedures used for collecting data and the levels of the

dynamics encountered.

Ostrem and Godshall (1979) reported an overview of the

common carrier shipping environment. The characteristics

18

of the vibration environment have been studied in greater

detail than any other shipping hazard. Reasons for this

include: shippers have little control over the operation
of commercial vehicles and the resulting vibrations
transmitted to the lading; vibration measurement requires

sophisticated equipment so testing is usually precise and
detailed; and access is often not available to the

carriers’ vehicles.

Forms of reporting the data include: Power Spectral
Density (PSD) plots; Root Mean Square (RMS)
Accelerations; RMS versus Frequency; Peak Acceleration
versus Frequency; and Distribution of peaks within given
frequency ranges. Acceleration levels reported are
dependent on the bandwidths of the filters used for data
analysis. These filters separate the complex signals into
discrete frequency bands. Ostrem and Godshall (1979)
indicated that current technology relies on constant
bandwidth filters used in compression real-time analyzers
(RTA) and Fast Fourier Transform (FFT) analyzers. Both
analyzers produce the same results. Samples of
amplitude-time histories are analyzed to produce
amplitude-frequency spectra at a rate dependent on the
frequency resolution needed. The various spectra are then
combined to generate an averaged spectrum. Current
analysis typically relies on l-hertz bandwidths and

presents data in the form of PSD levels in units of

19

g2 /Hz. This is considered to be the only format suitable
for describing random vibration. It is used as both a test

specification and an analytical parameter.

2.1 VIBRATION IN COMMON CARRIER VEHICLES
The following section discusses the results of vibration
studies on trucks and railcars. Both have implications for

TOFC and CIWC investigations.

2.1.1 Over-The-Road Truck

Figure 2-7 shows a typical series of vibration profiles for

over-the-road trucks, plotting peak acceleration against
frequency. Such data indicates that probabilities can be
used to compare the levels of severity for a given
vehicle. The advantage of the statistical format is in

identifying the distribution of acceleration below the

peaks. However, a large data base is needed to produce

that summary.

Schlue (1966) reports finding peak PSD levels at 2, 10, 11,
and 50 Hz in truck trailers. The low-end frequencies were
determined to reflect input from the suspension system.
The 50-hertz peak was identified as a structural response.

Such structural responses were found to contain little

energy.

Kusza and Sharpe (1973) found the vibration levels at the

vertical rear location, over the rear wheels, to be the

.Aamofluuo>v xosuu Mom muuoomm mocmsvoum .nlm madman

a... ..l‘ >uzu30u¢u

 

 

“569 095 09 9 Q.
4444l14 4 1 144444 4 4 4 —4444 4 4 4 4 «4444 4 4 1 4 «44444 4 1 4 .q
A
L
L
L
n m
T. L... n
. .. 806. n
.. .. 1.0.». U
. . 804. L v
.. .. $0.00 41’ \x/ \\’ ii.
241.» mama inn. , 4 ( ( ’ L W
x ,
u, ..\
I ..
\ ’ \I,’/ A "O
0
\\ < .
x48 \ H
.. '
u
A
L
L
_ _ r _ u
.o.

 

 

huoumuonmq muuscoum umouom "mounom

21

most severe. Mid-trailer points were the least active, and
the forward area, over the king-pin, was in-between.
Lateral acceleration motion was half that of vertical
input. Speed increased the amplitude of vibration, but did

not change the frequency profile.

Silvers and Caruso (1976) report that the worst case for
shipping is to place cargo directly over the rear wheels of
a lightly loaded trailer. With regard to road type, high

speed highways generated the highest amplitude levels.

In Figure 2-8, the PSD profiles for the above three studies

are compared. The levels of peak acceleration are roughly
equal, but differ as to the related frequencies. They show
the variation in vehicle and load combinations. A

conservative test profile could be defined by enveloping
the peaks in a straight line giving a constant PSD. This
is shown in Figure 2-9. The peaks and notches do exist,
however, and should be included in the test profiles since

it could potentially reduce overpackaging.

Random test procedures, such as ASTM D-4728 (1987),
indicate that PSD plots are an effective format for
carrying out the data analysis and test design. However,
package-related applications are better served by
descriptions of acceleration and frequency levels. This is

because transmissability characteristics of cushioning

22

.mumc 9mm mo unmEESm I muuoomm hocosvmum xosue .mim musmflm

4: ll. >uzu30u¢u
950. g.

 

  
  

 

    

  

 

 

 

 

 

 

tom 2 9. ..o
.214. 4 12:4. :3 1::— _::_ _ 1.22:4 _ :o.
2.. u
. ...mm ...m m l
...m. ...x m t
. n u 1 L w
..11. u i. A
_ .. . 3
W” m \ m H
II ..c II... I s
.7. ~ no.“
2. 3
1 L n
“ § L W
l 1 a
‘ 1.1.. in. u
.. l .h
[Qua-AA
t ~
......... Edfifi .LP
llllggfim 1 m
9256 1
r
— — — .oO.

 

 

 

Shoumuonmq muoscoum umouom "mousom

23

go.

.>HmEE5m xosuu I m>uso omo~o>co cowumunfl>

4: its >UZU30u¢u

.mum musmflm

 

 

 

Rd. .6. o. o; to
4444 4 1:2.— 7 12:14 4 1:2. 4 4 142—4. 4 .00

L a
il a
.l 3
l 1
M . 3
u
o
I. N
\ U 9
l ‘0
fil LuQ. m
d
3
I. '
x

1

L

n

,r _ e F to.

 

 

huoumnonmq nuoscoum umouom “mousom

24

materials are defined with respect to the natural frequency

of the cushion under load.

Cargo placement has been shown to influence the measurement
of trailer floor dynamics. This is particularly noticeable
with light loads of loose cargo and cargo in resonance. It
has been shown that cargo may not respond to some inputs
due to natural damping conditions. Silvers and Caruso
(1976) address this issue for light loads by removing the

cargo and reporting the severest environment as the empty

truck. This eliminates the effects of bouncing loose
cargo.
Ostrem and Godshall (1979) discuss some of the

inconsistencies in the reporting of resonant frequencies
found in trucks. They cite the most significant differences
as occurring in the frequency range above 50-hertz. The
most likely cause for this is poorly maintained or

functioning suspension systems.

2.1.2 Rail Cars
Ostrem and Godshall (1979) observed that past studies on
railcars focussed much attention on the periodic input from
rail joints. These joints occurred every 19.5 feet,
alternately on both left and right rails, because the
standard rails were 39 feet long (Figure 2-10). The car’s

geometry generates periodic input also, with respect to the

25

Joint

/

fl
—_

 

Track

 

 

/
>1
(4 3...... 44

Standard Half-Staggered Jointed Tradk

 

Track

/
\

 

Continuous Welded Track is Fused in 1/4-Mile Sections

Figure 2-10. Track‘compariSons.

26

joints. The overall axle wheelbase distance, the interior
wheelbase distance, the truck centerplate distance, and the
truck axle wheelbase all determine the periodic inputs to
the car. Inputs may also come from eccentric wheels or

flat spots on wheels.

Large responses occur while any of the forcing frequencies
are the same as those within the railcar system. These
include low-speed rolling or rocking and the effects of
wheel lift-off. Frequencies causing the above are

typically below 20 Hz for train speeds up to 75 mph.

Continuously welded track has eliminated periodic input
from rail joints, but the unevenness and roughness of the
rail system still cause significant vibrations to occur.
Responses of the railcar to irregularities on welded track
are at system natural frequencies of the car truck and
track. Vertical inputs also come from rail intersections
and road crossings in the form of transient loadings.
These will cause the car to vibrate at its natural

frequencies.

A study by Sharpe (1972) indicated major peaks in a
Hy-Cube, 70-foot, 100-ton boxcar, loaded with 30 tons of
scrap iron, occurred at frequencies of 20, 40, 60, 80, 120
and 400 Hz, with the maximum peak observed at 60 Hz. These

peaks remained even though the speed was varied and the

27

suspension altered. This would indicate the responses are
a function of the car body natural frequencies, and are
excited by impulsive loadings at the interface of the wheel
and rail. It was also found that increased speed resulted
in increased levels of vibration and vibration on welded
track was lower than jointed track. Welded track
essentially eliminated the peaks. Welded track PSD data

appear very similar to wide-band random vibration.

Guins (1974) reported very little response above 15 Hz in
the preliminary report on a 5,000 mile test conducted by
the Association of American Railroads. Most of the

significant peaks were below 10 Hz.

Foley (1972) describes frequencies below 10 Hz as the
result of transient impulses. Byrne and Anderson (1976)
observed "truck hunting," lateral instability within the
flatcar, occurs between 40 to 50 mph for half-staggered,
jointed track, and above 50 mph for continuous welded
track. Major peaks in the vertical axis related to rail
joint frequencies, while secondary inputs related to truck

geometry. All inputs were below 6 Hz.

