Professional Documents
Culture Documents
of Impact Loading on a
Three Spoke Steering Wheel
BY
William Aitenhof
1999
1*1 National Library
of Canada
Bibliothèque nationale
du Canada
Acquisitions and Acquisitions et
Bibliographie Services services bibliographiques
395 Wellington Street 395. rue Wellington
Ottawa ON K1A ON4 OnawaON K l A û N 4
canada Canada
The author has granted a non- L'auteur a accordé une licence non
exclusive licence allowing the exciusive permettant à la
National Library of Canada to Bibliothèque nationale du Canada de
reproduce, loan, distribute or sel1 reproduire, prêter, distribuer ou
copies of this thesis in microform, vendre des copies de cette thèse sous
paper or electronic formats. la fonne de microfiche/film, de
reproduction sur papier ou sur format
électronique.
The focus of this dissertation is based upon two objectives. Firstly, experimental
tests were conducted with collisions between a three spoke steering wheel and a
deformable chestform. A large amount of testing was completed which examined the
effects of variations in the column angle and wheel angle of the steering wheel, as well
as the velocity of the chestform as it collided with the steering wheel. Secondly, detailed
finite element models of the testing apparatus were developed, and simulated under
identical testing conditions to see if explicit finite element simulations may be used to
predict the performance of the three spoke steering wheel in a collision type situation.
Through an analysis of the experimental data, it was observed that the calculated
energy absorbed during the impact process by the steering wheel and deformable
chestform depended significantly upon the column angle, wheel angle, and impact
velocity. Surface contours have been developed which illustrate the effect of the three
independent variables on the peak impact load, the peak impact displacement, and the
calculated energy absorption. Furthemore, a met hod to quanti the energy absorption
Explicit finite element simulations were conducted for twelve different testing
testing was found, indicating that finite element modeling may be used as a predictive
To an understanding and loving family, without which I would never have been able to
complete this work, to al1 of them, even those not with us now, al1 my love. You have
been, and will continue to be one of the best ways of attaining knowledge.
The author would like to take this opportunity to express his deepest gratitude and
appreciation to Dr. Nader Zamani, and Dr. Robert Gaspar. With the aid of these two fine
professors the author has grown intellectually, and attstined talents which \vil1 better his
life. Thank you for your time, efforts, guidance, recomrnendations, and encouragement
staff at K.S. Centoco, especially, Mr. Saverio Paonessa. You have al1 provided
ABSTRACT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iv
ACKNOWLEDGMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vi
LISTOFFIGURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xii
LISTOFTABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . u
NOMENCLATURE . . . . . . .. . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . mii
.. *
CONVERSION FACTORS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxrii
1 . INTRODUCTION . . . . . . . . * . . . . . . . . . . . . . . - . . . . . . . - - . - - . - - - . . . . . . . . . .1 - .
2. LITERATUREREVIEW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.1. Documentation Regarding Experimental Testing . . . . . . . . . . . . . . . . . . . . . . 3
2.1.1. Existing Relevant Standards and Accident Statistics . . . . . . . . . . . . . . . . 4
3.1.2. Testing Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2 - 1 3 The Prototype Testfng Machine . . . . . . , . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.2. Other Considerations in Experimental Testing . . . . . . . . . . . . . . . . . . . . . . 10
23. Documentation Regarding Finite Element Mode1ing and Experimental
Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1
2.3.1. Simulation of Body Block Steering Wheel Impacts Using Dyna3d . . . . 12
2.3.2. Designing Energy Absorbing Steering Wheels Through Finite Element
Impact Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.3.3. An Investigation of Steering Column Collapse Behavior Using Finite
Element Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.4. Documentation Regarding Crashworthiness Testing o f Automobiles and
Associated Vehicle Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.5. Documentation Regarding Occupant Modeling and Crash Test Dummies . . 16
2-6. SumrnaryofPastResearch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
4 . EXPERIMENTAL TESTING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . - 2 1
4.1. Terminology Associated with Steenng Wheel Testing . . . . . . . . . . . . . . . . . 21
4.1.1. The Column Angle and Impact Velocity . . . . . . . . . . . . . . . . . . . . . . . . - 2 1
4.1.2. The Wheel Angle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . - 2 2
4.1.3. Peak Impact Load, Peak Impact Displacement, and Calculated Energy
......................................................... 23
4.2. The Steering Wheel Selected for Experimental Testing . . . . . . . . . . . . . . . . 24
4.3. A Background Discussion of the Experimental Testing Conditions . . . . . . . 24
4.4. The Experimental Testing Machine Used in this Investigation . . . . . . . . . . . 27
4.4.1. A Basic Description of the Mechanisms in the Testing Machine . . . . . . 27
4.4.2 The Data Acquired During an Impact Test . . . . . . . . . . . . . . . . . . . . . . 30
4.4.3 Analysis of the Acquired Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
1.4.4 The Experimental Testing Procedure Used in this Investigation . . . . . 34
4.4.4.1 Checking on the Integrity of The Chestfonn . . . . . . . . . . . . . . . . . . 34
4.4.4.2 Experimentai Testing Procedure Used for Impacts on Steering Wheels
................................,.............. . 36
5.6.9. The Effect of Column Angle and Wheel Angle on the Coefficient of
Restitution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
5.6.10. Comments on the Variation of the Coefficient of Restitution . . . . . . . . . 75
5.6.1 1. The Effect of Column Angle and WheeI Angle on the Peak Impact Force
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
5.6.12. Cornments on the Variation of the Peak Impact Force . . . . . . . . . . . . . . 85
5.7. Data Acquired and Calculated for Each Experimental Test Conducted . . . 85
REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
APPENDIX B . Data Acquired and Calculated for Each Experimental Test Conducted
..................................................... 168
APPENDIX C . Material Properties and Sample Input Data for the Aluminum
Crosshead . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224
APPENDIX D . Material Properties and Sample Input Data for the Wooden Backing
Plate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226
APPENDIX E . Material Properties and Sample Input Data for the Low Density
PolyurethaneFoam . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228
TABLE OF CONTENTS .CONTLNUED
APPENDiX F . Material Properties and Sample Input Data for the Steenng Wheel
A!uminumAlloy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
APPENDiX G . Sample Daia From the LS-DYNA Input File Regarding Hourglass
Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232
APPENDIX H . Sample Data From the LS-DYNA Input File Regarding Contact and
MassScaiing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 334
VITAAUCTORIS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236
LIST OF FIGURES
Figure1 . Basic apparatus and setup for the SAE 5944 testing procedure . . . . . . . . . 5
Figure 2 . Droptower testing machine .................................... 8
Figure 6 . Typical crosshead displacement versus time curves from the Mathcad
worksheet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . - 3 1
Figure 7 . Typical "z-axis" load versus time curves from the Mathcad worksheet
Figure 9 . Experimental setup used to test the integrity of the chestform over the
course of the 420 impact tests considered . . . . . . . . . . . . . . . . . . . . . . . - 3 4
Figure 10 . Experimental test setup for a typical "chestform to steering wheel" impact
test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
Figure 1 I . Deformed geometry of the steering wheel and chestform after an impact
test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
Figure 14 . [Ilustration of the total energy the dropping assembly has at any point
during the experimental impact test . . . . . . . . . . . . . . . . . . . . . . . . . . . . - 4 3
Figure 1 5 . Typical crosshead velocity versus time curve illustrating the impact
velocity and rebound velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . - 4 9
xii
LIST OF FIGURES CONTINUED -
Figure 16 - Wheel and column angles versus peak vertical load at 8 rnph [3.58 m/s]
impact velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . - 5 2
Figure 17 - Wheel and column angles versus peak vertical load at 9 rnph 14-02 m/s]
impactvelocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
Figure 18 - Wheel and column angles versus peak vertical load at 10 rnph [4.47 m/s]
impactvelocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
Figure 19 - Wheel and column angles versus peak vertical Ioad ai 1 1 rnph C4.92 m/s]
impact velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.. 53
Figure 20 - Wheel and colurnn angles versus peak vertical load at 12 rnph C5.36 m/s]
impact velocity . . . . . . . . . . . . . .
... . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
Figure 21 - Wheel and column angles versus peak vertical load at 13 rnph [5.8 1 m/s]
impact velocity . . . . . . . . . . . . . . . . . . . . . . . . . .
.. . . . . . . . . . . . . . . . . . 54
Figure 22 - Wheel and colurnn angles versus peak vertical displacement at 8 rnph
f3.58 m/s] impact velocity . . . . . . . . . . . . . . . . . . . . . .
.. . . . . . . . . . . . - 5 7
Figure 23 - Wheel and colurnn angles versus peak vertical displacement at 9 rnph
[4.02 d s ] impact velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
Figure 24 - Wheel and column angles versus peak vertical displacement at 10 rnph
[4.47m/s] impact velocity . . . . . . . . . . . . . . . . . . . . . . .
.. . . . . . . . . . . . 5 8
Figure 25 - Wheel and column angles versus peak vertical displacement at 1 1 rnph
r4.92 d s ] impact velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . - 5 8
Figure 26 - Wheel and column angles versus peak vertical displacement at 12 rnph
15-36m/s] impact velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 9
Figure 27 - Wheel and column angles versus peak vertical displacement at 13 rnph
[5.8 1 d s ] impact velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 9
Figure 28 - Wheel and column angles versus peak calculated total energy absorbed at
8 mph l3.58 d s ] impact velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . - 61
..-
Xlll
LIST OF FIGURES CONTINUED -
Figure 29 - Wheel and column angles versus peak calculated total energy absorbed at
9 mph [4.02 d s ] impact velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 1
Figure 30 - Wheel and column angles versus peak calculated total energy absorbed at
10 mph [4.47 m/s] impact velocity . . . . . . . . . . . . . . . . . . . . .
.. . . . . . . 6 2
Figure 31 - Wheel and column angles versus peak calculated total energy absorbed at
1 1 mph [4.92 m/s] impact velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . - 6 2
Figure 32 - Wheel and column angles versus peak calculated total energy absorbed at
12 mph r5.36 m/s] impact velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . - 6 3
Figure 33 - Wheel and column angles venus peak calculated total energy absorbed at
13 mph [5.81 mls] impact velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 3
Figure 34 - WheeI and column angles versus the average energy absorption factor
(E.A.F.) at 8 mph [338 m/s] impact velocity . . . . . . . . . . . . . . . . . . . . . 6 6
Figure 35 - Wheel and colurnn angles versus the average energy absorption factor
(E.A.F.) at 9 mph [4.02 d s ] impact velocity . . . . . . . . . . . . . . . . . . . . - 6 6
Figure 36 - Wheel and column angles venus the average energy absorption factor
(E.A.F.)at 10 mph [4.47 m/s] impact velocity . . . . . . . . . . . . . . . . . . . . 67
Figure 37 - Wheel and coiumn angles versus the average energy absorption factor
(E.A.F.) at 1 1 mph l4.92 m/s] impact velocity . . . . . . . . . . . . . . . . . . . - 6 7
Figure 38 - Wheel and column angles versus the average energy absorption factor
(E.A.F.) at 12 mph [5.36 m/s] impact velocity . . . . . . . . . . . . . . . . . . . - 6 8
Figure 39 - Wheel and column angles venus the average energy absorption factor
(E.A.F.) at 13 mph f5.81m/s] impact velocity . . . . . . . . . . . . . . . . . . . - 6 8
Figure 40 - Wheel and column angles versus the coefficient of restitution at 8 rnph
[3-58 m/s] impact velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . - 7 2
Figure41 - Wheel and column angles versus the coefficient of restitution at 9 rnph
r4.02 m/s] impact velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
xiv
LIST OF FIGURES CONTINUED -
Figure 42 - Wheel and colurnn angles versus the coefficient of restitution at 10 rnph
r4.47 m/s] impact velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
Figure43 - Wheel and column angles versus the coefficient of restitution at 1 1 rnph
[4.92m/s] impact veloçity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
Figure 44 - Wheel and column angles versus the coefficient of restitution at 12 rnph
[5.36 m/s] impact velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
Figure 45 - Wheel and column angles versus the coefficient of restitution at 13 rnph
[S. 8 1 m/s] impact velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
Figure 46 - The coefficient of restitution versus the average energy absorption factor
for 8 mph 13-58 d s ] impact velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . -77
Figure 47 - The coeficient of restitution versus the average energy absorption factor
for 9 mph 14-02 mk] impact velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
Figure 48 - The coefficient of restitution versus the average energy absorption factor
for 10 mph 14-47 m/s] impact velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 8
Figure 49 - The coefficient of restitution versus the average energy absorption factor
for 1 1 mph [4.92 d s ] impact velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 8
Figure 50 - The coefficient of restitution versus the average energy absorption factor
for 12 mph [5.36 m/s] impact velocity . . . . . . . . . . . . . . . . . . . . . . . . . . - 7 9
Figure 51 - The coefficient of restitution versus the average energy absorption factor
for 13 mph r5.81 m/sJ impact velocity . . . . . . . . . . . . . . . . . . . . . . . . .
.-79
Figure 52 - Wheel and column angles versus the peak impact force at 8 rnph
[3S8 mis] impact velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
Figure 53 - Wheel and column angles versus the peak impact force at 9 rnph
E4.02 m/s] impact velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . - 8 2
LIST OF FIGURES .CONTINUED
Figure 54 . Wheel and column angles versus the peak impact force at 10 mph
t4.47 m/s] impact velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
Figure 55 . Wheel and column angles versus the peak impact force at 1 1 mph
[4.92 mis] impact velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
Figure 56 . Wheel and column angles versus the peak impact force at 12 mph
f5.36mis] impact velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
Figure 57 . Wheel and colurnn angles versus the peak impact force at 13 mph
[5.81 m/s] impact velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
Figure 59 . Finite element models of the wwden backing plate and the aluminum
crosshead . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . - 9 1
Figure 60 . The cross-sections used in development of the finite element model of the
chestfom . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . - 9 2
Figure 63 . Engineering stress versus engineering strain curve for the polyurethane
chestform material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
Figure 64 . Finite element mode1 of the three spoke steering wheel . . . . . . . . . . . . 100
Figure 65 . Stress versus effective plastic strain curve for the aluminum alloy of the
three spoke steenng wheel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
Figure 66 . Expenmental setup used to ver@ the finite element model of the
chestforni . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
Figure 67 . Finite element models for the chestform. wooden backing plate.
crosshead. and n'gid cylinder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
LIST OF FIGURES CONTLNUED -
Figure 69 - Cornparisons between the two modified finite element models and the
original finite element mode1 . .
.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 10
Figure 70 - Experimental setup used to ver@ the finite element model of the three
spoke steering wheel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I 12
Figure 71 - Finite element models of the three spoke steering wheel and the rigid plate
used to veriQ the model of the steering wheel . . . . . . . . . . . . . . . . . . . 1 14
Figure 78 - Al1 four flnite element models used in simulating the chestfoxm impacting
the deformable steering wheel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
Figure 79 - Three orthogonal axes used to il l ustrate the testing methodology for the
first seven numerical simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
xvi i
LIST OF FIGURES .CONTLNUED
Figure 81 . Deformation of the chestform finite element model with hourglass control
impIemented. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
Figure 82 . Contact without the proper values of the penalty factors for the chestfonn
and the steering wheel . . . . . . . . . . . . . . . . . .
.. . . . . . . . . . . . . . . . . . . 132
Figure 83 . Contact between the chestfonn and the steering wheel with the proper
values of the scale factors for the SURFACE-TO-SURFACE contact
algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
Figure 84 . Geometry of al1 the finite element rnodels before impact between the
chestfonn and the three spoke steering wheel . . . . . . .
.. . . . . . . . . . . . 134
Figure 85 . Geometry of al1 the finite element models after impact . . . . . . . . . . . . 134
Figure 87 . Experimental and numerical findings for simulation test if2 . . . . . . . . . 137
xvi ii
LIST OF FIGURES .CONTINUED
Figure 99 . Percentage of total interna1 energy absorbed by the three spoke steering
wheelandchestfonn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
xix
LIST OF TABLES
Table 1 Crosshead dropping heights required for the impact velocities considered
inthisstudy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
Table 5 Expenmental testing conditions considered for the ngid plate impacting
the three spoke steering wheel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 13
Table 6 The details of the five different finite element models considered at
experimental testing situation X 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 15
Table 7 Approximate computational times required for the five finite element
models considered for the three spoke steering wheel . . . . . . . . . . . . . 1 18
Table 10 Testing parameters for the twelve numerical simulations conducted with
the corresponding Figure # illustrating the cornparison between numerical
andexperimental findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
Table 1 1 Percentage difference of the peak vertical load and peak crosshead
displacement fiom numerical and experimental results . . . . . . . . . . . . 144
Table 12 Intemal energies and calculated energies from finite element simulations
(not considering strain rate effects) and expenmental tests . . . . . . . . . 149
Table 13 Intemal energies and calculated energies from finite element simulations
(consideringstrain rate effects) and experimental tests . . . . . . . . . . . . 150
LIST OF TABLES .CONTINUED
Table 14 . The true E.A.F. calculated based upon the interna1 energy within just the
steering wheel and the expenmentally detemined average E.A.F. . . . . 151
acceleration (length/time2)
xxiv
1. INTRODUCTION
automobiles which are light weight, fuel efficient, confonn to a level of safety outlined
by government regulations, and are available to the consumer at a reasonable cost. The
automobile industry has placed a significant amount of time and research funding into
faced with the requirement of meeting several more govemmental regulations and
corporations have developed their own "in-house" safety criteria which also must be
The steering wheel represents the most dangerous vehicle component for the
responsible for the loss of many lives in collision situations between the steering wheel
and the driver. Automotive manufacturers realize this problem and have devoted a
significant amount of time to the design and development of various safety components.
The seat belt and the driver's side air bag represent only a few of many engineering
accomplishments in the effort to minimize impact situations between the driver and the
automobile corporations found a method of testing proposed designs and simulating their
products before fabrication. This represented a significant advancement in the
engineering design phase of vehicle components, and significantly decreased the costs of
see how the changes would affect its performance, without actually performing costly
develop steering wheels which rnaximize the absorption of energy in an impact situation.
chestform and any of the steering wheels K.S. Centoco manufactures. The key feature of
the machine is its ability to investigate several independent variables, such as location of
impact on the steenng wheel, and observe how well the steering wheel mechanically
behaves in such a situation. Characteristics such as peak impact load between the
chestform and steering wheel, peak impact displacement of the steering wheel, and
e n e r a absorbed by the two cornponents are very important in the design of a "safe"
steering wheel.
Finite element simulations are also an integral part of steering wheel design.
Dynamic simulations using explicit finite element software have onty become a reality in
the past three to five years with the advent of faster personal computers. In addition,
only a handful of software companies are developing explicit finite element code. The
importantly, the experimental testing machine shall be utilized to examine the effects of
location of impact between the chestform and the steering wheel, as well as, the impact
velocity of the chestform ont0 the steering wheel and observe how these variables
influence the performance of a steering wheel during collision. Secondly, finite element
(FE) models of the experimental testing equiprnent sha1l be developed, and simulated
using explicit finite element code (LS-DYNA), and compared with experimental testing
steering wheels is to aIlow engineers and designers to predict how the steering wheel will
perform in an actual car crash situation. It is much more feasible to test a single steering
wheel in a destructive test than to perfonn a full car crash test, which unfortunately
destroys the entire vehicle. The author has completed prior work with the University of
Windsor and K.S.Centoco and has developed and fabricated a testing machine for
Reference [ I l dealt with three specific areas of concern in the design and
deveiopment of the testing machine. Firstly, a background search into past standards
associated with impact testing of steering wheels was performed. Secondly, a testing
was described. Finally, the development and manufacturing process of the machine was
detailed.