A comparison of the above rail car studies show an order of
magnitude reduction in PSD levels when compared with
over-the-road trucks. Scarcity of data requires more rail

testing, especially in the low frequency range from 0 to 10

28

Hz. Summary of the above information shows conventional
railcars have a maximum PSD level of 10"3 g2/Hz. See Figure
2-11. Figure 2-12 shows an envelope curve which summarizes
the peak accelerations measured in railcars. By using only
peaks, a conservative profile is defined. Figure 2-13
shows a composite of various rail frequency spectra.
Measured frequencies occurred in the vertical axis, and are
profiled as percentages of peak accelerations. A direct
relationship has not been found between the effects on
lading of a constant amplitude sinusoidal vibration and

random vibration.

2.2 Shock Inputs
One difficulty in analyzing shock data as an environmental
parameter is that it must be separated from vibration data
and treated accordingly. Shock events were not always
defined in the same way by the various investigators. Some
events were classed as a vibration input in one study and a

transient shock in another.

Shocks have been measured on both the lading and the lading
floor, in various vehicles. Inputs to lading are generally
reported as peak accelerations and those inputs to the
lading floor are presented as peak accelerations, shock

spectrum or spectral analysis.

Data reported as peak lading acceleration has several

advantages, since it can be directly related to product

29

.mumc Qmm mo aumfifism I muuommm hocoswonm Hmoaflmm .HHIN madman

4... I >02uaoucu

 

 

 

 

 

950.. 08. 8. o. 0.. ..o
4.44 a . _:4444 4 4 #:24_4 1I4 #:4444.4 4 .2..4_. _. to.
I m
I I m
l u .
.I a E .l.h: m
.
— .J’ ‘ l m
_ J ‘:.\< V
‘ _ . _ 4 I 1
..’. p. I
.u/ . < I m
a a. u a
= L II.
II __ I A
a. l ‘00-
...-I-I-I 3453 I 9
I'll EOE 1 N/I
II in II II ”23 l I
L
u
P P bi — 1 80°-

muoumuonmq muoscoum phenom «mousom

30

.GMOHHHMH I mm04m>cm c04umum4moom :04umunfl> .NHIN musm4m

4: ll >UZUDOUCK

R89 RB. 92 o. o; .6
4444114 4 1444444 4 1444—44 4 1444—4 4 4 144444 44 .oo

 

l

iuIIlI 1

3.9 ~ NOMVUTIJDIDV

.. \

uIIlI L,1
(xvzd-O)

 

 

p P p b 0..
anoumuonmq mposcoum umouom “condom

 

31

go.

.A4m04uwm>v cMOMHHMH How muuoomm xocmsvmum

a: I > UZNDOUCK

.m4Im musm4a

 

 

 

 

 

g. 8. O. 0.. ..0
444444 4 44444444 4 44444444 4 44444444 4 1444444 4 40.0
i .h.d
l
l
.I Inc;
I.
L
u
L
1
l
_ r _ _ o.

.Nuoumuonmq muoscoum umouom

«condom

( NV3d ‘0) ..9 '- NOIIVUTBDSV

31

.A4m04uuo>v cmon44mu now muuoomm moconvmnm

4: III >uzu30u¢h

.m4um musm4m

 

 

 

 

 

g9 g. 8. O. 0.. ..0
44444 4 N 1444444 4 144444 _ 4 144444 4 4 4444444 4 4 40.0
A
M
i .l ..O
.I Inc;
i
u
_ _ 4 C o.

Nnoumuonmq apostonm umouom

"condom

(8'34 '0) ..9 - NOILVU3‘I333V

32

fragility or "g rating. The ASTM D-3332 Standard Test
Method for Mechanical-Shock Fragility of Products, Using
Shock Machines describes the procedure for this test.
Packaging engineers can use such data to design appropriate
cushioning systems for their products. Appendix A reviews
the fragility concept. Measured peak acceleration to the
floor of the vehicle is not as useful to package cushion
design unless the product is simple and unprotected. More
sophisticated studies have translated shock inputs to
either shock spectra or peak acceleration versus natural
frequency of the responding system. Using this
information, the maximum value of the response, at the

dominant forcing frequency, due to the input motion at

points where cargo is located, can be determined.

Another manner of analyzing shock data is to do a waveform
analysis which plots peak acceleration as a function of
frequency. It has been suggested by Harris and Crede
(1976) that most transients in the distribution environment

could be simulated with a random vibration generator.

Some method of marking transients is required to separate
them from continuous input such as random vibration, if
recorded on the same instrument. Rail cars experience
vertical transients when cars cross intersecting track,
switch junctions, roadways or bridges. Longitudinal inputs

occur during switching or coupling operations. No humping

33

occurs with loaded flatcars (TOFC) or well cars (CIWC).

Run-ins and run-outs do occur, longitudinally, with the

above, and were measured as transients.

Studies on commercial trucks have indicated that transients
produce increased low-frequency vibration and 32; sharp
peaks. The general level of peak accelerations at given
frequencies are much higher during a transient event.
Kusza and Sharpe (1973) report levels of 2.7 and 5.0 g’s
for a truck crossing an expansion joint and road
intersection, respectively. They also conclude that if a
random noise generator is used to produce a test signal,

these higher levels will be produced automatically.

Johnson (1971) found that impact pulse shapes for heavier
packages, reported in truck studies, were nearly
half-sinusoidal, with durations of 3 to 5 ms, for more than
90% of the events recorded. These packages also showed

lower accelerations. Weights ranged from 50 to 2,200 lbs.

Sharpe’s (1972) data for rail crossings showed peak
accelerations varied from 0.95 to 1.4 g’s. Two types of
responses were noted. The first demonstrated uniform
increase and subsequent decline with time. The other
indicated a large initial acceleration followed by a slow

return to conditions prior to the event.

34

Package bounce and repetitive shock can be very damaging.
This has been demonstrated by glass breakage studies in
Britain. Lee (1969) performed studies using low cost,
highly fragile, glass bottles, bulbs and ampoules to rank
the severity of transients. The number damaged was used as
the scaling factor. Rear overhang areas were the worst
positions for cargo and corresponded to equivalent drop
heights of six inches or less. Cargo peak accelerations
versus the number of occurrences are useful if simulated in
the laboratory. Vibration parameters measured on empty
trailers may address this issue. However, this
investigator was not able to find any correlation studies

between empty truck vibration and bouncing packages.

CHAPTER THREE

TEST METHODOLOGY

Trailer/container instrumentation procedures involved the
use of eight piezoelectric accelerometers. Four

accelerometers measured transient shocks and four measured

vibration inputs. The shock and vibration accelerometers
were placed, side-by-side, in each of four different
locations.

Previous studies completed by Kenworthy (1979), suggested
the midpoints of highway trailer floors experienced the
least energy when placed on rail flatcars. In order to
record the most severe inputs, two vertically mounted
accelerometers were placed over the kingpin and the rear

wheels along the trailer floor centerline, Figure 3-1.

Placement of accelerometers on vehicle floors was found to
vary from study to study. A trailer floor is not a rigid
plane. It has various bending modes. In order to identify
the characteristics of this response, a modal analysis was
performed. Floor resonances were excited and analyzed.
Results for a 38-foot truck trailer floor show major

resonances fall within the 1 to 12 Hz range. These are
35

36

.xosuu cmoqusuIHm>

 

_§4§s§._

 

 

 

450°

 

 

 

 

 

5.35.4552»
233......44183 ..

35.45485 .
O

£52§>u§-RS£TI.

 

 

 

'3 363820.21 o

 

 

 

 

 

37

most pronounced over the rear one-third of the floor. See

Appendix C for details on these measurements.

The lateral axis was instrumented in the forward left
corner of the trailer. The study, referenced above,
indicated lateral accelerations of TOFC shipments were 1.5
to 2 times greater at the top than the bottom of the side
walls. A mounting height of six feet was chosen, since the
single pallet heights were approximately five feet, and

this would represent the most severe input cargo would

experience.

The longitudinal axis was also instrumented at a height of
six feet. Both lateral and longitudinal accelerometers
were placed in a corner support column, to minimize the

effects that noise from wooden inner panels would have on

the readings.

The trailer/container lading in this study was comprised of
liquid chemicals and photographic materials. Liquid

chemicals were placed in 55-gallon fiber drums and 5-gallon

polyethylene bags inside of corrugated shipping
containers. Photographic film and paper were also placed
in corrugated shipping containers. All pallet loads of

product were stretch wrapped for purposes of unitization.
The 48 by 40-inch pallets were "pinwheeled," or placed with

the 48-inch length staggered, to minimize load shift. Air

38

bags were also used to stabilize the cargo. Protective
wooden boxes, 8 by 8 by 2.5 inches, were placed over the
accelerometer mounting positions on the floor, as seen in

Figure 3-2.

All recording instruments were strapped to the top of the
forward-most pallet, as shown in Figure 3-3. The bump
recorders were contained in urethane cushioned blocks and
the tape recorder was placed in a large plastic container

lined with polyethylene foam.

3.1 VIBRATION EQUIPMENT
The device used for recording random vibration inputs was a
Bruel and Kjaer (B&K) Model 7007 instrument grade, FM,
reel-to-reel tape recorder. The data was stored on a

magnetic tape.