The development of the testing machine was based primarily upon two very
significant past and current testing standards; the Society of Automotive Engineers -
Steering Control System - Passenger Car - Laboratory Test Procedure SAE 5944 [SI
(which was canceled and not replaced) and a very similar European testing standard,
Annex III to Directive 74/297/EEC of the European Economic Community [3] (which is
cunently in use).
SAE 5944 involves experimentally testing the entire steering assembly and not
just the steering wheel. Figure 1 illustrates the testing apparatus and test setup.
deformabie
chestform or
DIRECTION O F LOAD CELL
bodyform. The LOCATION
SAE J944. The Figure 1 - Basic apparatus and setup for the SAE J944 testing
procedure.
standard is also
specific to the type of steering assembly being tested; the location of the chestfonn and
the column angle (as illustrated in Figure 1 ) are determined from the vehicle into which
Some details of the 5944 standard are vague. Fintly, the standard does not detail
the orientation of the steering wheel itself (the angle that the wheel is being tumed).
Secondly, and most importantly, this testing procedure does not specify a required impact
velocity.
The European standard, Annex UI to Directive 74/297/EEC, is much more
elaborate than SAE 5944. Although very similar to 5944, it considers several situations
which will significantly alter the loading profile between the bodyform and the steering
assembly. Firstly, the testing standard requires that the steering wheel be tested at the
most rigid and most weak positions on the wheel. This requirement considers the effect
of the orientation of the steering wheel. Secondly, if the steering system which is k i n g
tested utilizes an airbag, then the testing procedure must be conducted with the airbag
device inflated. Finally, this testing procedure places a requirement on the impact
velocity of the chestform on the steering wheel to be 15.0 * 0.7 mph [6.7 1 0.3 1 m/s].
with collisions between a driver and the steering wheel. Moms et al. 141 have found that
collisions between the driver's chest and the steering assembly represent the greatest
percentage of accident data. The second greatest percentage of accident data involved
Cohen [ 5 ] , has determined that the steering assembly is also responsible for the
most serious injuries inflicted upon the driver in an accident. Collisions between the
chest of the driver and the steering assembly represent the greatest percentage of serious
injuries caused by the assembly. The second greatest percentage of serious injuries is
due to interactions between the driver's abdomen and the steering assembly, similar to
methodology for impact testing of steenng wheels to simulate real world accident
situations. Based upon past accident statistics, it is very difficult or impossible to predict
the exact location of an impact between the driver and the steering assembly. Accident
statistics do indicate a probability of impact between the steering wheel and either the
driver's chest or the driver's abdomen. However, the orientation of the steering wheel
and location of impact on the steering wheel cannot be predicted. This is primarily
because al1 drivers are different in geometry and weight and al ter the steering assembly
(tilt steering) and seating position to the orientation they most prefer.
Altenhof [1] has developed a steering wheel testing procedure which shows that
in order to be able to predict steering wheel performance in an actua1, real world, impact
situation the experimental testing must investigate al1 the parameters associated with
velocity, location of impact (between the chestform and steering wheel), and the steering
wheel orientation to be able to simulate the most likely real world collision.
prototype testing machine. This machine is illustrated in Figure 2. The device uses
gavity to accelerate the chestform and provide an impact velocity to the rigidly mounted
steering wheel. The chestform is allowed only to translate in the vertical direction.
A triaxial load cell is used to determine the impact force acting between the
chestfom and the steering wheel. As well, a linear variable differential transformer
displacements of the
test-
LVDT
One axis of the load
experimental arnount of
Figure 2 - Droptower testing machine.
-8-
energy which was transferred to the steering wheel and chestform in an impact. The
where, 'E' represents the amount of energy transferred to the steering wheel and
chestform, 'F' represents the force acting in the interface of the steering wheel and
displacement at the beginning of the test, and 'x,' represents the final displacement
Only a finite number of data points are COllected by the testing mach ine so
where, Zload, represents the z axis load for the ith data point, Ivdt,., represents the LVDT
displacement measurement for the i+ 1 data p i n f and Ivdt,., represents the LVDT
The testing machine also incorporates three accelerometers, one located at the
head of the chestform, another located at the lett shoulder of the chestfonn, and a third
located at the lower right abdomen area. These devices are used to determine how the
acceleration in an impact test varies over the length of the bodyfom (from the head of
wheels, employing the SAE J944 standard or equivalent, were found. In addition, no
documentation regarding data, or the analysis of data, was found, based upon the 5944
However, experimental testing and data has been documented for droptower
testing of steering wheels from one of K.S. Centoco's clientele [6] and has k e n
provided as a source of reference. The testing procedure and testing equipment, from
K.S. Centoco's clientele, does not follow the 5944 standard. The chestfonn that is used
in this expriment is a rigid dome-like structure which does not deform, or deforms to a
the testing machine developed by Altenhof [ I l . Energy is calculated based upon the area
wheels has been completed and documented. Unfortunately, these types of experiments
focus on peak accelerations (or decelerations) which the head may see in an impact
situation and do not focus on the energy absorption characteristics of the steering wheel.
As well, full automobile testing has also been completed and documented. Yet this
information is proprietary to the Company which has developed the automobile. Similar
to the headforrn testing, full automobile tests do not focus on the energy absorption
has been demonstrated by Culver [7].Large amounts of data, acquired fiom typical
"field test" crash testing (completed on an entire automobile), was analyzed and
developed by Culver and used to determine the impact force, energy absorbed, and
Documentation regarding current and future trends for steering wheel matenals
and safety concems has been completed by Rodrigues et al. [8]. An analysis of the
geometry and surface hardness of the steenng wheel and its related compnents was
There are a number of documents which address finite element modeling and
esperimental testing of impacts between a chestform and the steering assembly. For the
experimental, the J944 standard, or its equivalent, is used as the basis for a testing
procedure.
2.3.1. Simulation o f Body Block Steering Wheei Impacts Using Dyna3d [9]
McKie 191 has cornpleted work using Dyna3d, an explicit finite element software
program (now named LS-DYNA), to simulate a European standard similar to SAE 5944.
The chestform utilized in this standard is identicat to the deformabte chestform used in
5944. As well, experimental testing was completed and used to compare results with
The method used by McKie was sornewhat different from the testing procedure
chestform, and simulated a spherical dome impacting the chest of the bodyform.
Experiments were completed at the same conditions and a good relationship between
experimental testing and numerical simulations was found for the velocity of the
McKie then developed a full FE model of the steering wheel which was
considered in the analysis. However, the testing procedure used for impacts on the
steerinç wheel significantly differed from that required by the European standard.
McKie replaced the deformable chestform with a rigid, non-deformable, chestform, and
is based upon peak impact load, and not energy absorption characteristics of the steering
wheel. Although experimental testing was completed using a deformable chestform, and
numerical simulations were completed using a rigid chestform, the results indicated a
characteristics of a steering wheel in an impact situation. The SAE J944 standard was
utilized as the procedure for conducting both numerical and experimental tests. It should
be noted that this documentation is very similar to McKie [9] with the exception that
energy is used as the bais for determining the performance of the steering wheel (in an
impact situation) rather than peak impact load. As well, a modified steenng wheel
The procedure implemented by Shyu et al. was identical to the approach outlined
in SAE J944 wïth the exception that a ngid chestform, instead of the required deformable
chestform, \vas used in the numerical simulations. Similar to McKie [9], Shyu et al. have
A very good relationship, in terms of impact toad versus time during impact, was
situations. Although this document focuses primarily on the steering column, and not the
steering wheel, it does provide some information regarding finite element modeling of
the apparatus used in SAE 5944. The 5944 testing rnethod was utilized in this
investigation.
Similar to the other studies, Gotoh et al. have chosen to use a rigid chestform as
the impacting entity on the steering assembly. However, the FE chestform modeled in
this investigation is very crude compared to FE models from the other references. The
mode1 does not conform to the geometry or shape of the chestform required by the 5944
standard.
As previously mentioned, this work dealt with steering columns (rather than
steering wheels) and unfortunately it did not present any data or analysis of data
regarding the steering wheel. However, Gotoh et al. have found that a good relationship
between experimental testing and numencal simulations was evident on the bais of the
crashworthiness has been reviewed for this dissertation. Omar et al. [ 121, Pries et
al. [ 133, Bennett et al. [14], Saha et al. [ l 51, and Scott Lui [16] have al1 completed
extensive work in the field of crashworthiness. These references did not focus on the
steering wheel but discussed the general principal of crashworthiness for automotive
design and addressed the use of explicit finite element software for the numerical
techniques and new element formulations are three areas of concem which have been
mathematically developed and are only now, in present day, being developed and
Rowberry [19], and Fan [20], have each addressed crashworthiness from an energy
which undergo massive deformation in a collision are a significant factor in the energy
testing is rollover. Rollover is defined as any maneuver in which the vehicle rotates
ninety degrees or more about its longitudinal axis such that the body makes contact with
the ground. Literature regarding rollover and the cause of this situation has been
docurnented by Gillespie [2 1 ] and Chou et al. [22]. Mathematical models for the quasi-
static and transient rollover of vehicles has been reviewed. Furthemore, Bathe [23],has
discussed the proper procedure for the finite element modeling of rollover simulations.
Bathe believes that quasi-static rollover simulations are best conducted using implicit
finite element software (or implicit integration techniques) due to the fact that unrealistic
artificial oscillations become present in the solution deterrnined from explicit integration
methods.
A literature review regarding occupant modeling and crash test dummies was
impact testing conducted on actual human cadavers and has cornpared results, in terms of
impact force and displacement, for males and females without any relationship found
between the two genders. This information was utilized for recommendations in the
testing of the Hybrid III crash test dummy has been reviewed [25-301. In general, a very
conducted on the entire Hybrid III body, and more localized testing, conducted on just
the head and neck structure of the Hybrid III, was observed. In addition, experimental
test cornparisons between older crash testing dummies (Fart 572) and the Hybrid III body
was conducted with a good correlation found between the two entities.
In an effort to determine the forces and pressures which act on the face of the
human body during a coHision between the head and the steering wheel Warner et
al. 1311 developed a load sensing faceform. This faceform contained extensive
instrumentation and high speed data acquisition capabilities which proved to be a very
significant advancement in experimental testing for collisions between a facefom and
2.6. SummaryofPastResearch
Based upon the literatue reviewed, several observations can be made conceming
experimental testing or flnite element simulations. SAE 5944 does not place a
wheel in an impact test. Furthemore, the column angle is specified as per the
âctual column angle within the automobile. This does not consider the option
of tilt steering assemblies, which are oflen a standard option in the majority of
automobiles. European standards are more robust in the fact that they do
an automobile will hit the steering wheel at a speed of 15.0 mph [6.71 rn/s].
European standards do address the weakest and stiffest part of the steering
The first observation Iisted above, indicates that even if experimentd testing is
completed, there exists a very good chance that this information may not be usefui in an
actual real world collision situation. Since, there is a very low probability that the dnver
will impact at a speed of 15.0 mph [6.71 m/s] and at either the weakest or stiffest part of
the steering wheel, then the experimental information obtained in these types of tests
may prove to be worthless if a n impact occurs at Iower (or higher) speed and at any other
position on the steering wheel other than the weakest or stiffest locations.
If experimental testing is conducted using the SAE 5944 standard or its European
impact between the steering wheel and chestform are required. However, in a real world
collision situation there rnay exist an infinite number of possibilities as to how the dnver
will hit the steering wheel based upon the location of impact, the orientation of the
steenng assembly, and the velocity at which the driver collides with the steenng wheel.
The need to hlly understand how a steering wheel will perforrn at any impact
position, velocity, or steering wheel orientation very much exists. This wilI enabte
engineers to properly estimate the abilities of a steering wheel in a real world collision as
wel 1 as modiQ their current steering wheel designs so as to improve on the performance
The work associated with this dissertation focuses on two specific concems
tests on that wheel. The experimental testing will consider variations in the
orientation and location of impact on the steering wheel, the impact velocity
of the chestform, and the column angle of the steering wheel. The variation
of these three independent variables will be used to examine how they affect
the resulting peak impact load, peak impact displacement, and calculated
of impact on the steering wheel, the impact velocity of the chestform, and the
There are a number of terms associated with steering wheel testing which need to
reference line to the centre line of the steering column. Figure 3 illustrates a steering
wheel and a bodflorm which are used to define the column angle. In measuring the
column angle, positive values are taken in a cIockwise sense from the horizontal
reference line. Furthemore, a zero degree column angle represents the centre line of the
Bodyform
i instant contact is made (directed
towards steering wheel,parallel
to the horizon)
only
translates
towards (or
away) from
steering
wheel
1 Column Angle - angle between
horizontal aGd steer& column
Figure 3 - The column angle and impact velocity.
Typically, in modem automobiles the column angle ranges from 25 degrees to 35
degrees. However, with the addition of tilt steering, the actual column angle, referenced
to the steering wheel, may not be the actual column angle specified by the automobile
manufacturer. Depending upon the preference of the driver, the column angle referenced
from the steering wheel may be significantly different fiom the actual column angle.
For this study, a "low column angle" indicates that the column angle is smali' in
the range of 15 degrees to 20 degrees. As weli, a "steep or high column angle" indicates
impacting chestform at the instant contact is made between the chestform and the
steering wheel. The impact velocity represents the maximum value of the velocity of the
chestform over the duration of impact. It should be noted that the bodyform only
translates in a direction parallel to the horizontal reference line. Thus, the impact
This angle represents the orientation or angular displacement which the steering
wheel is being tumed during an impact test. Figure 4 illustrates the wheel angle and
different wheel angle locations over the entire steering wheel. The wheel angle is
degrees. The wheel angle varies kom the 12 o'clock position ( O 0 position), which is the
top dead centre (TDC) of the steenng wheel (in normal driving conditions, ie. driving
"Normal" 12 o'clock or O
driving position. ..,/ -
4
; degree wheel angle
y'- \.
3 o'clock or
1L - 3 If---. an APUWP
straight ahead), to the 6 o'clock position (1 80" position), which is the bottom dead centre
(BDC) of the steering wheel (in normal driving conditions), and back to the 12 o'clock
position.
It should be noted that any collision between the chestforrn and steenng wheel
\vil1 always occur at the B K of the steering wheel. However, depending upon how the
steering wheel is turned, this may not represent the 6 o'clock ( 180") position. If the
wheel is tumed, then in general, impact will not occur at the 6 o'clock position. In
addition, al1 impacts on the chestforrn occur along the centreline (midline or a i s of
4.1.3. Peak Impact Load, Peak Impact Displacement, and Calculated Energy
The peak impact load and peak impact displacement represent the maximum
values of the force and LVDT measurements in the direction of the impact velocity or
number of data points attained in a test. Equation (2) basically determines the area
within the curve bounded by the force versus displacement profile observed in an
experimental test. The force and disptacement measurements are al1 taken in a direction
along the horizontal reference line (which is in the vertical direction on the testing
machine).
The steering wheel selected for experimental testing is a three spoke wheel which
is manufactured by K.S. Centoco Ltd., in Tilbury, Ontario, Canada. The three spoke
A die casting process is used to produce the steering wheel. Furthemore, the
for the steering wheel industry. To minimize differences in steering wheels due to any
variation in the manufacturing processes, al1 steering wheels tested were talcen from a
batch of wheels which were produced during the same manufacturing run with the same
die.
As previously indicated, one of the goals of this study involved the investigation
of how the location of impact, column angle of the steering wheel, orientation of the
steering wheel, and impact velocity of a chestfom ont0 the steering wheel effect the
peak impact load, peak impact disptacement and calcuIated energy during a collision
between the chestfonn and steering wheel. The orientation, column angle, and location
of impact on the steering wheel can be defined by the value of the wheel angle and
column angle. As well, the impact velocity defines the speed of impact.
The manufacturer o f the automobile, which utilizes the three spoke steenng
wheel, has provided information regarding al1 the possible values of the col umn angles
under normal driving conditions [32]. The actual steering colurnn angle is twenty eight
degrees (28"), however, the tilt steering provides ten ( 10) increments of two and one
quarter degrees (2.25" ) gaps, or spaces. Thus, the values of the cotumn angles with tilt
steering are 38. i X 0 , 35.875 O, 33.62s0, 3 1.375", 29. l25", 26.875", 24.625". 22.375 O,
20.1 X O and
, 17.875".
During an actual car crash situation the steering column generally has a tendency
to push upwards toward the driver and in effect increases the column angle [33]. Thus,
for experimental testing, the range of values of column angles that were chosen vaned
angle values which were considered in this study; 1 S O , 18", 2 1 O, 24", 27"- 30". 33"- 36",
Due to the symmetry of the wheel, when considering the variation in the wheel
angle, only half of the possible wheel angles were needed to be investigated. For
the 9 o'ctock position. Thus the range of wheel angles considered in this investigation
were from the 12 07cIockposition (O " wheel angle) to the 6 o'clock position ( 180" wheel
angle), in increments of 1 hour or 30". Hence, seven (7) different wheel angles were
considered: 12 o'clock, 1 o'clock, 2 o'clock, 3 o'clock, 4 o'clock, 5 o'clock, and 6 o'clock.
It has been shown that impacts at speeds greater than 13 rnph [5.81 m/s] tend to
develop a very large degree of deformation in the steering wheel, which does not actually
occur in real world dnver/steerîng wheel collisions [34]. This is based upon testing
conditions similar to which al1 the experimental tests for this study were conducted with
f5.81 mis] was selected as the high end for the range of impact vetocities. Based upon
past testing observations, a low end value for the impact velocity was selected as 8 rnph
[3S 8 m/s]. Impacts with speeds lower than 8 rnph i3.58m/s] and at a stiff location on
the steering wheel generatly illustrated deformations similar to low speed car crash
situations. The increment on the impact velocity was selected to be 1 mph [ M l m/s].
Thus, six (6)different impact velocities were considered in this investigation; 8 rnph
[3.58m/s], 9 rnph C4.02 rn/s], 10 rnph E4.47 mls], 11 rnph [4.92 d s ] , 12 rnph C5.36 m/s],
Based upon the nwnber of increments for the wheel angle (7),column angle ( IO),
and the impact velocity (6), the total nurnber of tests conducted in order to investigate al1
possible combinations were 420 experimental runs. This value represents the products of
the increments for each independent variable (wheel angle, column angle, and impact
The testing machine used in this study is the testing machine developed by K.S.
discussion of the components and devices incorporated into the testing machine is
bodyfom and LVDT are rigidly attached to a translating crosshead constrained to move
referenced or "zeroed" position. In al1 impact tests, completed in this study, the
reference position was set at a location such that a gap of 0.5 inches [12.7 mm] existed
Three spoke
steering wheel
fastened to the colurnn angle joint in a similar manner as it is fastened to the steering
column of an actual automobile. The locking column angle joint is allowed to rotate
with the aid of a hydraulic release mechanism. If hydraulic power is supplied to the
column to any required angle. The hydraulic power can then be disengaged which
deactivates the release mechanism and the column angle is rigidly locked into position.
Located near the column angle joint is a similar arrangement of devices that allow
rotation of the wheel angle. Similar to the column angle joint the wheel angle joint also
utilizes a hydraulic clarnping device to lock the wheel angle in a specified position.
Located directly below the column angle joint is the triaxial load cell. This
device measures forces in three mutually perpendicular directions. The most important
direction is the vertical sense or direction of impact and is identified as the --z-axis". The
other two directions of measurernent are termed the "x-ais" and "y-ais", which
measure the force in a direction along the width and depth of the machine respectively.
The " z - a d ' also represents the direction that LVDT displacement measurements are
made.
70 inches [1.78 ml. This distance, depending upon tubrication conditions between the
guideposts and bushings on the crosshead, can provide a maximum impact velocity of
roughly 13 mph [5.8 1 m/s]. In addition, a pneumatic cylinder can be utilized to provide
an "initial push" which can cause the crosshead to develop higher impact velocities, if
necessary. A calibration of dropping height versus impact velocity has been developed
[ X I . For the impact velocities which will be considered in this study the following
crosshead dropping heights, given in Table 1, are required. Table 1 also provides the
theoretical impact velocity calculated using equation (3), based upon constant
acceleration. without any fnction losses between the guideposts and bushings.
*
In equation (3,
,,,y is the theoretical impact velocity, h is the theoretical dropping
height, and 'g' is the local acceleration due to gravity. Note that the calibrated dropping
heights are, in general, higher than the theoretical dropping heights. This should be
Furthemore, at higher impact velocities, the difference between the calibrated dropping
-
Table 1 Crosshead D ~ O D D Heights
~ E Reauired
for the Impact Velocities Considered in this Study
interval requires the incorporation of a high speed data acquisition card in the testing
machine.
For al1 the experiments considered in this investigation, the testing machine
recorded five hundred data points of information during the impact period. The
observations recorded include the three directions of loading €rom the triaxial load cell,
measurements from the three accelerometers mounted to the back of the chestform at the
head, lefi shoulder, and lower abdomen region, and the displacements from the LVDT.
In addition, the high speed data acquisition board also kept a record of the time at which
Ail the information observed in a test was stored in an ASCII text file which was
detennine the peak loads, accelerations, impact velocity, and displacement of the
crosshead and steering wheel. As well, time history information for the loads,
Mathcad was selected as the s o h a r e for the analysis of the experimental data,
pnmarily due to its excellent abilities with unit conversions and mathematical
operations. Unfortunately, the data files generated by the testing machine require
reformatting in order for Mathcad to be able to properly read the obsewed information
from an ASCII data file. Thus, a Visual Basic program was developed by the author to
The Mathcad worksheet analyzes the data file from the droptower testing
machine and determines the impact velocity of the chestform, the maximum load in al1
three directions fiom the triaxial load cell, the peak crosshead displacement, the peak
acceleration measured by each accelerometer, and the calculated energy absorbed during
-.-
TEST H: Impact at 6-85 mph 1 3.06 m/s
- TEST #2: Impad at 8.82 mph 1 3.94 m/s
TEST #3: Impact at 10.77 mph / 4.81 mis
--
Figure 6 - Typical crosshead displacement versus time curves fiom the Mathcad
worksheet (note: each curve is offset in the time axis to better illustrate each
curve).
the impact. From equation (2), the calculated energy is detennined by integrating the
"z-axis" force versus LVDT displacement curve. Furthemore, the Mathcad worksheet
also provides a graphical output of al1 the loads, accelerarions, crosshead velocity, and
typical LVDT displacement versus time and "z-axis" load versus time curves frorn the
Figure 7 - Typical "z-axis" load versus tirne curves from the Mathcad worksheet. A
summary of the peak vertical loads is also provide in the upper Rght region
of the graph.
To determine the calculated energy a b s o h d during an impact test requires the
inteçration of the "z-axis" load versus crosshead displacement curve. If one cross-plots
the displacement versus time curve (as illustrated in Figure 6), with the ' ' z - a ~ i sload
~~
I I I 1 1 1 1 1
-4 -2 O 2 4 6
Crosshead Displacement (inches)
versus time curve (as illustrated in Figure 7) a typical "z-axis" ioad versus crosshead
should be noted that one of the curves, in Figure 8, is offset in the x-axis to illustrate
absorbed in the impact test is presented in the top lefi corner of the vertical load versus
this investigation to fully consider the variations in the column angle, wheel angle, and
impact velocity. The procedure for each individual test was identical. The only variation
from test to test was the settings (column angle, wheel angle and impact velocity)
The chestfonn is made from a low density polyurethane foam. To ensure that the
chestform was not degrading during the course of the 420 experimental tests, a rigd
cylinder was manufactured and used as a control to veriw if, over the span of the tests,
variations in the vertical load versus crosshead displacement curve developed. The
Adhesive Tape
(to hold chestjorm to wooden
bockplate)
above the cylinder and at a location across the chest of the bodyform. Figure 9 illustrates
the test setup. A leather covering was developed, for the chestfom, to minimize any
tearing that may occur on the surface of the chestfom. The total mass of the dropping
assembly , tncluding the chestform, wooden backing plate, and aluminum crosshead was
These "integrity checking" tests were completed after every fiftieth experimental
test. Over the entire 420 tests compteted, no significant variation in the load versus
The experimerital testing apparatus is illustrated in Figure 10. For each test a new
steering wheel was bolted to the steering column shaft. if necessary, the hydraulic power
supply was energized to unlock the column angle and wheel angle joints. The position of
the wheel angle and column angle was then modified to the required testing conditions.
The hydrauiic power supply was then de-energized to clamp down the column angle and
wheel angle joints. A one half inch [12.7 mm] gauge block was then placed on the BDC
of the steering wheel (where impact will first occur), and the chestfonn was lowered so
that it was just touching the gauge block. The gauge block was then removed fiom
Deformable cheslform
with leather wrap
Steering Wheel
fiflustratedas i m ~ a cat
t 6 O 'dock) other translations or rotations).
Figure 10 - Experirnental test setup for a typical "chestform to steering wheel" impact
test. The experimental testing conditions for this exact test involve a
column angle of 27", a wheel angle of 180" (6 o'clock), and an impact
velocity of 1 1 mph i4.92 d s ] .
l Wooden Backing Plate
A dhesive Tape
'Deformable c h e s t f m
with leather wrap
Figure 11 - Defonned geometry of the steering wheel and chestform after an impact
test. The testing conditions for this exact test are a coiumn angle of 2ï0,a
wheel angle of 180" (6 o'clock), and an impact velocity of 1 1 mph
C4.92 rn/s].
between the chestfom and steering wheel and the triaxial load cell, LVDT, and
accelerometers were zeroed (placed into the reference position). The chestform was then
raised to the appropriate dropping height and released. Figure 1 I illustrates a completed
test. Data acquisition was completed and the chestfonn t a s elevated to a safe distance
above the deformed steering wheel. The defonned steering wheel was replaced with a
Upon completion of the 420 expenmental tests, each data file was modified,
using the Visual Basic program, and analyzed in the Mathcad worksheet. One of the
focuses of this investigation was to examine the effect of colurnn angle, wheel angle, and
impact velocity on peak vertical (or "2-axis") Ioad, peak vertical displacement and
calculated energy absorbed in the impact. A discussion of the energy absorbed during
calculated energy profile is obtained. Figure 12 illustrates the graph of calculated energy
O 2 4 6
Crosshead Displacement (inches)
\vas conducted with a column angle of 2 î 0 , a wheel angle of 180" (or at the 6 o'clock
position), and an impact velocity of 12 mph [5.36m/s]. There are several key positions
on the curve that should be discussed. Figure 13 illustrates the identical curve presented
VJ a
O 2 4 6
Crosshead Displacement (inches)
First, the green vertical reference line represents the peak crosshead (or LVDT)
displacement. This value is the experimental measurement obtained minus the height of
the gauge block used in the "zeroing" procedure. Hence, the values along the horizontal
axis of the curve represent actua1 steering wheel and chestform displacements from their
original, undeformed positions. This value indicates the maximum displacement of the
chestform has little or a very small amount of kinetic energy, and is beginning to enter
Secondly, the peak value on the vertical portion of the curve, represents the
maximum amount of energy absorbed during the impact (this is illustrated with the red
horizontal reference line). Finally, the blue horizontal reference line illustrates the total
There are two distinct regions of the curve that need to be discussed. The
difference between the maximum energy absorbed and the total energy absorbed
represents the elastic response of the system (the chestform and steenng wheel). This
arnount of energy is &en back by the system in an impact test, and generaily causes the
chestform to move in the opposite direction of the impact velocity (Le., causes the
chestform to rebound off the steering wheel). The value of the total energy absorbed
represents the amount of energy transfened to the steering wheel and the chestform, to
plastically deform the two entities. The value of the total energy absorbed is the most
sipificant information as it indicates the amount of energy that both the steenng wheel
and chestform absorbed in an impact. This value is very important in the design and
In general, it has been observed that the maximum amount of energy absorbed
coincides with the peak crosshead displacement. This should be expected as in this
position, the chestform is beginning to rebound from the steering wheel. This can only
occur if energy is k i n g transferred back to the chestform. The transfer of energy back to
The calculated energy profile given in Figures 12 and 13 do not actually represent
the total arnount of energy absorbed in the impact process. These curves fail to consider
any energy absorbed in the other two perpendicular directions (Le. in the direction of the
direction). Wi th the experimental testing machine used, measuring the energy in the
other two directions is physically impossible. This is primarily due to the fact that
displacement curve does give an excellent approximation to the actual amount of energy
absorbed in the impacting process. This can be stated since the primary amount of
energy before impact, which is the combination of kinetic and potential energy, are both
5.2. Utilization of the Calculated Energy Profile for Development of More Energy
Absorbing Steering Wheels
As can be seen in Figure 13, and from the previous discussion, the task of
developing more energy absorbing steering wheels requires a maximization of the total
energy absorbed in an impact. This requires an investigation into the physicai geometry
steering wheel which can be more energy absorbing. Unfortunately, there are probably
constraints) but the principal idea is that modification of the geornetry of a steering
wheel can make it more or less energy absorbing. Although not considered in this
investigation, optimized section thickness, of the spokes, rirn, and hub of the steering
elastic response should be utilized in a steering wheel. The three spoke steering wheel
aluminum is a very ductile material, which indicates it requires a large amount of strain
(or deformation) to cause actuat failure of the material. Having a very small elastic
before failure, will aid in increasing the total energy absorption in an impact.
there exists only potential energy in the structure, which is equal to the product of the
- Position of chestform
before dropping (i.e.
dropping height +
displacement due to impact)
f e l o c i t y = O (no motion)
mass of the system, "m", the height from a referenced position, "Hzy, and the local
acceleration due to gravity, "g". Equation (4) is used to calculate the potential energy at
'p,ent,a/ = m . g o4 (4)
At position "3", just before impact, the dropping assembly wilf possess kinetic and
potential energy. The total energy at impact (or the impact energy) can be calculated
impact. The potential and kinetic energies at impact can also be hrther simplified to the
where, "H," is the height from the reference position (position at maximum deflection of
the steering wheel and chestform) and "v,"is the impact velocity of the dropping
assembl y.
I f one considers the ratio of the calculated energy absorbed to the amount of
energy available in the system before an experimental test, the following equation can be
generated:
vertical force versus vertical displacement curve, and 'E,,m," is the energy before
impact, either at the dropping height, "Hz" (E,,,), or just before impact (EimPCI).
procedures. It can be calculated using the potential energy, just before the dropping of
the crosshead assembly, or using the impact energy, just pnor to impact of the chestform
on the steering wheel. Theoretically, both methods should result in an identical value for
the E.A.F.. However, depending upon the amount of fnctional iosses that occur between
the bushings of the crosshead and the guideposts, along which the bushings travel, the
E.A.F. based upon potential energy (E.A.F.P.) may be higher than the value determined
from the E.A.F. based upon energy at impact (E.A.F.I.). In addition, if an inadequate
numerical algorithm for differentiating the displacement versus time curve (obtained
from the LVDT) is used, then the impact velocity (which is calculated using the LVDT
measurements) will have excessive numerical error and thus the E.A.F.I. will also have a
The value of the E.A.F. should always be less than, or equal to, unity. A value of
unity indicates that the calculated energy absorbed by the steering wheel and chestform is
equal to the energy availabfe before impact. This represents the ideal case for steering
wheel design. A "perfect" steering wheel is one in which it is capable of absorbing al1
the mechanical energy that the chestform possesses just before the collision. Steering
wheel designers can use the energy absorption factor (E.A.F.) as a numerical means of
Furthemore, an E.A.F. value of unity, indicates the situation of perfectly plastic impact,
where the coefficient of restitution between the two impacting bodies is equal to zero.
Ail the energy, which the dropping assembly possessed before impact, has been
transferred to the steering wheel and resulted in the plastic, or permanent, deformation of
the wheel. In the case where the E.A.F. equats unity, no rebounding of the chestfonn
occurs.
response (or an elastic portion of a stresdstrain curve), then the E.A.F. will always be
This is identical to the perf'ectly elastic impact situation, where the coefficient of
restitution between the two deformable bodies in impact is equal to unity. In this case,
where the E.A.F. is equal to zero, the rebound velocity of the dropping assembly is equal
to the impact velocity which it possessed before the collision. For steering wheel design,
this represents the worst design case. Al1 the energy that the dropping assernbly
possessed before impact is transferred back to the dropping assembly afier impact.
injuries. In a real world situation, the E.A.F. will generally be greater than zero but less
than one.
For experimental impact test #345, which had a column angle of X O a, wheel
angle of 180" (or impact at the 6 o'clock position), and an impact velocity of 12 mph
[5.36 m/s], the energy absorption factor based on potential energy (E.A.F.P.) was found
to be 0.69. The energy absorption factor based on impact energy (E.A.F.1) was
calculated to be 0.70. The difference between the two indicates losses due to friction or
errors in the calculation of the impact velocity. The percentage difference between the
two values is, however, very small; calculated to be 1.45 %. These calculated energy
absorption factors indicate that the steering wheel and chestform are absorbing
Unfortunately, for steering wheel designers, who are most concemed with the
steering wheel, this value is somewhat incorrect. The calculated E.A.F. gives an
indication of the total absorbed energy by the steering wheel and the chestform, not just
the steenng wheel alone. By using the deformable chestform in this experimental study
separating the amount of energy absorbed by the steering wheel and the amount of
enerw absorbed by the chestform is very difficult. A more robust method for
determining the E.A.F. would require the use of a rigid or non-deformable chestform in
the experimental test. Since the chestform would not deform, no amount of energy
available before impact could be transferred to the intemal energy within the chestform.
Hence, only the steering wheel would be able to absorb energy in the experimental
impact test.
chestform and the three spoke steering wheel can be determined. The coefficient of
restitution represents a ratio of the impulsive forces during the period of restitution to the
The period of deformation represents the duration of time in which the two
bodies (the steering wheel and chestform) deform and at the end of the period of
deformation both entities will have the sarne velocity. The period of restitution follows
the period of deformation and in this time the entities will attempt to regain their initial
shape.
Using the impulse-momentum equation, equation (8) is modified to consider the
velocity of two objects (the chestform and steering wheel) in impact, as presented in
equation (9):
chestfom aAer impact respectively. As well, ''vaenW . ,," and "vchafim ," are the
respective velocities of the steering wheel and chestform before impact In addition,
and "v am., ,, ,"are zero. Thus, equation (9) can be simplified to
equation (IO), which ultimatety can be used to determine the coefiicient of restitution
experïmental test and the rebound velocity (v,,,,, ,) is equal to the minimum veIocity
attained in the collision. Figure 15 illustrates the procedure to determine the impact
velocity and rebound velocity from an experimental test curve of velocity versus time.
completed using the velocity versus time curve and equation ( 10). Based upon
experimental data from test #345 the coefficient of restitution was calculated to be 0.40,
indicating a more plastic impact situation is occumng for this experimental test.
Crosshead Velocitv Versus Time
1;
I
Time (seconds)
Figure 15 - Typical crosshead velocity versus time curve iltustrating the impact velocity
and the rebound velocity.
Another important observation that should be discussed is the peak impact force.
During an experimental test, the steering wheel is tilted at a specific column angle, which
will sigiificantly affect the "y-axis" direction of load (which is the plane in which the
column angle lies, the y = O plane). Furthemore, the wheel angle of the steering wheel
may be set so that a symmetric impact will not occur. This experimental testing
parameter will affect the "x-axis" loading profile during an impact test.
To consider the effect of the x and y axes of loading, during an impact test, the
concept of the peak impact load has been developed. Mathematically, it is the square
root of the surn of the squares of the loads in the x, y, and z axes direction, as presented
in equation ( l 1 ).
The peak impact force is another method of quantiGing the effect of wheel angle,
column angle, and impact velocity on the performance of a steering wheel during an
impact situation.
For the above experimental test (test #345) the peak impact load was calculated
to be 2353.75 lbf [10.47 W], just slightly higher than the peak vertical impact force of
effectiveiy, a topographie, or surface plot, procedure has been selected to illustrate the
effects of coIumn angle, wheel angle, and impact velocity on peak vertical load, peak
Atthouçh three (3) independent variables are considered in this study, if one of
t hem is maintained constant, the development of surfaces il Iustrating the effects of the
completed. In the presentation of the data the impact velocity will be maintained
constant and variations in the wheel angle and the column angle are considered.
5.6.1. The Effeet of Column Angle and Wbeel Angle on Peak Vertical Load
Figures 16, 17, 18, 19,20, and 2 1 illustrate the effects of column angle and wheel
angle on the peak vertical load for impact velocities of 8 rnph [3.58 m/s], 9 rnph
[4.02 m/s], 10 rnph 14.47 mis], 1 1 rnph [4.92 m/s], 12 rnph [5.36m/s], and 13 rnph
[5.8 1 mis], respectively. It should be noted that the colour variation (lower values on the
peak vertical load axis indicated by red, and higher values on the peak vertical load mis
indicated by blue) on the surface plots is used only to illustrate the shape and contours of
- - - - -
--
2 1, the peak vertical load of impact varies over the location and orientation of the
steering wheel. In other words, the peak vertical load during impact does not remain
constant over al1 the different wheel angles and column angles considered in this study.
If the peak vertical load was to remain constant, then for each surface plot (at a specified
impact velocity) a constant, flat surface would be observed, which is obviously not the
case.
The peak vertical load alço varies depending upon the velocity of impact. Since
changes in the topography of the surface plots are observed frorn a low impact velocity
(8 rnph l3.58 mh]) to a higher impact velocity (13 rnph [5.81 mk]) then the peak vertical
load also depends on the impact velocity, which should be expected. As well, a greater
"valley" begins to occur, in the surface profile, for a wheel angle of approximately 40" to
5.6.3. The Effect o f Column Angle and Wheel Angle on Peak Vertical
Displacement
Figures 22,23,24,25,26, and 27 illustrate the effects of column angle and wheel
angle on the peak vertical displacement for impact velocities of 8 rnph [ 3 S 8 mis], 9 rnph
[4.02 m/s], I O rnph [4.47 d s ] , 1 1 rnph [4.92 d s ] , 12 rnph [5.36m/s], and 13 rnph
[5.8 1 d s ] , respectiveiy. Again, the colour variation (lower values on the displacement
asis indicated by red, and higher values on the displacement axis indicated by blue) on
the surface plots is used only to illustrate the shape and contours of the topographie
profile better.
M e e l and Column Anales Venus Peak Vertical Displacement
at 8 m ~ (3.58
h mis) lmpact Velocitv
i
Figure 26 - Wheel and column angles versus peak vertical
displacement at 12 mph [5.36 m/s] impact velocity.