In order to tape record portions of the entire trip, the
recorder was fitted with a programmable timing module. As
seen in Figure 3-4, the module has a series of dip switches
that control the frequency and length of recorded
segments. Using a tape speed of 38.1 mm/sec., the recorder
required nine meters of tape for a four-minute Segment.
Since the recorder was activated every two hours,
approximately 648 meters of tape was used in the six day
trip from Rochester, NY to Los Angeles, CA. This

programming procedure allowed a 690 meter tape to be used

effectively.

39

 

l

 

 

 

 

 

 

 

 

 

 

 

I—w—I

Figure 3-2. Pinwheel pallet loading configuration.

 

 

~ TRAPS

 

 

 

 

 

 

 

 

 

MR

Figure 3—3. Instrument placement and protection.

40

 

 

Timing Module

ON N BIQFF

 

 

 

 

RECORD

PERIOD

 

1
i

 

 

 

5 6 7 8

8i 8 8 8

l

 

 

TO!

38
68
98

 

 

 

 

128

158

 

 

 

 

278
388
338
368
338
428
458
488

 

 

 

 

 

 

 

 

 

 

 

l
l

 

 

2 3 4

El 1

 

1

1

l

1

Bl 8

 

 

1°!

 

2

 

 

 

18
12
14
16
18
28
22

 

 

 

 

 

 

 

 

24
.25

 

 

38
32

 

 

 

 

Tape recorder timing module.

Figure 3-4.

41

The recorder was powered by two lZ-volt gelatin batteries,
wired in parallel, as shown in Figure 3-5. Based on
information from B&K, the range setting for all four of the
charge amplifiers was 100 mv/m/sec . This translated to a
calibration factor of 3.22 g/volt. The frequency range was

0 to 120 Hz.

3.2 SHOCK INSTRUMENTATION
Shock data was gathered on B&K Model 2503 bump recorders.
The four units were individually powered by rechargable
nickel-cadmium batteries. The threshold level for a shock
event was set at 2 g’s. This figure was chosen after
preliminary data indicated 1 g would exhaust the tape too

soon, and 5 g’s would not generate sufficient data.

Analog signals generated by the shock transducers were
translated to a paper data tape by a thermal printer. Each
tape 'was capable of recording over 700 events. Events were
triggered when a 2 g shock occurred, and ended when the

analog signal fell below the threshold for 1 second.

Figure 3-6 shows the entire schematic for the shock and
vibration instrumentation. Cables were shielded and used
micro-dot connections. The cables were combined into an
insulated wiring harness to avoid kinks and abuse during

transit.

.u4souflo HoBOm Hocuoomu c04umun4> .mIm whamHm

mono uh

 

 

  

 

usmcH uo3om unmcH
on omumcu
> 3 3 m6 on > «4

42

 

 

 

omumco uo3om .uxm

 

 

 

BAO> NH

 

 

 

 

 

 

BAO> NH

I +

 

 

{ _.

 

 

 

 

 

 

43

A80"? RECORDBRS

 

Bruel s Kjaer
2503
VIBRATION RECORDER

 

 

Bruel & Kjaer 7007

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6 Bruel & Kjaer
2503
1 2 3 4 , _
1? ( I (I
7
Bruel s Kjaer -
2503
LAT
‘—‘——'—4:fl *me
LONG []
VERTR
{fl Bruel GI Kjaer
VERTF 2503 a

 

 

 

 

ml]. l.” “I: I.

Figure 3-6. Instrumentation schematic.

44

Mounting the accelerometers was accomplished using 10-32
UNF metal studs. Mounting holes were tapped in the
hardwood floor and the steel corner column of the trailer.
With this hard mount, mechanical filtering was not deemed

necessary I

3.3 TEST PLAN
The trailer/container was shipped through an intermodal
distribution system. It was placed on a rail flatcar in
Rochester, NY and line-hauled, with no stops for car
changes, to Chicago, IL on Conrail tracks. In Chicago, the
trailer was grounded and trucked to the North & Western
Railroad. The container "box" was removed from the trailer
chassis and placed on top of another similar container in a
double-stacked "well" car. It was then line-hauled to
Fremont, NE. The container was transferred to a well car

on the Union Pacific Railroad and hauled to Los Angeles,

CA. Upon arrival, it was grounded and trucked to a
regional distribution facility in Whittier, CA. The
recording devices were turned off, disconnected and

returned to Rochester, NY. See Figures 3-7 and 3-8.

The various railways are identified on the map shown in

Figure 3—9. Computerized summaries showing the train
movements were received. Through these records, all train
routings could be traced. By using this information,

running times were determined and plotted on the chart in

Figure 3-10.

45

 

I .HMU “NH.“ ICOIHOH ..HMHB

.AIm musm4m

 

 

 

 

 

 

 

 

 

 

 

 

 

 

46

m...

camJHmom

 

.umo HHmBICHIHocflmucou .wIm musm4m

    

     

:5 4592

...w

47

.mmE mason 44mm .mIm ousmam

       

to .ooaooc¢ a

s“..— UHhHO‘m ZOHZD
Hz .ouuaam canon
N2 .ucosuh
4 ‘0 0’0 0 4 4 ”do
\0 d A ,3 ..GBOhOr—U 033 Uddm

I .owa04nu
Htflxou
r! . uvnosuox . zxmamu3.a zaaoz

.mme4u 4m>mnu o34o\omoe .04Im m

as: case 950...
anunuaoueeesaxaueemnucm4n~ 4 o

b — P p p _ b b b — r bbbbbbbbbbbbb

E
, @§\\\\\\\\\\\\\ .5
\ \\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\ .5

NS

. 4S
\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\ .s
\\\\\\\\\\\\\\\\\\\\\\\\ s

\\\\\\\\\\\\\\\\\\\\ / ////////////////// \\ I
\\\\\\\\\\\\\\\\\\\\\.r s

in“. gaemfl .Eptsehl E3

 

 

 

 

 

 

 

 

 

 

 

49

Running times were important considerations, because train

vibrations were measured both at rest and underway. For

purposes of the laboratory test format, only "live" data

was desired. A method of isolating live data was based on

the timed "on-sequence" for the recorder and the

accompanying train status. See Table 3-1 for details.

1/9

1/10

g

DATE
1/10

1/11

1/12

1/13

1/14

50

Train Status.

*3
H

07:00
23:37

13:05
13:17
17:45
18:45
21:10
21:20

07:50
08:45

02:35
10:41
12:45

09:21
10:20

ETATUS

Recorder activated

Depart Rochester. NY (Conrail)

Arrive Cleveland. OH

Depart Cleveland. OH
Arrive Chicago. IL

Trailer grounded

Trailer transferred

 

(North a Western RR)

STATUS

Arrive North a western RR

Depart Chicago. IL

Arrive Fremont. NE

Depart Fremont. NE (Union Pacific RR)
Arrive North Platte, NE

Depart
Arrive
Depart

Arrive
Depart

Arrive

North Platte. NE
Cheyenne. WY
Cheyenne. WY

Salt Lake City. UT
Salt Lake City. UT

Los Angeles. CA

Hell car on unloading ramp
Container grounded

Container picked up by Golden Eagle Express
Container arrived at distribution center

in Whittier. CA

All times shown are Eastern Standard time. and are based on a
Dates listed are from 1/7/87 to 1/14/87.

24-hour cloak.

CHAPTER FOUR

DATA ANALYSIS

4.1 RANDOM VIBRATION DATA
Vibration data was processed on a Zonic 6088 Spectrum
Analyzer. The taped data was classified according to train
movements so that only live segments would be used to
define the laboratory test profile. Each of the live

segments was processed separately and then compiled for a

general profile. Processing was undertaken at a tape speed
of 381 mm/sec. A total of 16 averages were taken for each
nine meter segment. Comparisons are shown in Appendix D.

After determining the PSD profiles for each axis, the input
frequencies were identified using a peak selection
technique described in the Zonic 6088 Operations Manual.
This allowed the noise to be subtracted from the data.
Noise floor levels are shown in Figure 4-1. This process
was done on an axis-by-axis basis, since the laboratory

equipment could generate only single degree of freedom

motion.

4.2 TRANSIENT SHOCK DATA
Shock data recorded on paper tape was transcribed and

entered into an IBM PC XT computer. A LOTUS 1-2-3
51

52

NI

NU®G®.4

¥

Vt

.40044 om40c 0:43onm sump o4mEmm

 

4USO®.4

 

 

 

 

 

 

.4I4 ou5m4m

NIU4
mlu4
mlu4
Vlu4
mlu4
Nlu4

NIU4
mlu4
mlu4
Vlu4
MIU4
Nlu4

53

spreadsheet was used to manipulate the data and prepare the

statistical reports. Full printouts are listed in Appendix

E.

In the sample data entry shown in Figure 4-2, the shock

events at each location are tabulated and listed as to the

Day, Hour, and Minute of occurrence. The Velocity change
is calculated and listed. These steps are all made

possible through the macro-program which is used with the

spreadsheet.
Values were double-checked for accuracy of velocity
calculations. The presence of large values for

accelerations were tested to see if they indicated recorder

error .