- - -- - -
vertical displacement with ch'anges in impact velocity. All the surfaces (in Figures 32
through 27) take on a very similar appearance and possess very similar magnitudes of
peak vertical displacement, with changes in impact velocity. Thus, the impact velocity
has only a rninor affect in determining the peak vertical displacement. It should be noted
that, regardiess of impact velocity and column angle, at low wheel angles the peak
vertical displacement is much more predorninant than at high wheel angles. In addition,
at low wheel angles and high column angles, the peak vertical displacement is generally
the greatest for a given impact velocity. This should be expected since the tilt of the
column angle provides more vertical clearance between the outer rim of the steering
wheel and the inner hub of the steering wheel. At higher wheel angles, from the
considerably smaller than at al1 other positions, regardless of impact velocity or column
angle. This should also be expected as impact is occumng on a position of the steering
5.6.5. The Effect of Column Angle and Wheel Angle on Calculated Total
Energy Absorbed
Figures 28, 29, 3 0 , 3 1,32, and 33 illustrate the effects of column angle and wheel
angle on the calculated total energy absorbed for impact velocities of 8 rnph [3.58 m/s],
9 rnph [4.02 m/s], 10 rnph 14.47 d s ] , I I rnph [4.92 d s ] , 12 rnph [5.36 m/s], and 13 rnph
-- - --- -
Il
Figure 31 - Wheel and colwnn angles versus calculated total energy
absorbed at 1 1 mph [4.92 m/s] impact velocity.
M e e l and Column Anales Versus Calculated Energy
at 12 rnnh (5.36 m/s) Impact Velocity
Similar to the peak vertical load at impact, the calculated total energy absorbed
varies over the location of impact on the steering wheel. Furthemore, the calculated
total energy absorbed also varies significantly with impact velocity. This should be
expected as the energy at impact is proportional to the square of the impact velocity.
Slight increases in the impact velocity will cause significant changes in how much energy
the dropping assembly possesses before impact and will, in tum, directly affect the
amount of energy the steenng wheel and chestform absorb in an impact test.
The magnitude of the calculated total energy absorbed is related to the amount of
plastic deformation that occurs within the steering wheel and the energy absorbed by the
wheel, requires the investigation into the optimization of the steenng wheel geometry,
such as section thickness of the spokes, rim and hub, as well as the use of a very energy
absorbing material. The energy absorbing material would require a relatively low yield
stress, with a high Young's Modulus, and a significantly large plastic regime in the
5.6.7. The Effect of Column Angle and Wheel Angle on the Average E.A.F.
[n this section the effect of wheel angle and column angle on the average E.A.F.
is considered for impact velocities of 8 mph r3.58 m/s], 9 mph r4.02 m/s], 10 mph
14.47m/s], 1 1 mph [4.92 d s ] , 12 mph [5.36 m/s], and 13 rnph [5.81 mk]. The average
E.A.F. represents an average value calculated from the E.A.F.P. (energy absorption factor
based upon potential energy) and the E.A.F.I. (the energy absorption factor based upon
energy at impact).
M e e l and Column Anales Versus Averaae E.A.F. I
W at 8 mph 13.58 mls) Impact Velocitv
I
Figure 34 - Wheel and colurnn angles versus the average energy absorption
factor (E.A.F.) at 8 mph [3.58 mis] impact velocity.
l
m e e l and Column Anales Versus Average E.A.F.
at 9 moh 14.02 m/s) lmnact Velocity
-
- - - - -
Figure 35 Wheel and column andes versus the average energy absorption
factor (E.A.F.) at 9 mph [4.02 ds]
impact velocity.
M e e l and Column Anples Versus Averaae E.A.F.
at 1O mph l4.47 mis) Impact Velocity
Figure 36 - Wheel and column angles versus the average energy absorption
factor (E.A.F.) at 10 mph E4.47 d s ] impact velocity.
Figure 37 - Wheel and column angles versus the average energy absorption
factor (E.A.F.) at I I mph [4.92 d s ] impact velocity.
M e e l and Column Anales Versus Averaae E A F .
at 12 mph (5.36 m/s) Impact Velocitv
Figure 38 - Wheel and column angles versus the average energy absorption
factor (E.A.F.) at 12 mph i5.36 d s ] impact velocity.
Figure 39 - WheeI and column angles versus the average energy absorption
factor (E.A.F.) at 13 mph [5.81 m/s] impact velocity.
5.6.8. Comments on the Variation of the Average E.A.F.
The average E.A.F. can provide steerïng wheel designers and engineers with a
tool that they can use to develop extremely energy absorbing steering wheels. The E.A.F.
value gives an insight into how much energy the steering wheel and chestfonn absorbed,
based upon the total available amount of energy before the impact test.
A perfect steering wheel is one in which the value of the E.A.F. is unity. This
indicates that the steering wheel absorbed al1 the available energy (before impact) and
little or no energy will be absorbed by the driver of the automobile. Furthemore, for a
perfect steering wheel, this value should be independent of the steering wheel
wheel is hit by a driver or chestfom, it will absorb al1 the energy that the impacting body
possesses. For the driver of the automobile, this is the perfect impact situation.
for the range of velocities considered in this investigation accounts for the three spoke
steering wheel and the chestform. Ideally, one requires the E.A.F. for just the steering
wheel; however, isolation of the energy absorbed by just the steering wheel cannot be
quickly and effectively determined €rom the observations attained in the experiment. If
one assumes that the chestform absorbs a similar amount of energy in each impact,
regardless of impact location, impact velocity, or wheel orientation, then one can
visualize (from the surface plots in Figures 34 through 39) which combination of wheel
angle, column angle, and impact velocity represent the best or worst design cases for the
three spoke steering wheel. If designers and engineers are going to improve on the
design of this three spoke steering wheel then they should focus their attention on the
Experirnentally it was found that, for the range of impact velocities considered,
the average E.A.F. ranged fiom 0.85 to 0.61. Indicating that at the best situation, which
occurred with test #23 1 with a column angle of 15", a wheel angle of 60" (impact at the
2 o'clock position), and an impact velocity of 1 1 mph [4.92 rn/s], the steering wheel and
chestfom absorbed approximately 85% of the energy available before impact. In the
worst situation, which occurred with experimental test $49 with a column angle of 39", a
wheel angle of 120" (impact at the 4 o'clock position), and an impact velocity of 8 mph
13.58 mis], the steering wheel and chestform absorbed approximately 6 1% of the
From the observations obtained in this study, one finds that the average E.A.F.
varies depending upon the location of impact and steering wheel orientation, for a given
impact velocity. This is evident since each surface plot illustrates diffenng E.A.F. values
for changes in the wheel angle and the column angle. In addition, the E.A.F., for this
three spoke steering wheel, is also a function of the impact velocity. By cornparing the
surface plots over the range of impact velocities considered in this study, one notices that
Figures 40,4 1,42,43,44, and 45 illustrate the effects of column angle and wheel
angle on the coefficient of restitution between the chestforrn and three spoke steering
wheel for impact veloçities of 8 rnph [3.58 m/s], 9 rnph [4.02 mls], 10 rnph [4.47 m/s],
1 1 rnph [4.92 m/s], 12 rnph l5.36 rn/s], and 13 rnph [5.81 rn/s], respectively. The
Similar to the peak vertical load of impact, it appears, based upon the surface
plots in Fi y r e s 40 through 45, that the coefficient of restitution varies depending upon
the wheel and column angles. At lower impact velocities however, surface plots are
significantly flatter than at higher impact velocities. This indicates that at lower impact
velocities, the coefficient of restitution is less dependent upon the column angle and
appears to be more dependent upon the wheel angle than on the column angle. For
example, at an impact velocity of 12 mph [5.36 m/s], and if viewing the surface plot
alonç a line in the direction of the column angle axis, one sees a parabolic relationship
between the coefficient of restitution and the wheel angle. However, if viewing the same
plot, in a direction along the line of the wheel angle axis, one notices that the coefficient
of restitution appears to be constant (for a given wheel angle) for any value of the column
angle. This was also observed to occur for the experimental tests conducted at 1 1 rnph
to V a r y depending upon the impact velocity. Each individual surface plot changes shape
dependent upon the speed of impact between the two bodies in impact [36].
The value of the coefficient of restitution is always between zero and one. Zero
indicates a perfectly plastic impact situation, where the two bodies in impact physically
stay together after impact. Also, in this case, where the coefficient of restitution is zero,
the period of restitution is zero, implying that no impulsive forces exist afker the
then, in the case of the chestform impacting the three spoke steenng wheel, the impact
velocity will equal the rebound velocity. The case where the coefficient of restitution
equals unity represents the worst possible situation for a driver hitting the steering wheel.
The rebound velocity will equal the impact velocity and thus more energy will be
available to do damage or bodily h m to the driver. In addition, the driver's body will
also experience very significant magnitudes of acceieration, which also may play a very
important role in determining the amount of bodily harm which the driver may have to
endure. If the coemcient of restitution is equal to zero then the driver will have no
rebound velocity and thus a low amount of energy available to cause bodily harrn. As
well, the magnitude of the acceleration which the driver wiil have to endure will be
significantly less then the case where the coefficient of restitution is equal to one.
The relationship between the coefficient of restitution and the average E.A.F., is
illustrated in Figures 46 through 5 1. The blue colour filled circles represent actual
experimental data, and the smooth red curve (for each plot) illustrates a second order
polynomial curve fit to the experimental data. Each plot is for a specific impact velocity.
Coefficient of Restitution Versus Enerav Absomtion Factor
For 8 mph (3.58 m/s) Impact Velocity
0.6 - Experimental Data
- 2nd Order Polynomial Cuwe Fit
h
a
w e(E.A.F.) = 2.251 - 4.361 *(EA.F.) + 2 . 4 7 * ( ~ ~ . ~ . ) ~
t
O 0.5 - 'Y a
*-
u
7 .\.+a
.-
w
u
-a
V)
0.4 -
l .@
+
.o
.-:. a
' '
u
f
O
C
aa
-0
0.3 - O
*
*
ha 0
O
U
'
O.* -
0.60 O .ô5 0.70 0.75 0.80 0.85
Energy Absorption Factor (E.A.F .)
h
0,
0.6 - r
.
Experimental Data
- 2nd Order Polynomial Cuwe Fit
I
I
. *-
V
0.5 -
.-
c
O
1
. "y,.a *
.-
u
C
'
\
\a.
rn
-
8
O
0.4 -
O
a
a
.'
c.
E * O
al - a 0.
O
O
- . ' a
a.:
.
0.2
0.62 0.64 0.66 0-68 0.70 0.72 0.74 0.76 0.78 0.80 0.82
Energy Absorption Factor (E.A.F.)
--
3
C
u
cn
2 0.4 -
+
O
C * m
C
Q)
0.3 - O mm
m a *
O
O
0.2 -
0.64 0.66 0.68 0.70 0.72 0.74 0.76 0.78 0.80 0.82 0.84
Energy Absorption Factor (E.A.F.)
8 0.38 -
u
m
rt
O
2 0.32 -
.-al
O
5 0.26
O
-
0.20-
0.65 0.67 0.69 0.71 0.73 0.75 0.77 0.79 0.81 0.83 0.85
Energy Absorption Factor (E.A.F.)
- - -
.
For 12 mph (5.36 m/s) Impact VelociN
Experimental Data
0.45 -
..
- 2nd Order Polynomial Cuwe Fit
h
aa 0 e(E.A.F.) = -1.603+5.769'(E.A.Ç.)-4.237YE.A.F)'
Y
3 g
-0 0.40 -
C 0
C
3 a%; a .
.-
c
u
.
Y,
$ 0.35 -
0-45 - .
For 13 mph (5.81 mis) Impact Velocitv
Experimental Data
- 2nd Order Polynomial Cuwe Fit
~
1
O
0)
*-
C
3
.-
C
0.40 -
.. .
0 -
. b
-8y
-...
( E . A . F . ) = -0.679 + 3 196'(E.A.F.) - 2.467'(E A.F )'
O. . O
0 0.35 -
u
-
E
O
E 0.30 -
O
aa
-3
k 0.25 - I a
. .
O a
. a
experimental data appears to agree significantly better with the second order least
squares curve fit approximation. At higher impact velocities, the experimental data is
more scattered, indicating that there is little or no relationship between the average
It should be noted that, regardless of the impact velocity, the maximum E.A.F.
(for each set of tests in the range of the impact velocities considered) appears to occur
impact closer to the "perfectly plastic" situation. This should be expected since a lower
value of the coefficient of restitution causes a slower rebound velocity, which ultimately
synbolizes a test in which the steering wheel is absorbing the maximum amount of
energy it can. This value varies slightly for each specific impact velocity. Most
significantly, it has been observed that the maximum E.A.F., for each specific impact
velocity, occurs with very similar testing conditions. Table 2 lists the tests with the
maximum E.A.F., for each impact velocity considered, the corresponding coefficient of
restitution, the wheel angle, and column angle specified in the test.
Table 2 - Maximum E.A.F. with Correswndine Coeficient of Restitution
and Wheel and Column Andes for Each Test Conducted
Within the Range of Immct Veloçities Considered
5.6.1 1. The Effect of Column Angle and Wheel Angle on the Peak Impact Force
Figures 52, 53, 54,55, 56, and 57 illustrate the effects of colurnn angle and wheel
angle on the peak impact force for impact velocities of 8 rnph [3.58 m/s], 9 rnph
[4.02 m/s], 10 rnph [4.47 m/s], 11 rnph [4.92 m/s], 12 rnph [5.36 rn/s], and 13 rnph
[5.8 1 d s ] , respectively. The peak impact force is calculated using equation ( 1 1).
M e e i and Cdumn Andes Versus Peak lm~actForce
at 8 mph (3.58 m/s) lmpact Veiocity
Comparing the surface plots for the peak impact force to that of the peak vertical
Ioad, one finds that the topography of the surfaces are almost identical for each specific
impact velocity. This indicates that magnitudes of the forces in the "x-axis" and "y-awis"
are generalty considerably smaller than the load along the "z-axis" direction.
Similar to the peak vertical load surface plots, the peak impact force also varies
over the location of impact on the steering wheel and the orientation of the steering
wheei. Furthemore, the peak impact force is also dependent upon the impact velocity,
since changes in the impact velocity significantly affect the magnitude of the peak
impact load.
5.7. Data Acquired and Calculated for Each Experimental Test Conducted
Appendix B documents al1 the experimental and calculated data for each test
conducted in this investigation. The tests are arranged based upon the test number
(Test 8) corresponding to a specific dropping height, impact velocity. wheel angle and
column angle.
6. FINITE ELEMENT MODELING AND SIMULATION OF THE
EXPERIMENTALTESTS
6.1. The Software Selected for the Simulation of the Impact Experiments
FE software program for the simulations that it must conduct. For the purposes of this
investigation, the FEA software must incorporate highly non-linear material models,
large deformation capabili ty, dynamic analyses, and contact between deformable bodies.
However, the experimental testing utilizes a polyurethane foam chestfom that absorbs
energy, and an elastic-plastic alurninurn steering wheel. Thus the FEA software must
incorporate some matenal models which will consider both foam and an elastic-plastic
alurninurn alloy.
As well, a large number of FEA programs also consider contact between bodies.
deformable steering wheel. Thus the numerical contact algorithms, within the FEA
situations.
generally refers to the requirement of an explicit finite element code. Explicit FE codes
do not require the inversion of a stiffness matrix; however, since a central difference
scheme is generally used to detemine velocities and accelerations, corn calculated
found, a small timestep, with which to integrate (or move) through in time in the
dynamic analysis, must be used. Unfortunately, this last requirement generally increases
A vanety of explicit FEA programs are available in indus- today. For the
purposes of this investigation LS-DYNA was selected, due to its excellent contact
capabilities. non-linear material models (including both foam and highly non-finear
LS-DYNA is simply a numeric solver to an input file (or input "deck") which the
user specifies. Thus, a pre-processor is required for the developrnent of the finite
contact algorithms, and time history information of specific nodes, al1 of which are to be
analyzed in a typical simulation. The pre-processor utilized in this study was Finite
Element Model Builder (FEMB), which also served as a pst-processor for analyzing the
6.2. The Finite Element Models Developed and Used in This Study
To best simuiate the experimental testing procedure, finite element models were
developed for the three spoke steering wheel, the chestform, the wooden backing plate
The aluminurn crosshead serves two purposes on the droptower testing machine.
First, it connects the guideposts to the chestform, which provides a means of motion for
the chestfom. The crosshead allows the bodyforrn to translate along the yideposts, only
verticalty upward or downward. Secondly, it adds rnass, and hence weight, to the
dropping assembly (consisting of the crosshead, wooden backing plate, and the
chestfom). Thus, the aluminum crosshead provides a means of constraining the motion
of the chestform, and adds a considerabIe amount of mass to the entire dropping
assembiy. The mass of the entire dropping assernbly (including the crosshead, wooden
backinç plate, and chestforrn) on the experimental testing machine, has been found to be,
104.06 Ibm [47.18 kg] (corresponding to a weight of 104.06 lbf or 462.87 N).
9.86 Ibm [4.47 kg3 (corresponding to a weight of 9.86 lbf or 43.88 N). Experïmentally,
rneasuring the masses (or weights) of the individual crosshead and ~voodenbacking plate
was impossible. However, the di fference between the entire dropping assembly mass
and the chestform mass will give the mass of the crosshead and wooden backing plate.
This was calculated to be 94.16 Ibm [42.71 kg]. Thus, for the FE models of the
al uminum crosshead and wooden backing plate, the combined masses are to total
aluminum plates, steel plates, bushings, bushing holders, bolts and rubber bumpers. To
mode1 the entire aluminum crosshead is unnecessary since the crosshead serves only as a
means of constraining the chestform and adding mass to the dropping assembly. Thus, a
Based upon the geometry of the actual alurninum crosshead a FE model was
developed and is illustrated in Figure 58. The FE modet, identical to the actual
aluminum crosshead has a rib that is rigidly fastened to the wooden backing plate. Also,
in the FE model of the aluminum crosshead, there are a number of large cubic elements
near the centre of the structure, which act as the lumped mass used in this FE model.
the aluminum crosshead was moditied so that, afier inclusion of the wooden backing
plate, the entire dropping assembly (excluding the chestfonn) had a mass of 94.16 Ibm
142.71 kg]. Appendix C lists the materiai properties of the aluminum crosshead and data
FE mode1 so that oniy translational motion is permitted in a direction toward the steering
wheel. These boundary conditions simulate the constrained motion of the actual
which occurs within the structures of the droptower testing machine, involve distortions
of the chestform and steering wheel. Although, deflection of the wooden backing plate
may be occurring, they are probably negligible with respect to the magnitude of the
deformation occumng within either the three spoke steering wheel or the chestform.
Thus, in the FE model of the wooden backing plate a rigid material mode1 was utiI ized.
Young's Modulus, and an estimate for Poisson's ratio (for wood) was taken to
respectively be 1595.4 kpsi [ I l GPa], and 0.33 from reference [37]. An approach sirnilar
to the aluminum crosshead was completed for the determination of the density of the
wooden backing plate. Appendix D lists the material properties of the wooden backing
Figure 59 illustrates the wooden backing plate FE model (with the aluminum
crosshead FE model included). The creation of the FE model of the wooden backing
plate \vas conducted after completion of the FE model of the chestform. This is
pnmarily because the profile of the wooden backing plate is identical to the profile of the
Figure 59 - Finite element m d e l s of the wooden backing plate (green) and the
aluminum crosshead (blue).
back portion of the chestform, thus, simple element copying (with an offset) was used to
determine the profile of the wooden backing plate. Finally, a drag mesh operation was
completed to produce the solid elements which define the w d e n backing plate.
It should be noted that the wooden backing plate and aluminurn crosshead are
rigid bodies with some nodes which are shared (merged) by the two parts. In LS-DYNA,
been utiiized in al1 simulations conducted where the wooden backing plate and
the geornetry of the FE chestform model, and experimental testing to determine the
material properties of the polyurethane foam (which is the material of the chestform).
The geometry of the FE chestform model was developed using the dimensions
and illustrations provided in the SAE J944 laboratory testing procedure. The two
dimensional outline of the chestfiorm's front and side profile was optically scanned, using
a typical persona1 computer scanning device, and saved as a bitmap image. The bitmap
image was then imported into Corel OCR-Trace, which is a software package which
vectorizes Iine segments fiom bitrnap outlines. Vectorization is a process which takes
Figure 60 - The cross-sections used in development of the finite element model of the
chestform.
the small pixels from the bitmap image, interpolates the connectivity of the pixels and
vectorizes or makes an entire line from the individual pixels. The bitmap file was
vectonzed. manually modified, and saved as an Autocad drawing file (a " . D W file).
The Autocad drawing file was imported into Cadkey 7.5 1, where cross-sections of the
chestform were developed. Figure 60 illustrates the cross-sections taken for only half of
the bodyform (the remaining half was completed by rnirroring the sectional splines). The
principal reasoning why al1 the spline drawing was completed in a CAD software
package (Cadkey) was due to the fact that FEMB (the pre-processor used in this
investigation) has considerable limitations in drawing and editing spline entities. Afier
the final spline drawing was developed in Cadkey, the file was exported as an IGES
(Initial Graphics Exchange Standard) file, which was then imported into a FEMB
session.
After importing and setting up the IGES file, development of the finite element
model took place. The splines, which were created in Cadkey, were used to develop the
finite eIements from drag meshes, and solid meshes. A finer mesh distribution was
developed from the neck portion to the lower abdomen region of the chestform. The
reason for this was to ensure a good discretiziation of the actual bodyform in the region
where contact will occur between the steering wheel and chestform which, ultirnately,
will give more accurate results in FE simulations. Figure 6 1 illustrates the completed
bodyform FE model (with the wooden backing plate and crosshead also shown).
Figure 61 - Finite element models of the chestform, wooden backing plate, and
crosshead.
It should be noted that the nodes in the intefiace between the FE models of the
chestform and the wooden backing plate are not coincident or merged nodes. There
exists a very smail gap between the two rnodels. The nodes along the very back of the
chestform, directly adjacent to the \vooden backing plate, are constrained to move with
the motion of the two merged rigid bodies (the crosshead and wooden backing plate).
command in the LS-DYNA input file. Using this command places constraints on the
nodes on the chestform, closest to the wooden backing plate. in experimental tests, the
bodyform is significantly secured to the wooden backing plate, using adhesive tape and
velcro strapping- Constraining the nodes along the back of the chestfonn simulates the
adhesive tape and velcro strapping which occurs in experimental testing. No other
constraints are placed on any other node in the chestform FE model. The entire final
bodyform model consists of 15251 nodes, and 12517 elements, 12290 of which are
hexahedron type elements, and 227 of which are wedge type elements.
The SAE 5944 standard outlines the required material characteristics of the
of the bodjorm. These experimental tests were conducted at the University of Windsor
the material properties, of the chestform, in a dynamic situation is necessary, since this
would provide matenal properties, for the FE model, at similar experimental testing
chestform in a dynamic situation, were conducted with the droptower testing machine.
The chestform was dropped fiom three different heights and collided with a rigrd
cylinder (usinç the same setup for checking the integrïty o f the chestform, illustrated in
Figure 9). A vertical load versus vertical displacement curve was experimentally
determined and used to deveiop an engineering stress versus engineering strain curve.
The approximate area of impact on the rigid cylinder was determined, which provided a
cylinder and the bodyform. The engineering strain was calculated from equation (13)
based the displacement measwements (Al) obtained by the LVDT, and the initial
Experimentai results and equations ( 12) and (13) were utilized to generate an engineering
stress venus engineering strain cuve for the chestform matenal, which is illustrated in
Figure 62.
100 - - - - .
1
I1 [ 600
.-v,
h
Q
80 - 1 .- 500
I'
Y
V)
Y,
E 60 - 400
5 /
0,
C
/'
/ 300
-5 40 -
0, /'
.-C ,
, 200
:- 100
. O
0.0 O. 1 0.2 0.3 0.4 0.5 0.6
Engineering Strain ( M n ) (m/m)
first selected as a possible candidate for the polyurethane foam was LS-DYNA matenal
Additionally, optional information may be provided, but for the purposes of this study
these material characteristics are not needed and cannot easily be determined [38].
The density of the material was determined expenmentally by taking rnass measurements
and finding the volume of the FE model. Young's Modulus was determined with the aid
of the engineering stress versus engineering strain curve. Thus, only determination of the
hysteretic unloading factor and the shape factor are needed to be determined.
conducted and compared with experimental tests to find these two material
characteristics 1391. AAer preliminary simulations were conducted and compared with
experimental tests, the optimal values of the hysteretic unloading factor and the shape
factor were found to be 0.39 and 5.00 respectively. Appendix E summarizes the material
properties of the polyurethane foam and includes data from the LS-DYNA input file
The development of the FE model of the three spoke steering wheel was
the FE model was completed then the material properties were determined.
A cornplete wirefiarne drawing, for the three spoke steering wheel, was provided
by Centre Tool, who is responsible for the development and manufactunng of steering
Figure 63, was imported into a Cadkey session. There existed a number o f duplicate
-98-
lines and splines in the CAD model, al1 of which were deleted in the Cadkey session-
Afier considerable work the CAD rnodel was exported as an IGES file and imported into
Since FEMB provides no "automatic" solid mesh generation procedure, the entire
FE model of the three spoke steering wheel was completed manually, which
and solid rneshing were completed several times to eventually produce the completed FE
model of the steering wheel. The six thin ribs on the steering wheel, which stiffen the
joint between the 3 spokes and the outer rim of the steering wheel, were initially al1
modeled using shell elements. A later version of the FE model was created, in which the
ribs were modeled using solid elements. Afier initial verification testing, which is
discussed in section 6 . 3 2 3 , it was observed that there was no difference in modeling the
ribs as either solid or shell elements. Figure 64 illustrates the completed FE model of the
three spoke steering wheel. The completed FE rnodel of the three spoke steering wheel
consists of approximately 2248 1 nodes and 16676 elements, 15406 of which are
2-
Figure 64 - Finite element model of the three spoke steering wheel with the rim, three
spokes and hub in the colour red and the thin ribs in colour green.
As already mentioned, the tbree spoke steenng wheel is made from a proprietary
aluminum alloy used specifically for steering wheels. The majority of the material
properties for the aluminurn alloy were available from K.S. Centoco. Tensile testing was
selected for the numetical rnodeling of the aluminum alloy. The basis for this selection
was due to the fact that McKie [9] and DuBois [39] used this material model as a means
Two parameters needed for strain rate considerations (if these two pieces of
information are not provided, strain rate effects are not considered).
Al1 the information necessary for this material model either was available or
could be calcuiated. The only two points of concem were the strain rate constants and
model. The stress is scaled according to the values of the strain rate and the Cowper-
test, D and q are the strain rate parameters. For ahminum alloys, Bodner and Symonds
have found that the vaiues for D and q are 6500/s and 4 respectively 1401. If necessary,
strain rate effects can now be considered in the simulation of the chestform impacting
To develop the stress versus effective plastic strain curve requires two main steps.
Firstty, equations (1 5) and (16) are used to transform the engineering values of stress and
and "e," is the engineering strain. Once the true stress and true strain have been
determined, then the stress is simply equal to the tme stress (in the plastic regime) and
The basic idea behind the conversion from the engineering stresdstrain curve to
the stresdeffective plastic strain curve 1s to eliminate the elastic regime of the true
stressktrain curve. Figure 65 illustrates the stress versus effective plastic strain curve
used in the material model for the steering wheel. Appendix F summarizes the material
properties of the steering wheel aluminum alloy and includes data from the LS-DYNA
Figure 65 - Stress versus effective plastic strain curve for the aluminum
alloy of the three spoke steering wheel.
6.3. Verification of The Cbestform and Three Spoke Steering Wheel Finite
Element Models
Rather than combining al1 the FE models together and begm simulations, a
verification of the two deforrnable FE modcls was considered. Firstly, the dropping
assembly, consisting of the chestform, wooden backing plate, and crosshead, were
cylinder. The principal idea behind these simulations is to investigate if the correct
materiai properties and degree of discretization were utilized in the chestfonn FE model.
Cornparisons between the experimental and FEA predictions for the vertical load versus
crosshead displacement curves were used to detect any differences in either the loading
or displacement profiles existing between the experimental and numerical tests. Since
the chestform is the only deformable body in this situation (with the chestform impacting
the rigid cylinder), the only variable that need be considered, in this situation, is the FE
model of the chestform, and not the FE model of the three spoke steering wheel.
Experimental and numerical simulations were then conducted with a rigd plate
impacting the deformable three spoke steering wheel. The polyurethane foam chestform
was replaced with a thick, rigid, alurninum plate, which simuIates a rigid body irnpacting
the deformable steenng wheel. The procedure only considers the steering wheel as a
deformable body and thus will verify the material properties and discretization of the
where conducted with the deformable chestform impacting a rigid cylinder. A rigid
cylinder, with a holding stem, was fabricated and rigidly mounted in the droptower
testing machine.
chestform. The setup and procedure is identical to the process used to veriw the integnty
of the chestfom dunng the 420 experimental tests. The procedure involved dropping the
chestform, and attached accessories, from three different heights to generate three
Adhesive Tape
00 hold chesrform io wooder
backplate)
-RÏgzd Cylinder
-Rigid Cylinder stem
Figure 66 - Experimental setup used to veriw the finite element model of the chestform.
different impact velocities. The impact velocities for the experimental tests were
6.85 mph [3.06m/s], 8.82 mph [3.94 d s ] , and 10.77 mph [4.81 m/s]. The data files
acquired from the tests were then modified, using a Visual Basic program, and anal yzed
in a Mathcad worksheet.
The chestform, wooden backing plate, and crosshead FE models were combined
with a FE model of the ngid cylinder. Figure 67 illustrates al1 four parts used in the
process of veri@ing the chestform FE model. Al1 nodes of the rigid cylinder, which were
modeled using shell efements, were totally constrained from motion and a rigid material
chestform and the rigid cylinder. The chestform, and the attached accessones, were
differences in the three experimental impact velocities (6.85 mph [3.06 m/s], 8.82 mph
values of the vertical load versus crosshead displacement curves are presented in
Fi y r e 68. Curves at different impact velocities are offset in the x-axis (crosshead
vertical Load versus Crosshead Disdacemerit
Crosshead D i c e m e M (rnetres)
425 4.10 4.15 4.10 4.05 0.00 0.05 0.10 0.15
should be noted that in detennining the percentage difference between experimental and
Percentage Djfference =
(LS-DYNA Resuit - Experimentul Resrrlr )
Experirnental Result
Table 3 - Percentaee Di fference Between Ex~erimental
and Numerical Simulation Results for
Peak Vertical Load and Peak Crosshead Dis~iacement
Impact
Velocity 1 Peak Vertical Load (lbf
kW)
Peak Crosshead
Displacement (in
% Diff.
on Peak
Vertical
% Diff. on
Peak
Crosshead
Load Displacement
Exp. LS-DYNA Exp. LS-
(%) (Oh)
Experimental testing was not conducted at higher speeds as there [vas concern
that, at higher impact velocities, damage may occur to the testing machine. Typically,
peak vertical Ioads, for impacts between the deformable chestform and deformabte
steering wheei, generally are not higher than 3500 lbf [lS.57 W. This value is
considerabty less then the peak vertical load, with the chestform impacting the rigid
cylinder, at 10.77 mph [4.8 1 m/s]. Hence, the range on impact velocities in this
vertical ioads that will occur in actual experimental impacts between the chestforrn and
Based upon the findings presented in Table 3 and Figure 68, an excellent
at lower impact velocities, the percentage difference in the peak vertical load between
- 108-
experimental and numencal simulations approaches 16%. It is felt that this value is
acceptable due to the high degree of non-linearity in the material model of the
pol yurethane foam. In al1 three cases, the magnitude of the percentage difference in the
peak vertical displacement of the crosshead is very low, indicating a good relationship
elements (mesh sensitivity study) and a fulty integrated solid element formulation
The FE model of the chestform was modified so that an additional 1328 elements
were added to the chestform, in the area where contact occurs between the bodyform and
rigid cylinder, and an impact simulation was conducted at 8.82 mph [3.94 m/s]. It should
also be noted that only hexahedral elements were used in this chestform FE model.
integration points within each solid element, was also varied. In past numerical
simulations the LS-DYNA standard, under-integrated, solid element was used (element
velocity of 8.82 mph [3.94 d s ] the element formulation was changed so that a fully
integrated solid element, with 8 integration points, was used. To illustrate if any
differences were an'sing due to changes in the FE mode1 and the element formulation,
three curves of the vertical load versus crosshead displacement were plotted on the same
Vertical Load Versus Crosshead üisolacement
For lm- Velocihr of 8.82 mph 13.94 mlsl
JFEd e l Cornnarison)
Crosshead Displacement (rnetres)
-002 000 002 004 O06 008 O10 O12
t
-Element Fomnilabon #1 mtfi Hex. 8 Wedge Elements
-Element Formulabon #l wiih al1 H e n Uamnts
-Element ForrnulatJon #2wrth Hex & M d g e Elemnts
-
-1 O 1 2 3 4 5
Crosshead Displacement (in.)
Figure 69 - Cornparisons between the two modified finite element models and
the original finite element model.
graph, as illustrated in Figure 69. The black colour curve illustrates the findings from the
normal simulations, the red curve illustrates the findings for the entire hexahedral mode1
with added elements, and the green curve illustrates the findings for the FE model using a
From Figure 69, it was concluded that, firstly, adding more elements to the
chestfom FE model did not effect the results, and, secondly, the element formulation
had no effect on the results. However, there was a significant increase in computation
time observed. Table 4 summanzes the chestfom FE model specifications, the number
of elements, the element formulation implemented, and the approximate run time to
-
Table 4 Summary of the Three FE Models lnvestieated
Element Formulation
FE model Number of Elements Associated with Each Approx,mate
Specifications Type of Element l
Run Time
--
,
Since little or no changes were evident by altering the FE model, and simulation
tirnes were significantly tess, the first FE m d e l developed was selected as the model for
between experimental and numerical results was observed for this FE model.
6.3.2. Verification o f the Three Spoke Steering Wheel Finite Element Model
To verifL the three spoke steering wheel FE model, experimental tests and
nurnencal simulations were conducted with a rigid plate impacting the deformable three
spoke steering wheel. A rigid plate, fabricated by the author, was designed such that it
would replace the deformable chestform and rigidly mount to the aluminum crosshead.
the three spoke steenng wheet. The deformable chestform was replaced with a rigid
plate and fastened to the aluminum crosshead. Testing was conducted in a similar
manner as with the deformable chestform impacting the rigid cylinder. Different impact
~WoodenBacking
Plate
Deformable chestform
replaced with rigid
aluminum plate
Experimental tests were conducted at four different conditions. Table 5 lists the
/ Espenmental
Testing
Conditions #
Wheel Angle
(degreedo' dock
position)
Column Angle
(degrees)
Impact Velocity (mph
[mjsl)
7
1 1 *O0 / 6 o~clock
90" 1 3 o'clock
25
25
O
O
7.18 mph [3.2 1 mls]
7.3 1 mph [327 m/s]
The experirnents were completed on the steering wheel impact testing machine and each
data file was modified (using the Visual Basic Program) and analyzed in the Mathcad
worksheet.
A FE model of the ngid plate was developed and combined with the FE model of
the three spoke steering wheel, which is illustrated in Figure 71. A rigid material model
was used for the non-defonnable plate FE model and material properties were based
upon aluminum (see Appendix C). Boundary conditions were applied to nodes of the
rigid plate FE model, to permit only translational motion in a vertical direction, identical
speci@the initial veloçity of the FE rigid plate model. Furthemore, a surface to surface
contact algorithm was used to rnodel contact between the rigid plate and the three spoke
steering wheel.
(wheel angle, column angle, and impact velocity) as specified in Table 5 . The first
involved a wheel angle of t 80" (or impact at the 6 o'clock position), a column angle of
25 O , and an impact vetocity of 7.18 mph [3.2 1 rnls]. Five different FE models were
considered for evaluation at this testing condition. Table 6 details the five different FE
Finite Element
Description
Mode1 #
- Normal FE model.
- Used as a benchmark for other comparisons.
- Ribs of steering wheel modeled as shells with shell nodes
merged to solid nodes at interface.
- Identical model as outlined in model # 1 except mass scaling
was incorporated into FE model.
- Total mass scalingcaused wheel to have 1.2 times itsactual
mass.
- Identical model as outlined in model ff 1 except mass scaling
was incorporated into FE model.
- Total mass scaling caused wheel to have 2.0 tirnes its actual
mass.
Mass scaling is a technique used in explicit finite element codes to lower the
computational time of the problem being considered. The critical timestep, which is the
model geometry and the material properties of the modeis. If the mass (or indirectly the
density) is increased artificially, then a solution can be found in a much lower time.
Unfortunately, the effects of mass scaling are sometimes unknown and this is why FE
model # 1 was simulated, with no mass scaling, to see if artificially scaling the mass of
of the three spoke steering wheel. By using shell elements, and merging the nodes of the
shell eIements to the solid elements (at the interface between the two dir'ferent element
types), no rotational constraints are placed on those shell element nodes. Ultimately, this
elements. However, this cornes at a cost. This contact algorithm also requires
cornputational time. Thus, to see if the effects of the rotational constraints are significant
Finally, to see if solid elements could be used to model the rib sections, FE
model #4 was simulated. By modeling the ribs as solid elements, nodes could be merged
without any concerns for rotational constraints. However, since only one and two layers
of solid elements where used in the modeling process, fully integrated solid elements
(element formulation CC2 in LS-DYNA) were implemented for these solid elements only
To venQ the FE model of the three spoke steering wheel, corn pansons between
the experimental tests and the numerical simulations were conducted on the basis of the
graphical methods only, the percentage difference (between experimental and numencal
tests) was not calculated for the rigid plate impacting the three spoke steering wheel as
significant numerical noise was present in the resultant contact interface forces between
the FE ngid plate model and the FE three spoke steering wheel model. Numerical noise,
which is generally common in FE simulations with materials that exhibit little damping
(which iç the case here), make it difficult to approximate the peak vertical load and
Figure 72 shows the vertical load versus crosshead displacement curves for two
Testing Conditions:
-
Colurnn Angle 25 degs
t O
-
Wheei Angle 180 degs (6 o'clock impact) - -2500
-
Impact Velocrty 7.18 rnph / 3.21 mis
I I I 1 1
-2 -1 O 1 2 3 4 5
Crosshead Displacement (inches)
including mass scaling in the FE model of the steering wheel has only a minor effect on
the results. However, a significant difference in computational times was observed and
these are listed in Table 7. Thirdly, results are independent o f whether the ribs are
modeled as shell elements o r solid elements, since no change in the load versus
38 hours
Based upon the comparisons between the experimental and numerical fmdings,
FE model #4 was selected as the model which would be used in the simulations
between the rigid plate and the steering wheel. Although FE model #3 did exhibit a
lower computational tirne, FE mode1 #4 was selected since al1 elements are similar (i.e.,
solid elements), with the exception that fulty integrated elements are used in modeling
Figure 73 shows the vertical load versus crosshead displacement curves for an
conditions #2 (outlined in Table 5). As discussed previously, FE mode1 type #4 was used
in the simulation. Again a very good agreement between experirnental and numerical
findings is evident.