4.3 VIBRATION RESULTS
Vibration data on the TOFC portion of the study revealed a
broad-band maximum PD level of 1.5 x 10'4 gZ/Hz, on the
lateral axis, from 5 to 10 Hz. This indicates suspension
resonance. Kenworthy (1979) found a resonance at 1 Hz,
with a PD value of 2.5 x 10"3 gZ/Hz, for TOFC. The current
study also found a resonance at 2 Hz. Neither of the
studies show significant energy at higher frequencies. See

Figures 4-3 and 4-4.

Also of interest is a study on the Canadian Pacific

Railroad reported by Naylor (1987). It was noted, in TOFC

54

 

 

.QOu must umcuooou mfism .mlg oasm4m
or... _ _ _ _
2944454304 :63“; use...

55

L b 5 b b P P

.204umun4> I ommoaro on umummnoom .omoe .mIv musm4m

«carcasses:
04 was.

b ppphhbb p P

 

l

.\ . /
/.<.><sfl

 

.4"

wmv I a 1cvvvtv v v
09‘.“
.e
O

.4"

.\
. , , ../. .\\
4 / c \\

zH/zb

\

"V' V I IIVVUV I i
8'

 

 

Acum3uomv 4m04uuo> Hammond 4MC4csu4mcoq on fi.

 

 

56

.coflumunfl> I mo4mmcm mOA ou ommoflsu .OBHO .¢I¢ whamHm

awry Econ—vet
9

pbppb

. 3.0
a b L

 

 

 

 

 

b -

 

zH/ZB

 

 

 

57

shipments, the vertical response over the king pin, on the
trailer floor, demonstrates a larger vibration response if
the trailer air brakes are set as it rests on the flatcar.
In the current study, the trailer brakes were pg; set
during the TOFC portion of the journey, and thus explains

the lower vibration level.

It should be noted that the cable connecting the vertical
vibration accelerometer over the rear wheels failed during
the test. The only vertical vibration data presented was

measured in the forward location.

The CIWC analysis indicates quite a different profile.
Little energy is present at low frequencies of 0 to 8 Hz.
The vertical axis shows readings of highest energy levels
in the 10 to 13 Hz range, with the peak at 7 x 10"5 gZ/Hz.
Removing the truck suspension system altered the low
frequency response of the system. However, the well car
suspension is observed to be the input frequency in this
range. The overall difference between the random vibration
characteristics of TOFC and CIWC modes is a flattening of
vertical and lateral responses, below 10 Hz, for the

latter. The longitudinal input remains the same.

4.4 SHOCK RESULTS
Shock data recorded for the two modes indicated a shifting

pattern for input shock ranges. Each axis and monitoring

58

position experienced a specific shock level distribution.

For TOFC, the results were:

Vertical, front 2-5 g’s 76%
Vertical, rear 4-6 g’s 50%
Longitudinal 0-2 g’s 75%
Lateral 11-12 g’s 50%.

For CIWC, the results were:
Vertical, front 2-3 g’s 83%
Vertical, rear 2-7 g’s 75%
Longitudinal 1-4 g’s 13%
7—10 g’s 20%
15-20 g’s 44%

Lateral 6-13 g’s 79%.

The results summarized above are taken from Figures 4-5 to

4-15a

Since the distribution range for shock levels varied
axis-by-axis, a representative test procedure must also
vary these inputs. The test format explained in the next

chapter will address this issue.

m VorIIcaI Axis (Front)
[38 totd data points shown

59

 

 

 

9

I

“’3’

4
”a? I

I
ACCELERA'IION (9's)

4'
..p.

it:

s
00,

a,
/\\/

9’0

 

.p

 

so Percentage o: Shocks (z)

 

TOFC - shock distribution, vertical (front).

Figure 4-5.

60

.Aumouv 4m04uno> .coflusn4uum4c xoocm I Oboe .mIv madmam

 

 

 

 

 

 

 

 

4932345602

I e§.owam. (beb:.vb.o.+ :9 9.94.39.94.44 :4... «on... a» \o o
\

a 4 -.
\ \
\ \

““ ““ .Io—
\ \

\ \ In.
\ \

v “ -2
a \

52.. £828.22;
on

 

 

cone .3 3t; .N. E 8.85 t .852...

61

.4mcflcsuflmco4 .c04u5n4uum4c Moocm I Oboe .NIc musm4m

c. 8 22254304

Lg“? fix? %¢\9\V\9:\»;9\0\¢9\11\9:\009 \f \0 4‘0

  

A“!

0—

ON

on

04

on

5.2.. 28 .22 a
«3 3:35:04 N ARV 8.005 3 engaged

on

62

.4muoDMH .c04usn4uum4c xuocm I Oboe .mIc oasmflm

 

 

 

 

 

 

 

4.3 20552
\L? \Qs A. \ \ \ \ \
a. late.» skate. $.99 94.94.99 4.... e... WIN/4.49....“a o... no a... a... «a as...» «a o

m -2
\
\ I on
\

.. \ a
\ I On
\
\ I 3
\
a
8 I on

mm
:32? 44:. 20v 342 L on

.3 .223 U 3 8.02m 40 3252...

63

.994909500 .co4usnauum4o xoonm I omoe .muv wusm4m

6. is 223.483
933.4... 9&9999 9 :999m9991999:99.9 I...» «9 s 4...... «9.4.9..

fl.“ 4 . .4“ a ,
. mi .. _n _. __

. __.
.

 

. -\\‘

: \“\\“““‘\“"\“\“\\‘\‘\\\‘\“““‘ I}

_
_
_
_
._. .8

a _\_‘—‘_‘

' \\\“\“\\‘\\“\‘\“\“\\

Ion

I04

 

 

 

eoé§<8=3>fl .32e3Nu $863.38.... on

£55 .34 note, 3 .34 3523...... §

64

.Aucoumv HMUfluHm> .QOflufinHHumfin xuosm I UZHU .OHI¢ musmflm

 

 

 

 

 

  

          

 

 

  

Achy 20.5303
E99999................9.,,,...9.. ...,... ................. .. .
. . . . w
..
x - a
m
w
R .. e
w
m
\\x. 1 an
8
oo—

 

 

Ataé .3. _8_to>

‘E_

65

.Aummuv Hmofluum> .coflusnfluumflc xuoam I UZHU .HHI¢ musmflm

  

 

  

   

    

 

 

$3 5533
9.99.99.99Aehwého.9939.99.99.99WWWWV.Voww... «99.9.... .9 ... o
. ....u«u*““\\ .

v\\\ In

x “ \ .
“\\ -2

.. .. A“
a “ .. m.
\ 2

“ u cu
m r 2

 

 

 

.5533... E 9.39.3.5... a...

66

...,..9 9....

.3. 353.993 §

.9

.Hmaflwsu9mcoa

.co9usn999999 xoonm . ozHo .NH-9 musm9m

$3 zo=<fia8<

\ ‘6 \
xv .v ax» 4.0.

9.9.9

 

I
9x9.baxmxoa>.koaxynx.\w Am.

3 9.85 3 3252.. mu

67

 

9.59.... 9.... 9...”..99...

. u p p p p .

.Enifinu

 

.HmumumH .coflusnfluumflo xoosm . ozHo .mHI9 9.99.9

$3 2054633

.9939... 9.... .99....oo.9 .. ...9 9..

     

p c

O—

m—

3 9.85 3 .552... 3

68

.m mIo mufimomeoo .cofiusnflnumflw xoonm I U3HU .vHIv mnsmflm

4.3 2225.394
Tu Tu .Io oIm mi. T»

'0

IN ul— plo

 

 

 

4.44 4 44 ..44 4 4. 4... 44.4 4444 . 44 .. -.
‘M A. ‘ s u\ fi
4 \ \ a.
m s s a. \ .L: A“
\ .. m .. 3
a m r on
\
m I av
.
4 - a.
m T a.
4
m r on
4 9
.4 on
3
4.84.43.82.38 aifizsnu $3433.34... 8

€95 .3 3t.>! I .34 3533.43 §

69

.m mmlm muflmomEoo .coflusnfluumflc xoosm I DZHU

4...: 2055.394
.19 .5? v

 

.mHI9 9.5049

c—

9—

ON

4.8533328 3.223U 88.854.39.52: 3

44.3.5 .3 3:; Q .3 2.3.9.3 §

CHAPTER FIVE

LABORATORY TEST PROF I LE

The two PSD plots shown in Chapter Four are the summary
profiles for TOFC and CIWC modes, and the basis for the
laboratory test. One of the main objectives in this test

was to generate a simulation which incorporated both the

 

random vibration and the transient shocks. This goal was
achieved through the use of a digital control system and an

e lectrohydraul ic vibration table .

Instrumentation for the digital control system included a
Scientific Atlanta Spectral Dynamics SD 420 spectrum
analyzer. The coordinates for the profiles were programmed
onto the operations disk with the computer. A real-time
display identified the input signal and the tolerance band
established, as in Figure 5—1. The hydraulic system was
operated through an MTS Servo Controller 406. This unit
allowed the vibration table to be driven directly from tape
as a closed loop, sine wave input, a random signal or a

transient shock generator. Figure 5-2 shows the equipment

used.