I I l l ! 1
-- FE
Experirnental Test
Simulation
1
4
7500
f 1
Tesbng Conditions
1 -
Colurnn Angle 25 degs
-
Wheel Angle 90 degs (3 o'clock impact)
-
l m p a d Valocrty 7 31 mph 1 3 27 m h
'4 2
.< u
4.74 mph (2.12 m/s] which is slower than previous tests, yet, a good numerical prediction
O 1 2
Crosshead Displacement (inches)
experimental and numerical tests completed based upon testing conditions #4 (outlined
in Table 5). Again a good relationship between expenmental and numerical results is
evident. However, the peak vertical deflection differs from expenmental findings by
4 inches [IO1.60mm]. For the amount of deformation occuming, and the highly
non-iinear material modefs used in these types of simulations, this amount of difference
Experimental Test
FE Simulation
Two other investigations, concerning the verification of the FE model of the three
spoke steering wheel, were addressed in this study. First, the effects of strain rate on the
material model of the steering wheel were considered. Secondly, a sotid deformable
plate model was created and tested at identical conditions to see if modeling the rigid
By incorporating the strain rate parameters for the aluminum alloy into the input
file for the simulation, strain rate effects are automatically considered. Figure 76
illustrates the vertical load versus crosshead displacement curves for two experirnental
tests, the FE simulation conducted without considering strain rate effects, and an
additional cuve with strain rate effects considered in the numerical simulation.
Vertical Load Versus Crosshead Dis~lacement
Riaid Plate Impactinca 3 S ~ o k eSteerina Wheel
Com~arisonsBetween Ex~erimentaland
Finite Element Simulation Results
Crosshead Oisplacement (metres)
-0.04 -0.02 0.00 0.02 0.04 0.06
-Experimentaf Test #l
Expanmental Test #2
FE Sim. #l (without Strain Rate Effects)
- FE Sim #2 (with Strain Rate Effects)
-
Column Angle 25 degs
-
Wheel Angle 180 d e g s (6o'clack impact)
Based upon the findings presented in Figure 76, it is evident that strain rate
effects are a factor in the FE mode1 of the three spoke steering wheel. Results are closer
to actual experimental findings and thus indicate that strain rate effects are a necessary
deflections.
The ngid plate was also modeled as a true solid entity using solid elements and an
elastic, deformable material model was incorporated into the modified FE model of the
utilized, and resuits, based upon the vertical load versus crosshead displacement curve,
were compared with the experimental and past simulation findings. Figure 77
graphical ly il lustrates the results frorn two experimental tests; one simulation completed
-Experimental Test #1
- -- Experimental Test #2
FE Sim. #l (Rigid Plate as Shells)
- FE Sim. 12 (Rigid Plate as SolidsIDef.)
Testing Conditions-
Column Angle 25 degs -
-
Wheel Angle 1 8 0 degs (6 o'cfock impact)
-2 -1 O 1 2 3 4 5
Crosshead Displacement (inches)
completed with the rigid plate modeled as a solid, deformable, elastic body.
The resuits, presented in Figure 77, indicate that only a slight difference in peak
crosshead displacement occurs if the rigid plate is modeled as a solid and deformable
entity. Hence, the approximation of modeling the experimental rigid plate as a rigid
with a deformable chestfonn. Both the deformable chestfonn and the deformable three
spoke steering wheel FE models were verified by perfonning numerical simulations and
spoke steering wheel were then conducted and compared with the past experimental
testing results.
The experimental test setup \vas as previously illustrated in Figure 10. The
LS-DYNA input file consists of the deformable steering wheel, deformable chestform,
the wooden backing plate, and the aluminum crosshead FE models. Figure 78 illustrates
al1 the FE models used in simulating the chestform impacting the three spoke steering
wheel. It should be noted that mass scaling was invoked for al1 simulations only to an
extent where the mass of the steering wheei was only adjusted. Alteration of mass of the
chestfonn was not perfonned in any numerical simulations (see Appendix G).
- - --
Figure 78: Al1 four finite element models used in simulating the chestform impacting
the deformable steering wheel. The testing conditions for this exact
numerical simulation involve a column angle of No, a wheel angle of 120 O
For this investigation, two dual Pentium-I'I based machines, running at interna1
dock speeds of 300 MHz and 450 MHz were utilized. Based upon initial simulation
testing (completed on the dual Pentium-11450 MHz machine) it was observed that the
computational time to evaluate a single impact test was approximately 85 hours. This
included mass scaling as a means of decreasing the computational time. Thus it was
proposed to simulate twelve different experimental testing situations and compare the
that the first seven FE analyses should investigate the effects of wheel angle, column
angle and impact velocity in a FE simulation. Numerical simulations were to be
conducted by varying the impact velocity, while holding the column and wheel angles
constant. This same procedure was to be repeated by varying the wheel angle, while
holding the impact velocity and colurnn angle constant, and by varying the column angle,
while holding the impact velocity and the wheel angle constant. Figure 79 illustrates 3
orthogonal axes, one axis representing impact velocity, another axis representing the
column angle, and a third mis representing the wheel angle. This illustration is an aid to
velocity, column angle, and wheel angle. Test #2 is completed by increasing the impact
Figure 79 - Three orthogonal axes used to illustrate the testing methodology for
the first seven numerical simulations.
velocity, and holding the column and wheel angles constant. Test # 3 is completed by
decreasing the impact vetocity, and holding the column and wheel angles constant. This
procedure is repeated to also consider variations in the wheel and column angles.
Table 8 lists the first seven testing situations and an additional five situations that were
In attempting to simulate the impact between the deformable chestfonn and the
deforrnable steenng wheel, two significant problems were encountered. The first
problem dealt with hourglassing occumng within elements of the chestfonn, and the
second problem involved concems regarding the contact between the chestfom and the
of elements displace but produce zero strain energy. Hourglassing occun in fini te
eIements which are underintegrated, or in the case of LS-DYNA, elements that possess
only a single integration point. For this study, both the chestfonn and three spoke
control is çenerally less computationally expensive and for this reason was implemented
For the chestfonn FE model the Flanagan-Belytschko stiffiiess form with exact
control are recommended for deformations occumng at lower velocities (the impact
velocities considered in this study fa11 within the definition of "lower velocities").
Furthemore, for solid elements, the exact integration provides some advantage for highly
distorted elements, which was the case for elements in the chestform [44].
"negative volume" error. This error arises when nodes on one face (or surface) of an
element pass through another face (or surface) of the same element, thus inverting the
element and leading to a "negative volume". This enor was due to the amount of
deformation and hourglassing occumng within the chestfonn. Simulations were also
conducted with a further refined mesh of the chestfonn, nonetheless, the error
model, simulations terminated normally and the deformation occumng within the
chestform appeared more realistic. Figure 80 illustrates a section of the chestfonn dunng
has been removed for illustration purposes). Hourglass control was not implemented in
the chestform as it is seen in Figure 80. To illustrate the effects of the hourglass control
Figure 8 1 shows the defonned chestform, under similar testing conditions (except
Figure 81 - Deformation of the chestforni finite elernent rnodel with hourglass control
implemented (the three spoke steering wheel has been removed to better
illustrate the deformation ofjust the chestform).
considering the entire chestform FE rnodel), with hourglass control utilized. Note that
the deformation appears more realistic in Figure 8 1. Appendix G provides sample input
The contact type used in al1 simulations (invofving contact between the chestform
and the steering wheel) was a SURFACE-TO-SURFACE algorithm that is based upon
the penalty method. This method consists of placing normal interface springs between
al1 penetrating nodes and the contact surface. The stiffness of the spnngs is related to the
material properties of the two entities in contact [45]. Since the material properties of
the chestform and the steering wheel significantly Vary, the penalty factors, which adjust
the stiffness for the springs used in the contact, had to be significantly modified from the
Figure 82 illustrates contact without the proper penalty values used in the
chestform and steering wheel FE models. Note that the elements from the bodyforni are
Figure 82 - Contact without the proper values of the penalty factors for the
chestform and steering wheel. Note that the chestform elements
are penetrating into and through the steering wheel.
Figure 83 - Contact between the chestform and the steering wheel with the proper
val ues of the scale factors for the SURFACE-TO-SURFACE contact
algorithm.
actually penetrating into the steering wheel. By implementing more appropriate values
of the penalty scale factors, which can only be determined on a trial and error basis, a
better and more realistic contact is achieved as illustrated in Figure 83. Appendix H
provides sample input data regarding the implemented contact algorithm frorn LS-DYNA
simulations.
To illustrate the geometries of the FE models before and afer the impact,
Figure 84 and Figure 85 show the three spoke steering wheel and the chestform prior to
the collision and after the impact respectively. In addition, the computational times to
complete al1 simulations for each specific test are presented in Table 9. It should be
noted that the addition of strain rate effects did not significantly affect the computational
-- --
Figure 84 - Geometry of al1 the finite element models before impact between the
chestform and the three spoke steering wheel. This numerical simulation
was completed with a column angle of N o , a wheel angle of 180 and an O,
-134-
Table 9 - Com~utationalTimes Reauired to Corndete
the Twelve Numerical Simulations
Note: * - indicates simutations conduaed o n Dual Pentium II 450 MHz machine with 384 Megbytes
RAM with serial version o f LS-DYNA
** - indicates simulations conducted o n Dual Pentium i I 300 MHz machine with 384 Megbytes
R A M with serial version of LS-DYNA
experimental testing results with the predicted numencal simulation findings. These
curves illustrate differences in either the vertical loading profile ancilor the crosshead
displacement profile.
Numerical simulations were completed two times for each test situation as
presented in Table 8. The first simulation was completed without considering strain rate
effects in the aluminum alloy of the three spoke steering wheel and a second simulation
was conducted with strain rate effects included in the numerical analysis.
Table IO lists each numerical simulation, with its corresponding testing
conditions, and the figure number, il lustrating the comparison between the experimental
6
33"
24"
120" / 4 o'clock
60" 12 o'clock
12 mph 15-34m/s]
12 mph [5.36 m/s]
90
91
1
7 24" 180" / 6 o'clock 12 mph [5.36 d s ] 92
*
8 18" 150" / 5 o'clock 13 mph [5.81 m/s] 93
9 21 O 30" 1 1 o'clock 10 mph [4.47 mis] 94
-
- Experirnental Test
- FE Sirnulauon (without SIR Etfacts)
1
- FE Simulation (with SIR Effect.)
i
C
0
e
1000
6
-
:Il /i
O
2
-O
8
> 50:;
Tasong CondRions
Column Angk - 24 degs
-
Whaol Angim 120 doos (4 a'clock imp8ct)
-500
-
Impact Voloc~ty t 2 mph 1 5 36 mls
-2 O 2 4 6 8
Crosshead Displacement (inches)
i
Tesbng Condibons:
- --
Column Anglo 24 dogs
Whoal Anglo 1 2 0 dogs (4 o'clock impact)
-500 . Impact Volocity 13 mph 1 5 . 8 1 ml$ - - -2000
2000
O
>
-2 O 2 4 6 8
Crosshead Displacement (inches)
J --- --
Experimental Test
FE Stmulation (wtthout SIR Efïccts)
FE Simulation (with SIR Effects)
Testing Conditions
-
Column Angk 15 deus
O 2 4 6
Crosshead Displacement (inches)
-138-
Vertical Load Versus Crosshead Dis~lacernent
Chestform Impactinri 3 Snoke S t e e r i n ~Wheel
Cornparisons Between Exnerimental and
Finite Element Simulation Results
Expenmental Test
F E Sirnuktton (wdhout SIR Effects)
F E Simulation (wrth SIR Effccts)
2500 -
- Experimental Test 1
1 - FE Simulation (without SIR Enect.)
2000 - - FE simulation (with SIR ~ t k c t s ) , i
-- -- --
A- -- -----
Testing Conbnions
- ,Column Angk 2 4 digs -
-
Y
1500 -
W h i e l Angle 60 digs Q o'clock impact)
-
u -tmpacl
--- - VolocRy
- 2 mph
-1- -
15 36 mls
(L1
-2 O 2 4 6 8 10
Crosshead Displacernent (inches)
- --- --- - - - - -- -- - ----
- Exparimental Test
- FE Simulation (without SiR Eflacts)
- F E Simulation (wcth SIR Effects)
.- --- --
,
Tosting CondRions
- --
Column A n g k 24 degs
Whsal Angb 180 degs (6 o'cbck impact)
impact Vobcrty 12 mph 1 5 36 m l s
u
m
-
O
4 1000 1
O
E
> 500 ;
-500
-2 O 2 4 6 8
Crosshead Displacement (inches)
I
- Experimental Test
--FE Simulation (without SIR Effects)
- f E Simulation (with SIR Effects)
.- . .
--
Taating Condlions:
Column AngP - 18 dogs
-
Whosl Angk 150 dogs ( 5 o'cbck impact)
-
Impact Velocity 13 mph 15.81 mls
--
- 140-
1 Crosshead Displacement (metres)
- - - - - -
Expanrnmntal Test
FE Simulation (without S I R Effects)
FE Simuhtion (with SIR Effects)
Cohrmn Angb
: Whoel Angle
- 21 dous
- 30 degs (1 o'elack impact)
Impact Vakcity - 10 mph 1 4 . 4 7 mis
- 4000 5
3
- 3000 t
B
O
- 2000 -8
A
O
-e
- 1000 3
-- O
-
- -- ..- - .
Exparimental Test
FE Simulation (without SIR Effects)
FE Simuiatmn ( r a h SIR Effacti)
t
1
Experimentrl Test
F E S i m u l i t ~ o n(wdhout SIR Effects) I'
F E Simulation (wdh SIR Effects)
- 1
Tosting Condittons 1
Column Angle - 38 degs
Wheel Angb - O digs (12 o'clock imp8ct)
-
Impact Vibcity 1 1 mph 1 4 92 m h
6000
Z
- 4000 g
O
- 2000
.-
5
>
O
2 4 6 8
Crosshead Displacement (inches)
-142-
Based upon the data presented in Figures 86 through 97, for the majority of
numerical predictions is evident. The effect of strain rate is ctearly a significant factor in
simulating the impact of the chestform on the three spoke steenng wheel. The results
found from numerical simulations indicate that strain rate effects should be considered in
these impact type situations. In general, numerical simulations, which considered strain
rate effects, better predicted the vertical load versus crosshead displacement curve.
Since little numencal noise is present in the finite element simulations, which
should be expected due to the presence of the viscous foam materia1 mode1 for the
chestform, the peak vertical Ioad and peak crosshead displacement are easy to define.
Thus, the percentage difference, based upon experimental tests, has been calculated and
With regard to peak crosshead displacement, for al1 the testing situations
considered, a very low percentage difference between experimental and nurnerical results
has been found. This observation indicates that FE simulations can effectively and
The average of the percentage difference for the peak vertical load is acceptable
for the degree of deformation that is occumng in both the experimental and numerical
tests. However, two individual numerical simulations (tests #6 and 3 12), whether
considering strain rate effects or not, have a considerably high percentage difference
from expenmental findings. The testing conditions of these two simulations are
somewhat similar. They both involve impact at low wheef angles and at high impact
predict actual experimental results of the peak vertical load. In both cases the
experimental value of the peak vertical load was considerabiy less than the value
Test #
% Difference on Peak Vertical
without strain
Load
with strain rate
Il % Difference on Peak Crosshead
Displacement
without strain with strain rate
rate effects effects rate effects effects
1 9.92 % 2.38 % 7.47 % 3.41 %
-3
3
4
5
6
7
8
9
1O
11
12
For numerical simulations completed at any impact velocity and wheel angles
greater than or equal to 90" (or the 3 o'clock position), or numerical simulations
completed at any position on the steering wheel and at lower impact velocities (10 mph
C4.47 m/s] or less), LS-DYNA accurately predicts the peak vertical load in the collision.
regardless of the impact velocity, orientation, or location of impact on the steering wheel.
Another advantage in numencal simulations is the fact that the total intemal
energy within any FE mode1 can be computed at any instant in time. LS-DYNA, if
requested, will output a file that contains time history information regarding the internal
energy of any FE model in a simulation. The internal energy is the actual energy within
the structure at any given time. It differs significantly from the previous methods of
calculating energy, which required the evaluation of an integral of the vertical force
only an approximation to the actual energy within the structure. However, with FE
simulations, one can easily get information on the actual intemal energy within the
structure. Furthermore, the internal energy is determined based upon each different FE
model. Thus, the energy that just the steering wheel absorbs can now be determined.
This information can be used to determine the "True Energy Absorption Factor", since it
will express the amount of energy the steering wheel absorbs relative to the amount o f
Figure 98 incl udes five curves of energy versus crosshead dispiacernent. The
dark red curve gives the experimental findings for energy (the testing conditions are
identical to simulation test # 1), calculated by means of evaluating the integral of the
O 2 4 6 8
curve is identical to the procedure followed for the experimental test. The remaining
three curves (green, blue, and yellow) give the internal energy calculated in LS-DYNA
for onIy the steering wheel (the blue curve), only the chestform (the green curve), and the
remaining curve is the sum of the internai energies of the steering wheel and the
For the test situation illustrated in Figure 98, the approximation of calculated
energy evaluated for the experimentai test (evaluate the integral of vertical force over
crosshead displacement) appears to be valid, as the sum of the intemal energies within
the steering wheel and chestform are approximately equal to the calculated energy.
However, using the same method to calculate energy, based upon the vertical toad and
crosshead displacement curve from the LS-DYNA simulation, introduces a greater error.
The energy that just the steering wheel absorbs, which is a significant piece of
information for steering wheel designers, can be determined from the blue curve in
Figure 98. The end point of the curve represents the value of the amount of internal
energy, and thus the energy absorbed, by the steering wheel in the collision. For this
testing situation, approximately 3000 Ibf-in [339 JI of energy was absorbed by the
steering wheel. This value is significantly different fiom the value calculated
experimentally, which is approximately 5000 Ibf-in CS65 J] for this situation, a difference
With knowledge of the amount of energy absorbed by just the steering wheel, a
more representative value of the E.A.F. for just the steering wheel can be determined.
Equation (5) can be modified just to consider the absorbed (internal) energy of the
steering wheel and the energy available before impact, as given in Equation ( 19).
The true E.A.F. will give engineers and designers a better value of the actual amount of
To calculate the true E.A.F. of the three spoke steering wheel, requires
information regarding the internal energy within the steering wheel at the end of the
collision. Table 12 gives the intemal energies of the three spoke steering wheel, the
chestform, the surn of the intemal energies within the steering wheel and the chestform,
as well as the calculated value of the energy, based upon integration of the vertical force
without considering strain rate effects) and the corresponding experimental tests. Table
13 presents the identical information for the twelve FE simulations conducted Mth strain
interna1 Energy (lbpin [JI) from Finite Calculated Energy (Ibf-in [JI) of
Element Simulations Steering Wheel and Chestform
Test # three spoke
steering wheel
chestform steenng wheel
and chestform
From Finite
Eiernent
Simulations
I Experimental
From
Tests
-
7
1O
II
With the information presented, regarding the intemal energy of the three spoke
steenng wheel at the end of the impact, in Table 12 and Table 13, the true E.A.F. can be
determined and is listed in Table 14. Aiso presented in Table 14, is the average E.A.F.
calculated for each corresponding experirnental test, and the percentage difference
between the two values (using the experimental values as the accepted values).
-
Table 14 The True E.A.F. Calculated Based U-wn the Internai
Enerw Within Just the Steerina- Wheel and the Ex~erirnentally
Determined Average E.A.F.
1
without fiom Expenmentai Tests "thout with
with strain
strain rate strain rate
rate effects rate effects
eEis
1 1
1
3
3
4
6
7
8
9
10
11
12
Based upon the data presented in Table 14, one observes that a large percentage
difference exists between the true E.A.F. and the experimentally detennined average
E.A.F.. There is only a very small effect on the inclusion of strain rate effects in the FE
mode1 sim dations. Generally, considering strain rate effects will only change the
percentage di fference between the true E.A. F. and the average E.A. F. by approximately
3 %, based upon the percentage difference calculated without strain rate effects.
The results presented in Table 14 are very important. They indicate that by
wheel, the experimental results, fail to determine the amount of energy absorbed by just
the steering wheel or just the chestform. Experimental observations can only provide a
caiculated amount of energy absorbed by both the steering wheel and the chestform.
The true energy absorption factor should be utilized by engineers and designers
when developing more energy absorbing steenng wheels. Thus, the need for finite
simulations, is the percentage of total internal energy absorbed in the collision by just the
three spoke steering wheel or by just the deformable chestform. Equation (20) and
equation (2 1 ) were utilized to determine the percentage of total internal energy absorbed
by the three spoke steering wheel and the deformable chestform respectively:
where, "LE.,-," and "I.E.,hd-" represent the intemal energies of the steenng
wheel and the chestfonn aAer the impact, respectively. Note that the sum of the percent
steenng wheel and the chestform is given in Figure 99 for the twelve dif3erent finite
observed that, regardless of the location of impact, velocity of impact, or steenng wheel
orientation, the percentage of the total energy absorbed by just the steenng wheel is
approximately 70%. Furthemore, the chestform absorbs approxirnately 30% of the total
amount of energy absorbed. Results which are presented in Figure 99 are based upon
simulations conducted with and without strain rate effects being considered. Frorn the
information provided, the percentages of energy absorbed by either the three spoke
This information can also be used by steering wheel designers and engineers to
develop a steerîng wheel that absorbs a greater percentage of the total energy absorbed.
Percentaae of Interna1 Enerav Absorbed
bv 3 Spoke Steerinq Wheel and Chestform
in Finite Efement Simulations
1 2 3 4 5 6 7 8 9 10 11 12
Test #
Figure 99 - Percentage of total interna1 energy absorbed by the three spoke steering
wheel and chestform.
By increasing the amount of energy the steering wheel absorbs, the arnount of energy
absorbed by the chestform automatically decreases. Which implies, for actual drivers of
automobiles, there is less energy transferred to the driver to cause bodily harm.
6.5. CD-ROM Containing Related Files from Finite Element Simulations
and nine different animated results files. Each animation file can be executed from any
persona1 computer with the ability of reading an Audio Video Interleave (*.AVi) file.
sideview 2 of impact.avi
I Side view of the impact between the chestform and the steering
wheel FE modeis.
7.1. Conclusions
1. The completion of extensive experimental tests, and the analysis of the data
acquired from these tests, dealing with impacts between a deformable chestform
2. The finite element modeling and simulation, using explicit FE code, of the
Conclusions dealing with the experimental tests will be presented first, followed
1. The vertical load profile during the impact between the chestform and the three
spoke steering wheel is a function of the wheel angle, column angle, and impact
velocity. With increases in impact velocity the peak vertical load also increases.
associated with the experimental test. Experiments have shown that only a minor
variation in the crosshead displacement occurs for changes in the column angle or
impact velocity.
3. The calculated energy absorbed is also a complex function of the wheel angle,
introduced, and is termed the energy absorption factor (E.A.F.). Two methods to
calculate the E.A.F. have been presented, both methods gïve nearly identical
results. An average E.A.F. is calculated as the mean value of the E.A.F. based on
energy at impact and the E.A.F. determined fiom the potential energy before the
experimental test.
5. The E.A.F. is a complex function of the column angle, wheel angle, and impact
veiocity considered in the experimental test. Based upon testing observations the
three spoke steering wheel had an approximate maximum average E.A.F. of 0.83.
This generally occurred for impacts dealing with collisions occurrïng with a
6. The coefficient of restitution for impacts dealing with the chestform and the three
spoke steering wheel has been calculated. The coefficient of restitution varies
more significantty with a given column angle and wheel angle at higher impact
considered.
2. The effects of mass scaling on the finite element models considered in this
3 Hourglass control and the contact penalty scale factors must be selected
appropriately for contact types involving two entities with significantly different
material properties.
determined. The tme E.A.F. utilizes the intemal energy of the steering wheel to
determine a more representative value of the E.A.F.. The true E.A.F. should be
wheel.
From finite element simulations, as the sum of the intemal energies of the steering
Thus, finite element analyses should be an integral part of the design and
7. Based upon finite element results, the steering wheel absorbs approximately 70%
of the total amount of absorbed energy in an experimental impact test and the
chestform absorbs the remaining 30%, regardless of the wheel angle, column
Although every attempt has been made to present the most important information
observed from the experimental impact tests, there still exists a considerable amount of
information that has not k e n presented. Acceleration data of the head, shoulder, and
lower abdomen region of the chestform was acquired in each test considered in this
investigation. An analysis of this data may provide steering wheel engineers and
Airbag development has become a very significant part in the design of steering
wheels. LS-DYNA also provides numerical algorithms for simulating airbags and the
deployment of such devices. Experimental testing, with the inclusion of airbags, has
also be conducted.
The need for finite element simulations has been proven in this investigation.
Finite element simulations provide information which may be very difficutt, costly, or
tests should be performed at different testing conditions in order to ver@ the FE models.
This investigation has proven that if detailed FE modefs are developed, they can
i4I- Morris, J. B., Stucki, L.. Morgan, R. M., Bondy, N., Occupant Protection from
Impacts with the Steering Assemblv, Society of Automotive Engineers. paper
no. 826025, National Highway Traffic Safety Administration. ( 1982). pp. 175-190.
Rodrigues, J. C. B., Pina da Silva, F., Tonelo, V., Steering Wheel and Dashboard -
Eroonomic and Material Safetv Trends, IMechE (Institution of Mechanical
Engineers) Seminar Publication on Automotive Passenger Safety, paper no.
C498/32/075/95, Institut0 Superior Tecnico, Portugal, (November 19951,
pp.47-52.
McKie, T., Simulation of Bodv Block Steering Wheel Impacts Usino DYNA3D,
First European LS-DYNA Con ference, The Motor Industry Research Association,
England, (March, 1997), pp.9.1-9.14.
[10]. Shyu, S., Mani, A., Krishnaswarny, P., Conroy, R., Shermetaro, M., Exner, G.,
Desiming Enerov Absorbina Steering Wheels Throueh Finite Element I m ~ a c t
Simulation, 1993 SAE Future Transportation Technology Conference, paper
no. 93 1844, EASi Engineering and United Technologies Automotive, (August
1993). pp. 1- 1O.
[1 11. Gotoh, E., Ohsawa, S., Satoh, Y., Yasuda, S., An Investigation
- of Steering
Column C o l l a ~ s eBehavior Using Finite Element Analvsis, SAE International
Congress & Exposition, paper no. 92039 1, Nissan Motor Company, Ltd.. and
Fujikikou Company, Ltd., (February, 1992), pp. 1-9.
[ 131. Omar, T. A.. Eskartdarian, A., Bedewi, N. E., Crash Analvsis of Two Vehicles in
Frontal Impact Usine Ada~tiveArtificial Neural Networks, Crashworthiness and
Occupant Protection in Transportation Systems - 1998 (annual meeting of
American Society of Mechanical Engineers), FHWANHTSA National Crash
Analysis Center, The George Washington University, VA (November 1W 8 ) ,
pp. 1 15- 129.
[ 131. Pries, H., Sinnhuber, R., Zobel, R.,The C h a n e i n ~Rote of Crash Simulation for
Crashworthiness Desim of Passenger Cars, iMechE (Institution of Mechanical
Engineers) Serninar Publication on Automotive Passenger Safety, paper no.
C498/32/ l87/95, Voikswagen AG, Gerrnany, (November 1995). pp.7 1 -78.
(141. Bennett. J. A., Lust, R-V., Wang, J.T., Ovtimal Design Strate~iesin
Crashworthiness and O c c u ~ a nProtection,
t Crashworthiness and Occupant
Protection in Transportation Systems - 199 1 (annual meeting of American Society
of Mechanical Engineers), Department of Engineering Mechanics. General
.Motors Research Laboratories, Warren Michigan (December 199 1 ), pp.5 1-66.
[15]. Saha, N. K., Mahadevan, S. K., Midoun, D. E., Yang, J. S., Finite Element
Structure-Dummv Svstem Mode1 for Side Irn~actSimulation, Crashworthiness
and Occupant Protection in Transportation Systems - 199 1 (annual meeting of
American Society of Mechanical Engineers), Vehicie Systems CAE, P&MS Car
Product Equipment, Ford Motor Company, Dearborn Michigan (December 199 1 ),
pp. 1 69- 179.
[16]. Scott Lui, D. S., Vehicle Rear Barries Impact Simulation bv using Nonlinear
Finite Element Av~roach,Crashworthiness and Occupant Protection in
Transportation Systems - 1991 (annual meeting of Arnerican Society of
Mechanical Engineers), CPC Engineering General Motors Corporation, Warren
Michigan, (December 199 1 ), pp.207-2 15.
[ 1 71. BeIytschko, T., On Com~utationalMethods for Crashworthiness, Proceedings of
the Seventh International Con ference on Vehicle Structural Mechanics (SAE
conference), p a p a no. 880893, Depts. of Civil and Mechanical Engineering
Northwestern University, (April l988), pp.83-92.
[18]. Weinhold. V., Baccouche, M. R., Hasan, A., Crash Enernv Management
- of
Alurninum Comuonents, Crashworthiness and Occupant Protection in
Transponation Systems - 1998 (annual meeting of American Society of
Mechanical Engineers), Ford Motor Company, Alternative Fuel Vehicle Centre
(AFV), (November 1998), pp. 17-29.
[19]. Rowberry, P. J., Imuact Management and Passenger Safetv, TMechE (Institution
of iblechanicai Engineers) Seminar Publication on Automotive Passenger Safety,
paper no. C498/l8/CW/W, Warwick Manufacturing Group, United Kingdom.
(November 1 9 9 3 , pp.23-30.
[XI. Chou, C. C., Wu, F., GU, L., Wu, S. R., A Review of Mathematical Models for
Rollover Simulations, Crashworthiness and Occupant Protection in Transportation
Systems - 1998 (annual meeting of American Society of Mechanical Engineers),
Safety and Biomechanics CAE Department. Ford Motor Company, (November
1998), pp.223-239.
[23]. Bathe, K., Cmsh Simulation of Cars with FEA, Mechanical Engineering,
(November 1998), pp.82-83.
[24]. Neathery, R., Analvsis of Chest Impact Response Data and Scaled Performance
Recommendations, Hybrid III: The First Human-Like Crash Test Dummy (SAE
Publication), paper no. 74 1 188, General Motors Research Laboratories ( 1994),
pp.255-27 1.
[ 2 5 ] . Bacon, D., The Effect of Restraint Design and Seat Position on the Crash
, Hybrid m:The First Human-Like Crash Test
Dummy (SAE Publication), paper no. 896052, Motor Industry Research
Association, Crash Protection Centre (1994), pp.64 1-647.
[XI. Wismans. J., Herrnans, J., Madymo 3D Simulations of Hvbrid iIi D u m m ~Sled
Tests. Hybrid m:The First Human-Like Crash Test Dummy (SAE Publication).
paper no. 880645, TNO Road-Vehicles Research Institute (1994). pp.735-744.
Obergefell, L., Kaleps, I., Steele, S., Part 572 and Hvbrid Ui Dummv Cornparison
Sled Test Simulations, Hybrid JII: The First Human-Like Crash Test Dummy
(SAE Publication), paper no. 880639, Armstrong Aerospace Medical Research
Laboratories and Systems Research Laboratories ( 1994). pp.76 1-776.
Le, J., Yang, K., Finite Element Modeling of Hvbrid Di Head-Neck Cornplex,
Hybrid DI: The First Human-Like Crash Test Dummy (SAE Publication), paper
no. 922526, Wayne State University ( l994), pp.777-79 1.
Barbat. S. D., Jeong, H., Prasad, P., Finite Element Modeline and Development of
the Deformable Featureless Headform and Its Apvlication to Vehicle Interior
Head Impact Testinq, an SAE special publication - Technologies for Occupant
Protection Assessment. paper no.960 104, Ford Motor Company and EASi
Engineering, ( 1W6), pp.6 1-69.
Kanno, Y., Masuda, M., Matsuoka, F., Evaluations of the Recommended New
Parts for the Hvbrid IiI D u m m ~- Neck Shield and Hio Joint, an S A E special
publication - Technologies for Occupant Protection Assessment, paper no.960450.
Toyota Motor Company, ( 1 W6), pp.77-87.
Wamer, C. Y., Wille, M. G., Brown, S. R., Nilsson, S., Mellander H., Kock, M..
A Load Sensing Face Form for Automotive CoIIision Crash Dummv
Instrumentation, Passenger Comfort, Convenience and Safety: Test Tools and
Procedures (SAE symposium), paper no. 860197, Collision Safety Engineering,
Mechanical Engineering Dept., Brigham Young University, and Volvo Car
Corporation, (February 1986), pp.85-94.
Ziarco, T., Column Anrrle se tu^ for the NS Steering Wheet, (personai
communication) Chrysler Technology Center (CTC), Chrysler Corporation. (May
1998).
Altenhof, W. J., Paonessa, S., Gaspar, R., Zamani, N., I m ~ a c tSimulation with
Exoerimental Verification of a Chestform For Use in Standardized Steering
Wheel Impact Testing Procedures, 5" International LS-DYNA User's Conference,
Southfield, Mich., (September 1W8), 10 pages.
Beiytschko. T., Non-Linear Finite Element Analvsis Short Course, Pa10 Alto,
California, U.S.A., (December 1997). 1 pape.
Kennedy. J., LS-DYNA Customized In-House Training Course, KBS2 Inc., Burr
Ridge, nlinois, U.S.A., (October 1997). 1 page.