In order to generate a shock-on-random simulation, the two

signals were combined on the operations disk. A random
70

71

PD gE/Hz

—2
10 I,

 

 

 

Frequency Hz

Broken line defines the tolerance band used in a
random test.

Figure 5-1. Random vibration input.

 

Spectral
Dynamics
415

 

 

‘53 Q

 

 

Servo
Controller

72

I—I

Test
Control
PC

 

 

Tape Drive

 

Spectral
Dynamics
420

Real Time Diaplay

 

 

Dual

C221

Disk Drive

 

 

 

 

 

 

 

Electrohydraulic
Vibration System

 

 

Figure 5-2. Electrohydraulic vibration

system.

 

73

vibration profile is shown in Figure 5-3. This represents
a portion of the measured data. By utilizing the peak
selection technique mentioned in Chapter Three, the X-Y
(coordinates for one channel were determined and are listed
:in Table 5-1. X-coordinates refer to frequency levels
Ineasured in Hertz, and Y-coordinates identify the
czorresponding Power Density levels, in units of gz/Hz. The

)(-Y coordinates for the random signal were programmed on an

(aperations disk. Individual transients were then
Iprogrammed on top of the profile. In the current test, a
15 g, 11 ms half-sine shock pulse was programmed to occur

(every 10 seconds for a period of 15 minutes. See Figure
5-4. It is possible to program each shock as it occurred
in the measured environment. However, the test level is
sset to generate the maximum normal dynamic inputs to the
Jproduct. The flexibility afforded by a digital control

ssystem allows each test to be varied, if desired.

'Ihe vibration system used for this test utilized a
SOC-pound table. The levels of energy shown in the TOFC
Iand CIWC profiles were too low to drive the table. For
ipreliminary evaluation, the test level was amplified four
Orders of magnitude, so that the table would respond as
(desired. A feedback accelerometer was used to monitor the
‘table. The inputs to product/package systems were
generated on an axis-by-axis format, as the table had only

saingle degree of freedom capabilities.

74

NI
.m

uuaaw
_§

at

b

.cowuomamm xmmm

b

Euama.m

..-m mus...

 

/

 

 

 

 

 

 

 

 

 

 

 

 

 

hqu

mime

mlud

VIMa

Mind

NINA

 

mmq 03¢ Du

Nm\mm

an

Table 5-1 .

.8?SE+B
.125E+B
.3?SE+B
.DS4E+1
.219E+1
.SSZE+1
.687E+1
.B?SE+1
.831E+1
.556E+1
.781E+1
.844E+1
.3?SE+1
.656E+1

P‘F‘UJQJGJalalUluluJUJthJfiJ

x-axis is measured in hertz

Y-axis is measured in gZ/Hz

INDEX

.ZDBE+1
.SDBE+1

BOBE+1

.SBDE+1
.SBBE+1
.DDBE+1
.4DBE+1
.DDDE+1
.SDBE+1
.SDDE+1
.SBBE+1
.180E+1
.DBBE+2
.17BE+2

75

Peak Coordinates.

X

HERTZ

uJthJththF'h‘h-F'h-UJGJU)

flwflerw-m--v

Acceleration G's

+14.962

76

15 9'5 11 ms 1/10 sec,

15 min

..1- ‘Q- --‘-

 

 

—h\J

{l
I'
I'

J

I
'I
I
i/

 

 

-14.962 50

TOFC shock,

Figure 5-4.

100 150

Peak G 14.962

TOFC half-sine shock.

200

ms

CHAPTER SIX

OBSERVATIONS

The following observations were made during the process of
literature review, data collection and analysis. They are
intended to provide additional insight to those readers
conducting environmental measurments or seeking information
on the characteristics of the dynamics found in

transportation networks.

1. Test methodology development and data measurement took
place between March, 1986 and February, 1987. During that
period five rail trials were undertaken between Rochester,
NY and Los Angeles, CA. These included: one shipment of
over-the-road truck from Rochester to Chicago, IL combined
with TOFC from Chicago to Los Angeles; one CIWC run from
Los Angeles to Rochester; and three runs from Rochester to
Los Angeles in the TOFC - CIWC format described in this
study. Data presented in this study are from the second

such trial.

2. Multiple channel data collection requires many cables,
lead wires and accelerometers. A wiring harness was used
to eliminate tangles and to reduce set-up time. It also

77

78

reduced the amount of abuse received by the cables. Prior
to using the harness, two cable shields were broken and one
center hot wire failed. It is also suggested that
micro-dot connecting links be kept on the ends of

disconnected cables at all times.

3. Unattended instrumentation is often powered by
rechargable nickel-cadmium batteries. Useful charge-life
is uncertain with these systems. Power failures occurred

 

on some of the instruments during the first three trials.
Nickel-cadmium batteries tend to develop a short "charge
memory" if they are recharged frequently after only partial
discharge. The solution to this problem involved the use
of 12-volt gelatin batteries. They can be shipped in any
position without the spillage of lead-acid batteries.
Additional parallel units can be added for longer trips, or

to include periods of storage.

4. In situations where test instruments travel one-way to
collect data, personnel must be trained to remove them and
prepare them properly for return. A training session with
warehouse personnel in Los Angeles was effective in
-minimizing improper handling of instruments. Also, after
the first three runs, permanent containers were developed
using large plastic shells with hinged lids, made by the
Hardigg Co., to encase the instruments, wiring harnesses,

batteries, accelerometers and a few small tools.

79

Polyethylene foam was fabricated as interior cushioning

material.
5. Accelerometer mounting was accomplished with 10-32 UNF
metal studs. No problems were encountered with tapped

holes in either the hardwood trailer floor or the steel
support columns. One temperature-controlled shipment
required a refrigerated trailer. Such trailers have
grooved floors which allow accelerometers to be epoxied in
the recesses, out of reach of the forklift and pallets.
During this trial shipment an angle iron bracket mounted to
protect a forward accelerometer was sheared off due to load
shifting. That accelerometer failed. The one mounted in

the groove was undamaged.

6. It is standard practice to use spectrum analyzers to
process random vibration data. A sensitivity or range
setting of 100 mv/m/sec was a fair compromise between a

less sensitive range that minimizes cross-talk, and one
more sensitive which amplifies the noise levels making it

more difficult to select peak resonances.

7. Cargo placed in a trailer/container, over the rear
wheels or on the rear overhang, will experience the most

severe shock and vibration inputs.

8. Sources of transient loadings for rail cars, in the

vertical axis, would include road crossings, switch

80

junctions and rail joints. These contain both short

duration, 30 ms, and longer duration, 100 ms, pulses.

9. The number of shocks over 2 g’s was fairly large, over
700 in each axis, for the 6-day trip. Also, longitudinal
shocks had higher g-levels than expected with the
articulated couplings of CIWC shipments. Review of the raw
data revealed that over 90% of those shocks were less than
30 ms duration. Therefore, they generated little energy.
The only long duration shocks, those over 100 ms, occurred

during the loading and unloading operations.

10. Coupling shocks occur as train cars are joined in
classification railyards. An average impact speed for
coupling operations would be 8 miles per hour. Shocks of
10 g’s, 30 ms duration are typical of those generated with
standard draft gear couplers. Cushion underframe draft
gears have a travel range of 24 inches during coupling, and
experience shocks of 2.3 g’s and 250 ms duration. TOFC and
CIWC shipments are not supposed to be coupled with the

trailer/container units in place.

11. Shocks in empty or lightly loaded vehicles were found
to be much higher than heavily loaded ones. The majority
of the shocks increased from 5 to 15 g’s. Ninety percent

of the durations were from 3 to 5 ms.

81

12. By far the most common vibration frequencies found in
all forms of commercial carriers are between 1 and 12 Hz.
These inputs are attributed to vehicle suspension systems.
This characteristic is the same for both over-the-road
truck suspensions and rail truck assemblies. The CIWC
environment generates peak amplitudes at the high end of

this range, near 12 Hz.

CHAPTER SEVEN

SUMMARY AND CONCLUSIONS

1. The collection and analysis of data in this 2,877-mile
rail distribution network marks the first published study
of a commercial, non-experimental, CIWC shipment of more
than 500 miles. Studies of such length are also scarce for
TOFC modes. Soroka (1986), of the Ontario Research
Foundation, indicated that a 1985 study of a TOFC shipment
from Toronto, Ontario to Vancouver, British Columbia, on
the Canadian Pacific Railroad, was never formally analyzed
and published. In addition, Grosso (1987), of the
Association of American Railroads, reports that a
5,000-mile boxcar vibration test in 1974 was never
completely analyzed. Therefore, it has been determined
that environmental data on long distance rail shipments in

general, is limited.

2. The use of unattended instrumentation, described in the
methodology chapter, was successful in recording the shock
and vibration dynamics in all three axes. This technique
has been established as the standard procedure for
measuring all distribution networks utilized by the company

sponsoring this study. Collected data is now used to
82

83

establish test parameters for each respective test
simulation.

3. Evidence indicates the TOFC and CIWC rail transport
modes are increasing in use. They offer a more

cost-effective means to transport products for longer
distances, especially distances greater than 1,000 miles.
CIWC offers a significant economy of scale in freight

volume transported versus fuel consumption.

4. Mainline rail systems have been improved through the
replacement of jointed track with continuous welded track.
Measured inputs to rail vehicles indicated the sharp
vibration peaks have been eliminated as a result of this
change. Input responses to rail vehicles traveling on
continuous welded track were found to be more like
wide-band random vibration. CIWC PSD profiles demonstrated

such behavior between 0.91 and 10 Hz.

5. The current study found CIWC random vibration input to
be one order of magnitude lower than TOFC. Maximum PSD
values were found to be 1.5 x 10"4 gZ/Hz in the 5 to 10 Hz
range for TOFC in the lateral axis. CIWC maximum was
reached in the vertical axis and had a PD value of 7 x 10"5
g2 /Hz. Most studies completed previous to the current
investigation focussed on the low-end frequency scale. Low

fequencies of 1 to 13 Hz are usually associated with

vehicle suspension systems.

84

6. Shock data indicates that each axis is different, with
respect to the g-level of the shocks occurring in that
axis. The vertical frog; location experienced 76% of the
inputs in the 2 to 5 g range for TOFC measurements, and 83%
of the shocks in the 2 to 3 g range for CIWC. The vertical
reg; location encountered shocks in the 4 to 6 g range 50%
of the time for TOFC and 2 to 7 g’s 75% of the time for
CIWC. Fifty percent of the lateral shocks for TOFC
occurred between 11 and 12 g’s. Seventy-nine percent of

those for CIWC were between 6 and 13 g’s. Longitudinal

 

inputs for TOFC were relatively low, with 75% falling

between 0 and 2 g’s. Longitudinal inputs for CIWC were
quite different. Only 13% fell at the low 1 to 4 g range.

20% occurred between 7 and 10 g’s, while 44% were between
15 and 20 g’s. The articulated coupling did not appear to

reduce longitudinal inputs.

7. The vibration PSD profiles in this study were the basis
for a laboratory simulation procedure. In addition to the
vibration input, the transient shocks were considered. A

preliminary test profile has been established which
incorporates a 15 g, 11 ms half-sine shock pulse along with
the random vibration signal. Observation of the test
profile’s effects on stretch-wrapped pallet loads of
product indicate definite displacement of tiers when the

transients are added.

85

8. Several aspects of this study have influenced
management decisions. Cargo in the first trial run
experienced significant load shift during the TOFC portion
of the trip. Documented slides and video tapes of the
difficult unloading process in Los Angeles altered existing
shipping policy. In addition to stretch wrapping pallet
loads, air bags were added to stabilize the lading. This
solved the problem without the need to implement blocking

and bracing techniques.

9. Another consideration involved product from Japan,
which was shipped to Rochester via the Panama Canal. These
products were hand-loaded overseas and used no pallets or
slip-sheets. Management requested information about the
CIWC environment to see if it was suitable for such cargo.
Data from the second trial, Los Angeles to Rochester by

CIWC, proved this method to be acceptable. This resulted

in time and cost savings.

10. Preliminary test results, with containers of liquid
chemicals, have shown the new procedure to simulate field
damage more effectively. This has been attributed directly
to the addition of transient shocks to the dynamic input.
As a result, policy has been implemented to incorporate the
test procedure as a means of package system evaluation for

all product shipped TOFC and CIWC.

86

11. A new servo control system has been ordered to allow
the electrohydraulic shaker to operate at levels actually
measured in the field. The system currently requires the
signal to be amplified in order to drive the table. A
study is also underway to examine the possibility of
installing a new vibration system with multiple degree of
freedom capabilities. Such a device will require a larger
foundation than the existing shaker. It is estimated a

system of this type will cost approximately $750,000,

installed.

CHAPTER EIGHT

RECOMMENDATIONS FOR FURTHER STUDY

More long distance studies should be undertaken in order to
build the data base for probabilistic determination of peak
vibration accelerations expected in the intermodal rail
environment. It is also important that shocks be monitored
as well, since they can have a significant effect on the

resulting dynamic input to rail vehicles.

In addition to the above, it is important that the dynamics

be measured in all axes. Nordstrom, et. al. (1986) of
Packforsk in Sweden, have developed a Computer Aided
Transport Simulation (CATS) test procedure. They stress
the importance of testing with multiple degrees of
freedom. Therefore all axes are monitored in field data
collection procedures. The electrohydraulic shaker used in

this simulation currently moves in vertical, horizontal and
rotational directions. The test described above, can be
amplified to reduce test time, if fatigue and abrasion are
not major considerations. This would be determined on a

case-by-case basis.

Current work utilizes software to sort vibration spectra

and transient shocks in a biaxial format. Both vertical
87

88

and horizontal field signals are processed and used. A
program has been developed to merge these files and create
the combined test. Initial trials with aeseptic food
packages show the technique capable of simulating field
damage more precisely than conventional sine-sweep

procedures, or random-only inputs.

New test instruments are being introduced for measuring the
combined dynamics of the environment and processing them
directly through the computer. Through use of such
devices, it will be possible to have tailor-made test

programs for each specific distribution system.

‘1

.—4 I1 ' A f

APPENDIX A

FRAGILITY

APPENDIX A

FRAGILITY

Fragility is the term used to identify the acceleration
necessary to damage the critical component in a product.
The procedure, developed by Newton (1976), measures the
combined effects~of acceleration and velocity change on the
product structure. Velocity change for a half-sine shock

pulse is described by the equation:

 

 

 

 

AV = AP 9 1' 2/11 (A.1)
AV = Velocity change (in/sec)
AP = Peak acceleration (9'3)
9 = Gravitational constant
(386.4 in/secz)
1' = Duration (sec)
Z/K = Shape factor.
D
g tum“!
\munwwuumuummHMEumm

Figure A-l. Fragility curve.

See also, Brandenburg and Lee (1985).

89

 

90

The relationship between the shock pulse and the cushioned

product response is shown in Figure A-2.

rum-mans:
ACC. E
9'5 -
mummm

 

llllL'Llllllll Time(sec)

Input shock and cushioned response.

 

Figure A-Z.
The product and its critical component can be represented

as shown in Figure A-3.

“component

 

 
  

  

-- spring constant

K

 

 

 

/c2 2
coefficients ‘ f—PrOGUCt
of ‘\\
damping c1 K1—- spring constant
INPUT
Figure A-3. Spring—mass system -- product and component.

Amplification of the input shock to the critical component

can be described in Figure A-4.

91

 

 

 

 

 

 

 

 

 

 

 

10.0 DIFFE [NT
(5
RATIO
MIPL IFICAT ION
0.1 I
0 .5 1.0 1.5 2.0 2.5

Figure A-4. Amplification of input shock.

APPENDIX B

DETERMINATIONS OF PEAK ACCELERATIONS

APPENDIX B

DETERMINATION OF PEAK ACCELERATIONS

Peak accelerations can be determined from‘ PD values as

follows:

 

i=N
jg:: (nwus>§ /1a
i=0 2
PD = 9 /Hz (from Chapter 2)
BW’ (2.1)

PD = Power Density (gZ/Hz)

RMS = Root Mean Square acceleration (g's)

N = Number of data samples averaged

BW = Bandwidth of the filter (Hz)
(usually normalized to 1 Hz).

Any random phenomenon will generate a normal distribution
of its values. In this case, the amplitudes of random
vibrations are assumed to be randomly distributed about a

mean value of zero. See Figure B-l.

 

 

 

 

 

 

 

 

 

 

 

 

 

>- .4 I
t: ,. 3
P O
25 a; ./’
z: :5 .2 a;
:3 == ,{’ 53.3za,‘\\
fl- ..1 //". ,\
O ’4 fl -A
-4 -3 -2 -1 o 1 2 3 4

a/sigaa
Figure B-l. Normal distribution of acceleration levels.

92

93

Distributions about the mean determine the level of energy
present at any given frequency. The Root Mean Square (RMS)

acceleration is then related to Power Density as the

standard deviation is to the variance of a measured
population. Therefore:
PD = (RMS)2 = variance
VPD-= RMS = standard deviation

It is possible to determine the range of peak accelerations

at a given frequency by using the following procedure:

10-2

   

 

 

l
in I
5; 10-3 I
O I
I
I
10“ I
I
I

l l 1 l 1 l 1 1 1 L l

0 IO 20 30 40 50
huwummru~

Figure B-2. Power Spectral Density plot.

‘At 5 Hz, the measured PD level is 10‘2 gZ/Hz. Therefore:
VPD = 10'1 g = RMS acceleration

RMS acceleration = 0.707 AP (B.1)

AP

AP = 0.14 9.

Peak acceleration (g's)

94
One standard deviation (one sigma from Figure B-l) for the
distribution of acceleration amplitudes, would be.: 0.14 9.
Three standard deviations would increase the range of values
toii 0.42 g. GRMS occurs 68.3% of the time and does little
damage. Three sigma accelerations occur only 4.3% of the

time, but do most of the damage.

Peak Accelerations
PD

gZ/Hz

 

A—A

 

Frequency (Hz)

Figure B-3. Peak acceleration range.

The plot in Figure B-3.identifies the frequency range that
has the highest energy level. This range may be chosen as
a suitable test profile for products shipped in this environ-

ment.

APPENDIX C

MODAL ANALYSIS OF A TRUCK TRAILER FLOOR

APPENDIX C

MODAL ANALYSIS OF A TRUCK TRAILER FLOOR

In order to describe the responses of a truck trailer floor
to vertical inputs from the physical distribution

environment, it is necessary to perform a modal analysis.

0.1 Test Set-up
Trailer dimensions and mapping grid are shown in Figure
C-l. A 38-foot Fruehauf trailer, model number FBX-F2-38,
was chosen for measurement. It was grounded and had tire

pressures ranging from 85 to 90 p.s.i.

C.1.1 Instrumentation Hardware
Test measurement apparatus consisted of the following:
B&K Force Transducer 8200
Force Hammer
B&K Response Accelerometer 4369
B&K Conditioning Amplifiers (2) 2635
B&K Spectrum Analyzer 2035

Hewlitt-Packard Computer 9816/9121

The force hammer was a 5-1b. sledge, with a force
transducer attached. A rubber tip was affixed to the end
opposite the transducer, giving a frequency range of O to
500 Hz.

95

96

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1 2 3 4 5
10 9 8 7 6
11 12 13 14 15
20 19 18 17 16
21 22 23 24 25
30 29 28 27 26
31 32 33 34 35
4011 39 38 37 36
41 42 43 44 45
5 49 48 47 46
5 52 53 54 55
6 59 58 57 56
6 _§2 63 854___,65
X-—-——-Impact point
66

 

Trailer Dimensions: 38 ft. by 8 ft.
(Not to Scale)

Figure C-l. Trailer floor mapping grid.

97

0.1.2 Instrumentation Software
Structural Measurement Systems (SMS)

Modal 3.0, Version 4.0

Basic 3.0, Version 4.0

Figure C-2 shows the grid orientation with respect to the
x, y, and z-axes. Numbering begins at point (0,0,0). The
' plane identified is the x-y plane. The display sequence

will allow each mode to be defined separately.

  
 

 

1 +—_(000:0)

Figure C-2. Grid orientation.

0.2 Measurement Methodology
The point of excitation is indicated by the X symbol in
Figure C-2. Instrumentation was set up as shown in the
schematic in Figure C-3. An impact hammer was chosen for
this test, as it is a field measurement study, carried out
in a distribution center warehouse. It was not deemed
possible to set up a vibration system, with stinger, to

excite the floor resonances.

98

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Charge
Am
8‘: Spectrum
2635 Analyzer
88K
2035
INPUT Charge
Channel A Amp
85K
2635
OUTPUT
Channel 8
H-P 9806
sax 8200
g! I H-P 9121
Response -
Impact Hammer Accelerometer
Hewlitt-Packard
with Force Bi! 4369 Computer

Transducer

Figure C—3. Instrumentation schematic.

99

Preliminary testing with the impact hammer indicated

sufficient energy to generate responses at each point on

the grid. The accelerometer was moved, successively, from
point 1 to point 65. Resonances were plotted with the
spectrum analyzer, until each point’s profile was

distinct. A sample profile is shown in Figure C-4.

Initial calibration on the input channel was set at 3.16

mv/N. The output setting was 10.0 v/g. Adjustments were
made, as required, during the data collection phase of the
analysis. All calibration settings are noted in Table
C-1. Table C-2 notes the various strong resonances

encountered and the frequencies at which they occurred.
Points 8,10 and 11, near the front of the trailer,
demonstrated resonant responses in the 5 to 12 Hz range.
From point 40 to 60, responses were noted in the 0 to 10 Hz
catagory. It is thus determined that response frequencies

are due to the nature of the trailer’s suspension system.

C.3 Estimate Modal Parameters
The curve fitting procedure was accomplished in two steps.
First, each profile was reviewed to determine whether
single or multiple degree of freedom techniques ‘were
required. Figure C-5 makes such a comparison. Second,
peak resonances were defined, and then curve-fit using one
of the above techniques. Then the rest of the measurements

were AUTOFIT, fitted to the same parameters.

100

-60.0

 

LOGMG TRANS #l
LOGMG TRANS #2
lO/DIV

 

 

-l40.0

 

 

 

J .1 j A L A J J

0.00 FREQ (5 HZ/DIV) 50.00

 

Figure C-4. Sample measurement.

101

Table C-l. Calibration Settings.

Initial Calibration:
Input (Channel A) 3.16 mv/N
Output (Channel B) 10.0 v/g

Point Channel Settin
6 1 mv7N

 

A
23 B 3.16 v/g
44 B 1.0 v/g
57 B 360 mv/g
66 B 21.6 mv/g
Table C-2. Resonances.
Point Frequencies
8 5 Hz, 8 Hz
10 5 Hz, 12 Hz
11 11 Hz, 12 Hz
24 0 to 10 Hz

41 to 60 0 to 10 Hz

102

 

Lightly Coupled Heavily Coupled
Single d.o.f. ultiple d.o.f.

 

FREQ. (HZ)

Figure C-S. Combining SDOF and MDOF methods.

103

C.4 Sort Modal Data
A sorting process was used to generate the necessary mode
shapes. An example, showing frequencies and damping
levels, is given in Table C-3. The mode shapes were

transferred into global coordinates.

0.5 Display Mode Shapes
The structure can be shown in animation. Each mode can be
viewed separately. Inferences can be drawn from the

animation that pinpoint potential problems with cargo

placement.

104

Table C-3. Modal Frequency and Residue (Sample).

*** FREQUENCY AND DAMPING ***

EBEQIEZ). 1381111315). 121111121112).

1. 7.52 9.17 5.37
2. 10.78 .60 2.41

APPENDIX D

PSD PROFILES: TOFC

AND CIWC

“H. Lu.- ‘J m p
.

105

up. 1
060. m
base: n
.03D- C‘U

N:

.88. I 8.03 omm

NUOQO.~

vvanUNn
OQ\QO\OO
m.

vvnwauNo
QO\OO\OO
No

vvimOUNo
OQ\OO\QO
a.

P

«u000.u .0

h

 

 

 

 

 

 

 

mommuuomm JNZZCIUHPJDI @000

mm 03¢ an

filUu
mlu«
mind
VIUq
MIUq
Nlu~

stud
qu«
quq
Vlua
Mluq
Nluq

hIUu
wlu~
mIU4
VIUu
Mluu
NIUq

.Ta 8ng

at

at

at

106

$3.85 0mm

 

 

 

mu"0«.q0
OD\OO\OO
NI
Nu§.q F D D ID b P b “as.” no
5. 3
_ ..
71$ ts . CH .W»J.ltflr_> .lx‘r
NOHMQHNQ . «k
DD\OO\QO
a.
..O. 1
ate. m UzHU
wane: «
.090. csu

mommuuomm JUZZCIUHhJDt

«0‘ 93¢ an

. «no madman

slug

mlu«

mind

VIUA

MINA

NIUA

at

APPENDIX E
SHOCK DISTRIBUTION: TOFC AND CIWC

DATA SAMPLES

~ IV

107

IN
fin
a—
Na
RN
in
I?
8p
8v—
In—
in
R—
$9
*9
8&—
Ian
8—
bzmommm

«ova—"mm
o-a

.N

ha

a

N

m

«—

nu

mu

m
mmozmazaooo
no 233232

llllfi?;:§:£jiu

mule"
muurg
ruin"
cunn—
manu—
N—qu
aquoq

COO
|~
a:

uunvmcahc?
I I
O—vNfl'lflQb-fl I

6.» .moz<¢
.amsa goozm

moduuusz cu «

no.~
so."
an.”
“van
no.0
oa.~
ul.¢:n

>mn Ohm

oa.N
a>.o
um.m
ch.N
co.—
cc.“

el.¢aa

m0<¢m><

. ae«uu«53 ou use-eons oduqoem meme:
.3: ace-noun on «no :2 .a30 on com «deacon
:0 ~60 use: as .8860 .ano ou.oou acne ones

o—.u
an.“
ad.“
oo.n
—¢.~
oa.u

u.00<
>Ha can

u N unannouAk

«0.0
an.“
an.a
nh.@
o~.a
50.9

a.ou<
ue<¢m><

<0 .hewuu«£3

>1 .ueueeaoou

usaaona coaaaxazxaanucou

«a soda. mux< AccufiauwuBOJ

“no aquauauca-6o¢«¢uaoo\oa09

.uuommm same

n>¢<tt2m

20~H<UOA
20nh<004
ZO~F<UOJ
ZO~b<UOJ

20~h<004

ulNflQ‘fl"

ZO~8<UQJ

Whaandx

"flabamuflhmno
"mam afloaamz<mh
«hark flJU—=u>

.Hlm mdnme

108

huuoo~e> IIIIIII
~13a0<

~suoo
.nlsz

Ov-‘Nnfilfltohma

dawn
3630

0666.66
966—.06
0vm0.6
6706.6
66v0.9
6690.6
N9V9.9
Nod—.6
6066.6
$666.66
«067.6
mem_.6
9666.0
men—.6
6666.66
9660.6
06—N.—~
6676.6
~6v6.9
6766.6
~06v.6
men—.6
v~vv.9
awn—.6
mwvv.6

:oquaaouauu<

p—oN«Nunfl—no—oNN—o—o-‘mN—nn—n—O—n—I—o

upcoooauqal
codauhsn

soogm

0660.6
6600.0
6606.0
6606.6
6606.6
9606.0
0666.0
6660.6
6660.6
6606.6
6666.6
6666.6
6666.0
6660.6
6606.6
6606.6
6666.0
6666.6
v~60.0
6606.0
9666.0
v—60.6
9660.6
v—06.6
6960.6

aucooea
sequeasa
soozm

0096.6
.6696.0
6696.6
0666.6
0690.0
06v6.6
6696.6
6096.6
6666.6
6690.6
6696.6
0096.6
0660.0
6006.6
6696.6
v096.6
6690.6
6666.6
6090.6
6090.6
6066.6
6696.6
6060.0
6696.0
6066.6

6066.6
6666.6
6666.6
6606.6
6060.6
6666.~
6600.6
6066.6
6666.6
6666.6
6666.6
6006.6
v666.u
6600.6
6666.—
6666.6
6000.—
6660.6
6606.—
6600.—
6006.6
6660.6
v600.~
6060."
6660."

agenda
caduuoaaoo

0a
60
66
66
69
66
69
06
an
66
66
06
66
66
66
06
66
66
99
06
66
06
69
66
v6

 

a 90.0 66 06 6
n 90.0 66 06 6
90.0 66 06 6
60.0 66 06 6
90.0 66 06 6
v0.0 66 06 6
90.0 16 06 6
90.0 V6 ON 6
6.0 66 06 6
a 90.0 N6 06 6
90.0 66 06 6
90.0 «6 06 6
60.0 «6 06 6
«.0 «6 06 6
m 90.0 66 06 6
90.0 66 06 6
a 96.0 66 06 6
60.0 «N 06 6
90.0 a 06 6
90.0 a 06 6
6.0 0 66 6
90.0 0 06 6
60.0 06 06 6
90.0 66 an 6
66.0 66 0— 6 a floauduOJ
Noe-xi huuoo~e> : 6 9 sanded-cu
couueu u 6 ¢
oneoo< 8 6
.#003m mumv Xbosm .Nlm OHQMB

BIBLICXERAPHY

BIBLIOGRAPHY

American President Company.

1987

American Society for Testing and Materials (ASTM).

1987

American Society for Testing and Materials (ASTM).

1987

"APC Dramatizes
1986

Ride Characteristics: APC Stack Train vs.
TOFC and Truck. July, 1987. H

ASTM D-3332 Standard Test Method for
Mechanical-Shock Fragility of Products
Using Shock Machines. Annua Book 0 A TM
Standards, Vol. 15.09. March, pp. 573-580.

...tmu‘. ‘

' m“.
‘0‘.

ASTM D-4728 Standard Test Method for
Random Vibration Testing of Shipping

Containers. Annual Book of ASTM Spandards,
V01. 150090 MarChg pp. 1160-11670

Stack Train’s Smooth Ride by Carrying
Table-for-Two With China." Enpific
Shipper. July 14.

Brandenburg, Richard and Julian Lee.

1985

Byrne, R. and J.

1976

Flint, J.
1986

Foley, J.T.
1972

Grimm, C.M. and
1986

Fundamenta o acka in amic . MTS
Systems Corp. Minneapolis, Minnesota.

A. Anderson.

Performance Characteristics of Freight Car
Trucks Determined Through Road Testing.
ASME Paper No. 76-WA/RT-4. December.

Here Come the Truckbusters. Fopbes.
June 30. pp. 86-90.

Transportation Shock and Vibration
Descriptions for Package Designers. Sandia
Laboratories Report SC-M-72 0076. July.

K.G. Smith.

The Impact of Rail Regulatory Reform on
Rates, Service Quality, and Management
Performance: A Shipper Perspective.

Lpgis inn & Irnngportntion Rgview. Canada,
MarCh, Pp. 57-68.

109

Grosso, Harry.
1987

Guins, S.G. and
1974

Hanes, R.S.
1983

110

Personal interview during transportation
seminar at the Rochester Institute of
Technology, Rochester, New York, May 1987.

L.L. Olson.
5,000 Mile Boxcar Vibration Test. ASME
Winter Annual Meeting. November.

Recording and Analysis of Truck Vibration.
Technical Report TA 77.0126. IBM.
September 30.

Harris, C.M. and C.E. Crede.

1976

Johnson, E.E.
1971

Kane, J.T.
1985

Kenworthy, M.
1979

Kusza, T. and W.

1973

Lee, N.W.
1969

Naylor, R.C.
1987

Newton, R.F.
1976

Shock and Vibration Handbook. McGraw-Hill.

 

The Shock and Vibration Environment of
Packages Carried on a 3-Ton General Purpose
Truck. Shock Environment of Packages in
Transport--Society of Environmental
Engineers. London, England. Feb., p. 27.

Accepting the Double Dare. H&SM. Nov.,
pp. 55-590

Transportation of Vibration Sensitive
Equipment by Highway Trailer on an Inter-
modal Railcar - Volume I. U.S. Department
of Transportation. Final Report. July.

Sharpe.

Preliminary Measurement and Analysis of the
Vibration Environment of Common Motor
Carriers. MSU School of Packaging,
Technical Report No. 20. September 1.

Glassware Breakage as a Measure of the

Transport Environment. 3rd Symposium of
Environmental Engineers. London, England.
April 15-18.

Canadian Rail Environment. Annual Safe

Transit Conference. Toronto, Canada.

Fpngility Assessment, Theory and Tent
Procedure. MTS Systems Corp. Minneapolis,
Minnesota.

111

Nordstrom, J., T. Trost, and S. Cavlin.
1986 Computer Aided Transport Simulation --
CATS. 5th IAPRI Conference. Papers -
Volume III. Bristol, England. pp. 1-20.

Ostrem, F.E. and W.D. Godshall.
1979 An Assessment of the Common Carrier
Shipping Environment. Forest Products
Laboratory, U.S. Department of
Agriculture, FPL 22.

Schlue, J.W.

1966 The Dynamic Environment of Spacecraft
Surface Transportation. Jet Propulsion
Laboratory. Technical Report No. 32-876.
March.

Schlue, J.W. and W.D. Phelps.

1968 A New Look at Transportation Vibration
Statistics. The Shock and Vibration
Bulletin, No. 37, Part 7. Jan.

Sharpe, W.

1972 Preliminary Investigation of Freight Car
Vibration. MSU School of Packaging,
Technical Report No. 20. Nov.

Silvers II, W. and H. Caruso

1976 Advances in Shipping Damage Prevention.

Shock and Vibration Bnlletin, No. 46,
Part 4. August.

Soroka, Walter.
1986 Personal interview conducted at the
Ontario Research Foundation, Toronto,
Ontario, Canada, February, 1986.

Wesselman, C.

1986 The Greatest Potential... Modern Railroads.
July.

GENERAL REFERENCES

GENERAL REFERENCES

Andrews, E.A.
1981 Generation of a Transportation Vibration
Test Specification. Scientific Atlanta,
Spectral Dynamics Division. DRC-009.

November.
Borzo, Gary.
1986 Are Arguments Against Concrete Ties
Crumbling? Modern Railroads. July.
pp. 19-23.

Burgess, Gary.
1985 Combined Shocks and Vibration...Bumps in
the Road. 4th International Conference on
Packaging, Reference Abstract #144.

September.
Foley, J.T., M.B. Gens, and C.F. Magnuson.

1972 Current Predictive Models of the Dynamic
Environment of Transportation. Proceedings
of the Institute for Environmental Sciences
pp. 35-44.

F d ental of Moda e tin .

1986 Applications Note 243-3, Hewlett-Packard

Co. May.

Gibson, G.L. and D. Passmore.
1970 A Study of the Efficiency of Various
Combinations of Air and Standard Steel
Spring Equipment in Reducing Cargo Damage
While Operating Over a Normal Highway.

New Way Air Ride Report. Neway Equipment
Co.

Greenfield, L.P. and R. N. Hodges.

1975 An Investigation of Container Flat Car Ride
Quality. ASME Paper No. 75-WA/RT-7. July.

Hammond, John.

1986 Industry Offered a New Rail-Road. Democrat
& Chronicle. Rochester, New York. p. 63.

LaCroix, L.

Luebke, R.W.
1970

Luebke, R.W.
1971

Soroka, W.
1986

Tustin, W.
1967

113

Spectrum Analysis for Fun and Profit.
Nicolet Scientific Corp. (no date).

The Box Car Dynamic Environment. The

Shock and Vipration Bulletin. Bulletin 41.

December..

Freight Transportation Vibration
Environment. Joint ASCE-ASME Transporta-
tion Engineering Meeting. July.

Rail Shock and Vibration. Ontario Research
Foundation, File No. ES/EPD -85-20.
February 18.

Vibration and Shock Tests Do Not Duplicate
the Service Environment. Test Engineering.
Aug. pp. 18-21.