An electronic mass scale and the steering wheel impact testing machine are the
Acceleration (Accelerometers):
Pelvis (Accelerometer # 1 ) - 0.0526 g's
Head (Acceler~meter#2) - 0.0548 gTs
Shoulder (Accelerometer #3) - 0.054 I gTs
APPENDK B
The data from experirnental tests is presented in tabular format. The first column
indicates the test number (which varies from 1 to 420). The remaining columns present
the measured and calculated observations from each test. There are 23 columns of
observations which cannot be presented in the width of a single page. Thus, multiple
repeatedly to properly present the experimental data with a corresponding test # for each
I 5.766 1
-
including page) (in) ( i n c l u d i n g m Accel. (pS) Acccl. (p's) Accel. ( g S )
5.266 1 -1.809 1 20.7768 -0.333
2 7.607 1 7.1071 -5.482 15.4034 -0.7577
3 8-2158 7.7 158 -5.2627 17.9809 -0.6495
4 7.8 129 7.3 129 -5.3 !75 18.6388 -0.7036
5 9-3617 8.86 17 - 1.8639 14.6917 -0.2 165
6 8.7259 8.2259 -0.4934 8.8808 -0.3247
7 8.99 16 8.49 16 -0.2 193 14.1984 0.2 165
8 9.1 145 8.6 145 -0.0548 14.6917 -0.1O82
9 8.7759 8.2759 -0.1096 9.5935 -0.054 1
1O 8.4387 7.9387 -0.274 1 10.963 0.054 1
II 6.22 12 5.7212 -6.2495 14.5273 - 1.2989
12 7.0942 6.5942 -5.756 1 15.5689 -1.5154
13 7.82 7.32 -3.3988 13.047 1 - 1.353
14 8.340 1 7.840 1 -4.1663 14.2532 -0.7036
15 8.2558 7.7558 -2.9603 1 1.9507 -0.7036
16 8.4 101 7.9101 -3.6181 1 3.8694 -0.6495
17 8.7287 8.2287 -3.1247 14.2532 -0.7036
1S 8.520 1 8.020 1 -1.0416 1 3.7598 -0.2165
19 8.5101 8.0101 -0.7675 1 3.5953 -0.2165
20 8.35 15 7.85 15 -0.2 193 1 1.1284 0
21 6.59 13 6.09 13 -7.3459 16.0622 - 1 .84O 1
--
33 6.374 1 5.874 1 -7.4007 16.3363 - 1.353
23 7.1528 6.6528 4.8242 12.2248 -0.3247
24 7.33 14 6.83 14 - 1 -206 9.0453 -0.487 1
25 6.9299 6.4299 -1.1512 8.66 16 -0.2165
26 7.1242 6.6232 -0.3837 1 1.3477 -0.5953
27 6.7942 6.2942 -0.6578 8.66 16 -0.433
25 6.7456 6.2356 -0.4934 9.8 128 -0.433
29 6.5298 6.0298 -0.6578 1 1.7315 -0.5953
3O 6-464 1 5.964 1 -0.5482 1 1.238 1 -0.54 1 2
31 5.8623 5.3623 -7.7844 14.6369 -0.6495
32 5.7263 5.2263 -9.6483 17.8165 -0.8659
33 5.8087 5.3087 -6.8525 17.7617 -0.333
34 5.7 125 5.2125 -0.603 1 3.376 1 -0.7577
35 5.7592 5.2592 -3.6181 14.6369 -0.920 1
36 5.8857 5.3857 -0.7675 15.9526 -0.7036
37 5.6355 5.1355 -0.93 I9 1 7.652 -0.54 1 2
38 5.6136 5.1 136 - 1-0416 13.6502 -0.7577
39 5.6603 5.1603 - 1 -2609 16.1719 -0.920 1
40 5.3909 4.8909 - 1 -3705 2 1 S442 -0.920 1
41 5.406 4.906 -4.879 16.0622 -0.9732
42 5.4307 4.9307 -6.4 139 16.5008 -0.5953
43 5.55 17 5.05 1 7 - 1 -0964 14.8014 -0.7036
44 5.4527 4.9527 -4.276 17.8713 -0.9742
45 5.7029 5.2029 -1.1512 14.9658 -0.9742
46 5.5599 5.0599 - 1.4253 16.1 171 -1.1907
47 5.3868 4.8868 - 1.4253 14.7466 -1.1907
48 5.4582 4.9582 -0.7 127 15.6785 -0.4871
49 5.3249 4.8249 -0.7 1 27 18.8032 -0.3788
50 5.4775 4.9775 -3.7826 18.9677 -2.0566
Tcst ?cak Displacernent (withou Peak Displacement Min Head Max Head Min Shouldei
including sape) (in) = ( i n c l u d $ a g e ) (in)
_. Accel. (g's) Accel. (g's)
51 4.9304 3.4304 - 1.5898 17.9809 -0.5953
52 -5.0983 15.6237 -0.5953
53 - 1 .O964 15.1851 -0.3788
54 -0.877 1 13.8694 -0.3247
55 - 1.6994 15.8978 -0.3788
56 -1.1512 1 3-7598 -0.3247
57 -0.6578 13.8146 -0.3247
58 -1.1512 1 3.5953 -0.3247
59 -2.52 17 1 5.5689 -0.9742
60 -2.63 14 13.8694 -0.7577
61 - 1.535 1 7.652 -0.8659
62 -0.6578 16.1 171 -0.487 1
63 -0.603 14.3628 -0.1624
64 -2.9603 14.91 1 -0.433
65 -0.5482 14.4725 -0.6495
66 -0.877 I 1 2.9923 -0.3788
67 -0.4386 15.2948 -0.54 1 2
68 -2.63 14 13.705 -0.8 1 1 8
69 -0.9868 12.6634 -0.7036
70 -2.9055 13.2664 -1.5154
71 -6.359 1 2 1 -0509 -0.5953
72 -5.8657 18.2002 -0.433
73 -5.5368 15.2399 -0.9742
71 -5.9754 1 5-4592 -0.8659
75 -0.4386 10.7995 -0.3247
76 -3.5085 16.2815 -0.6495
77 -0.2 193 7.894 1 O
78 -0.274 1 8.1682 -0.2165
79 -0.6578 9.758 -0.3788
80 -0.1645 13.21 16 O. 1 O82
81 -6.4 139 20.3382 - I .SUS
82 -6.5784 15.3496 - 1.6236
83 -4.4404 1 7.926 1 - 1 -786
84 -6.4688 1 8.0906 -0.920 1
85 -2.9055 12.8827 - 1.2448
86 -3.6729 13-4857 -0.5953
87 -0.8223 9.6483 -0.1624
88 -0.6578 8.66 16 O
89 -0.603 8.0037 -0.7036
90 -0.3837 8.4123 -0.054 1
91 -6.0302 13.102 -1.7319
92 -5.8657 15.4044 -1.7319
93 -6.5236 15.4592 - 1.6778
94 - 1.6994 10.25 1 3 -0.54 12
95 -1.535 9.9772 -0.54 1 2
96 -0.3289 9.8 128 -0.6495
97 -0.8223 1 1.5 122 -0.8659
98 -0.3837 1O. 1965 -0.8 1 18
99 -0.6578 15.843 -1.1907
1O0 -0.4386 14.4176 -0.487 1
Pcak Displacement (withot Peak Displacement Min Head Max Head blin Shoulde
including gage) (in) (including g a p ) (in) &cel.(& Acccl. (p's Accel. (g's)
6-3512 5-8512 -1 -3157 16.446 -0.9742
6.1841 5.684 1 -8.66 I6 20.94 12 -0.7577
6.7599 6.2599 -4.8242 17.5972 -0.920 1
6.3498 5.8498 -0.7675 15.6785 -0.7577
6.3055 5.8055 - 1.7542 15.3496 -0.54 12
6.2898 5.7898 -0.877 1 1 5.3496 -0.6495
6.0055 5.5055 -0.7 1 27 19.5159 -0.7577
6.0469 5.5469 - 1.809 1 14.5273 - 1.2448
6.044 1 5.544 1 - 1 -2609 20.1737 - 1.0824
6.1412 5.64 12 - 1.5898 19.5707 - 1.2989
5.9805 5.4805 -6. I398 19-I87 -1.1365
5.9572 5.4572 -1.0416 16.6653 -0.8659
6.0898 5,5898 -1.3 157 15.4592 - 1 -0283
6.1698 5.6698 -4.0567 16-2815 -0.9742
5.9297 5.4297 -9.4839 21.2701 -0.920 1
6.1241 5-6241 - 1 -6994 16.9942 - 1.0824
5.898 1 5.398 1 -I .9735 16.1719 - 1.2448
5.8802 5.3802 -1.9187 20.1 189 - 1.5695
5.821 1 5,321 1 - 1.809 1 17.9261 -1.1907
5.8238 5.3238 -3.3988 18.4733 -2.1 107
5.41 15 4.91 15 -16.1 171 32.4534 -2.1648
5.975 5.475 -0.877 1 17.1586 -0.54 12
5.9022 5.4022 -7.62 18.584 -0.8 1 I8
6.1369 5.6369 -0.9868 14.91 1 -0.3788
6.1041 5.604 1 -1.3157 1 8.6388 -0.5953
6.1612 5.66 12 -0.9868 14.2532 -0.487 1
6. I969 5.6969 -0.93 19 14.582 1 -0.3247
5.9902 5.4902 - 1 -3705 14.4725 -0.3758
6.1541 5.654 1 - 1.0964 1 5 -0207 -0.5953
6.0598 5.5598 -2.63 14 15.1303 -1.8401
5.7854 5.2854 -1.8091 19.7352 -0.9742
5.8 I83 5.3 I83 - 1 -4253 17.3779 -0.54 1 2
6.277 5.777 -0.6578 16.1 171 -0.487 I
6.1955 5.6955 -0.877 1 15.8978 -0.5953
6.5084 6.0084 -0.93 19 15.7333 -0.3788
6.4 127 5.9 127 -0.93 19 14.0887 -0.487 1
6.327 5.827 -0.9868 13.705 -0.433
6.2084 5.7084 -1.1512 16.6653 -0.3247
6.2255 5.7255 -1.0416 14.308 -0.487 1
6.1727 5.6727 - 1.6994 13.5953 -0.9742
6.7999 6.2999 -7.18 14 27.9034 -0.8 1 18
8.1672 7.6672 -4.6049 23.79 19 -0.6495
8.733 8.233 -5.59 16 15.0207 -0.7036
8.6959 8.1959 -6.085 16.9394 -0.920 1
8.91 16 8.41 16 - I .206 20.5575 -0.6495
10.4975 9.9975 -2.0832 16.1719 -0.6495
10.6018 10.1018 -0.6578 13.8694 -0.2706
10.2818 9.78 18 -0.3289 8.8808 -0.2706
10.2332 9.7332 - 1.3705 15.0755 -0.433
10.5333 10.0333 -0.1096 8.66 16 0.3247
Test 4 Pcak Disptaccment (withou Peak Displacement Min Head Mm Head din Shouldc
including gage) (in) (including gage) (in) Accel. (p's) Acccl. (g's) Acccl. (pSs)
151 7.1014 6.60 14 -7.894 1 26.0395 - 1.6236
152 8.2072 -6.9073 16.1719 -1.4613
153 8.2329 -2.0283 18.3647 - 1 -5695
1 54 8.3029 -6.6332 16.2267 -2.0566
155 8.963 1 -4.221 1 16.1719 -2.4896
I56 9.766 -2.2476 1 3.6502 - 1.4072
157 9.9 146 - 1.5898 1 2.8279 - 1.4072
158 9.816 -0.7675 9.4839 -0.433
159 9.143 1 -0.3289 8.8808 -0.7577
160 9.856 -0.6578 8.497 1 -0.433
161 7.5528 - 10.7995 18.7484 -2.3272
162 7.3657 -7.0 17 18.6388 -2.38 1 3
I63 8.1058 -7.8393 19.0773 -1.4613
164 8.5644 -4.7 145 15.514 - 1 -6778
165 8.36 15 -3.344 12.9375 - 1 -2989
1 66 5.3401 -0.4934 9.9772 -0.7577
167 7.99 -0.7675 9.9772 -0.54 I2
168 8. I286 -0.6578 12.2248 -0.920 1
1 69 7.9072 -0.603 15.6237 -1.7319
170 7.6786 -0.2 I93 15.9526 -0.5953
171 6.707 - 1.809 I t 8.0358 -2.4896
172 6.8327 -9.9773, 22.2569 -0.433
173 7.1 1 14 -0.877 1 16.1 171 -0.8659
173 6.7756 -0.6578 16.1 I71 -0,6495
175 6.6084 -0.7675 16.6653 -0.5953
I76 6.7656 -0.9868 16.0622 -0.3788
177 6.5498 - I -0964 14.3176 -0.54 1 2
178 6.267 - 1.8639 22.8599 - 1.2989
179 6.347 - 1.6994 20.72 19 - 1.5695
I 80 6.45 13 - 1 -7542 1 9.2966 - 1.5695
I81 6.224 1 -6.7429 19.6255 -0.9742
182 6.2855 - 1 -2609 17.652 -0.9742
183 6.2984 - 1.535 15.9526 - t .0283
184 6.2 155 - 1.6446 18.255 - I .0824
I85 6.393 1 - 1 -6994 16.9942 -1.1365
186 6.3984 -1.9187 17.4327 -0.920 I
I87 6.1883 - 1.206 17.5972 -0.9742
I88 6.134 1 -0.9868 15.584 -0.7036
1 89 6.06 12 - I ,8091 19.461 1 - 1 -0824
190 6.0 I 69 -2.3573 17.8713 -1.1365
191 5.6493 - I S898 19.8448 -0.541 2
192 6.2984 -7.4007 20.064 1 -0.7036
193 6.4398 -0.877 1 16.9394 -0.3788
194 6.44 13 -4.550 1 19.187 -0.54 12
195 6.454 1 -2.74 1 19.7352 -0.3247
196 6.487 - 1 -0964 15.5689 -0.5953
197 6.5227 -1.3157 15.9526 -0.3788
198 6.2998 - 1.0964 15.2399 -0.6495
199 6.457 -0.9868 t 5.6237 -0.487 1
200 6.424 1 -2.6862 16.1 171 - 1 -786
Test 'cak Displacement (withou Peak Displaccrnenr Min Head Max Head Min Shouldc
includinegaee) (in) (includingage) (in) A C C(e3:~ s Acccl. (g's)
20 1 6.0655 5.5655 -6.6332 2 1.2701 -0.9742
202 6.0483 5.5483 - 1-3705 20.2286 -0.54 1 2
203 6.494 1 5.994 1 -0.9319 17.7068 -0.433
203 6.4698 5.9698 -0.93 1 9 16.8845 -0.7036
205 6-5013 6.00 13 -6.6332 19.79 -0.5953
206 6.499 8 5.9998 -2.6862 16.9394 -0.7036
207 6.7584 6.2584 -0.9868 14.5273 -0.3788
208 6.6 156 6.1 156 -0.8771 14.1435 -0.1624
209 6.5784 6.0784 -0.7675 15.2399 -0.6495
210 6.434 1 5-9341 - 1 -6994 16.5556 -0.387 1
21 I 7.2957 6.7957 -7.9489 36.4553 - 1.2989
212 9.003 8.503 -6.3043 24.7238 -0.3237
2 13 9.60 17 9.1017 -6.2495 20.5027 -0.920 1
2 14 9.2559 8.7559 -6.7977 19-79 - 1 .O283
215 1 1.2076 lO.7076 -3.2892 19.1322 -1.1907
2 16 10.4204 9.9204 -3.5633 18.9677 -0.8 1 18
217 1 1.2391 10.739 1 - 1.9735 15.7881 -0.487 i
218 1 1 -0247 10.5247 - 1.5898 14.308 -0.6495
219 1 1.0076 10.5076 -2.1928 16.0074 -0.54 12
220 11.1891 10.6891 -1.0416 12.773 -0.1O82
22 1 8.8402 8.3402 -8.0585 19.0225 -0.920 1
222 8.5258 8.0258 -6.5784 20.5027 -0.54 12
223 9.5903 9.0903 -5.4272 15.6237 -0.7577
224 8.873 8.373 -6.9073 17.3231 - 1.353
235 9.6746 9.1746 -4.8242 19-2418 -2.1648
226 9.493 1 8.993 1 - 1.6446 19.187 -1 -2989
227 10.6547 10.1547 -4.0019 15.7333 -3.3555
228 1 O. 1975 9.6975 - 1 -4801 16.1719 - 1 .2W8
229 10.979 10.479 -0.6578 12.8827 -0.433
230 10.7547 10.2547 -0.7675 10.4706 -0.487 1
23 1 7.6 157 7.1 157 -9.429 27.1907 -2.706 1
232 7.9458 7.4458 -6.7977 2 1 -0509 -2.4355
233 8.493 7.993 - I -0964 1 3.5405 -2.38 13
234 8.7873 8.2873 -7.7296 20.393 -2.7602
235 8.8745 8.3745 -5.0983 16.6104 -2.7602
236 9.323 1 8.823 1 -0.8223 9.8676 -0.6495
237 8.9745 8.4745 -0.9868 9.8 128 -0.7036
238 8.8945 8.3945 -0.7675 10.3062 - 1.9484
239 8.5573 8.0573 -0.3837 15.843 -1.1365
240 8.2 144 7.7 144 -0.274 1 12.3345 -1.1907
241 7.37 14 6.87 14 - 1 -7542 20.0093 - 1.8942
232 6.9556 6.4556 - 1.5898 2 1.599 1 -2.0025
24 3 7.55 7.05 -0.93 19 18.3099 - 1 -2989
244 7.37 14 6.87 14 -1.3157 17.049 -0.6495
245 6.9756 6.4756 - 1.7542 17.926 1 -0.7577
246 7.49 14 6.99 14 -0.7 127 17.1038 -0.1082
247 7.2799 6.7799 -0.9868 16.8845 -0.54 12
248 7.2985 6.7985 -1.1512 24.01 1 l -0.5953
249 7.0842 6.5842 - 1.5898 20.0093 -0.920 1
250 6.897 6.397 - 1 -8639 20.7768 - 1.2989
Test 4 'cak Displacernent (withou Peak Displacernent Min Hcad Max Hcad Min Shouldei
sage)(in) ( i n c t w n ) Acccl.(gS) Accel. (p's) Acccl. (g's)
25 I 6.6484 6.1383 - 1.4801 19.6255 -0.8 1 18
252 6.5027 6.0027 4.8242 23.5726 -0.9742
253 6.8099 6.3099 -7.6748 18.8581 -1.1907
254 6.8885 6.3885 4.3308 19.5707 -0.920 1
255 6.6842 6.1842 -2.138 19.4063 -1.1365
256 6.7799 6.2799 - 1.8639 19.6804 - 1 -2989
257 6.427 5.927 - 1 -4801 20.4478 - 1 .O283
258 6.6556 6.1556 - 1.6446 17.4327 - 1 -4072
259 6.521 3 6.02 1 3 - 1 -6994 20.1 189 -1.5154
260 6.594 1 6.091 1 -2.4669 22.53 1 -0.920 1
26 1 6.0 126 5.5 126 - 1.535 20.72 19 -0.3247
262 6.5084 6.0084 -9.1549 25.2 172 -0.8 1 18
263 6.7784 6.2784 - 1.2609 18.9 129 -0.3247
264 6.3427 5.8427 - 10.4706 26.5329 -0.920 1
265 6.8599 6.3599 - 1.206 16.6104 -0.5953
266 6.977 6.477 - 1.7542 16.1719 -0.5953
267 6.7956 6.2956 -0.7675 20.94 1 2 -0.54 12
268 6.7099 6.2099 - 1 -3157 2 1.4346 -0.6495
269 6.957 6.457 - 1.2609 16.5556 - 1 -0824
270 6.6599 6.1599 -2.7958 19.1322 - 1 -6236
27 1 6.537 6.037 - 1 -8639 21.8183 -0.920 1
272 6.6 184 6.1 184 - 1 -5898 21.1057 -0.920 1
273 6.3427 5.8427 - 10.4706 26.5329 -0.920 1
273 6.8242 6.3242 -1.0416 19.2966 -0.7036
275 7.07 13 6.57 1 3 -2.0283 l8.2oO2 -0.3788
276 7.0785 6.5785 -2.4 121 17.8165 -0.6495
277 7.2528 6.7528 -1.0416 15.8978 -0.5953
278 7.3257 6.8257 -0.8223 1 6.7749 -0.5953
279 7.0 156 6.5 156 - 1 -4253 15.8978 -0.433
280 7.2057 6.7057 - 1.2609 14.2532 -0.54 12
28 1 9.1273 8.6273 -3.O699 30.8088 -0.9732
282 9.4 16 8.9 16 -6.6332 25.6009 -0.8 1 18
283 9.2402 8.7402 -4.1663 32.3986 - 1 .O283
284 1O. 2389 9.7389 -6.3043 19.0773 -1.1907
285 1 1.7434 1 1.2434 -3.8922 24.8334 -1.1907
286 11.1648 10.6648 -4.9338 23 .2985 - I .O283
287 1 1.802 1 1.302 -3.8922 2 1 -8732 - 1 -0283
288 1 1.5234 1 1 .O234 - 1 -9735 17.4327 -0.7577
289 1 1.3948 10.8948 -4.3308 19.79 -1.1907
290 1 1.7206 1 1.2206 -4.7 145 28.1226 -1.1907
29 1 8.2244 7.7244 -9.8676 33.7 143 -0.8659
292 9.37 17 8.87 17 -7.4007 22.53 1 -0.7577
293 10.0303 9.5303 -5.8 1O9 17.323 1 - 1.0283
294 10.1 189 9.6 1 89 -5.6465 20.5027 -0.5953
295 IO. 1946 9.6946 -4.9886 2 1.325 - 1 -6236
296 10.6004 10.1004 -2.138 2 1 -928 -1.8401
297 11.1448 10.6448 -4.6597 18.7484 -4.1 132
298 10.8933 10.3933 -3.4537 17.9261 -3.7885
299 1 1.6048 11.1048 -2,138 16.1 171 -1.1907
300 10.7076 10.2076 -6.3043 32.837 1 -2.0566
Test Peak Displacement (witi Peak Displacement Min Head Max Head Min Shoulde
- Accel. Cg's) Accel. (pS)
30 1 8.1901 7.6901 -5.756 1 26.9 166 -2.65 19
303 7.943 -5.9754 24.3949 -2.3272
303 9.0 173 8.5 173 -5.70 13 15.514 -2.273 1
304 9.6788 9.1788 -6.0302 17.7068 -2.273 1
305 9.2645 8.7645 -5.59 16 17.8165 -2.5437
306 9.2802 8.7802 - 1.8639 14.308 - 1 .5695
307 9.76 17 9.26 17 -1.0416 1 1.567 - 1.2989
308 9.71 17 9.21 17 -0.5482 16.3363 -0.7577
309 9.4 145 8.9 145 -0.4386 12.7 182 -1.5154
310 9.1816 8.68 16 -0.93 19 17-8713 -1.4613
31 1 7.8343 7.3343 - 1 -7542 22.6406 -2.1 107
312 7.9529 7.4529 -9.758 23 -4629 - 1.5695
313 8.3 144 7.8 144 -1.535 19.2966 -1.1907
313 8.0986 7.5986 -9.9224 23.9563 -0.7577
315 7.9486 7.4486 -0.93 19 19.8448 -0.7036
3 16 8.0643 7.5643 -1.0416 19.1322 -0.3788
317 7.8572 7.3572 -0.8223 18.4743 -0.54 12
318 7.7386 7.2386 -0.8223 17.87 13 -0.5953
3 I9 7.69 7.19 - 1.0964 20.393 -0.595 3
320 7.5643 7.0643 -0.4934 20.064 1 -1.1907
32 1 7 .O428 6.5428 - 1.7542 2 1.7635 -3.5 179
322 7.107 1 6.607 1 -2.3024 2 1 -6539 -1.5154
323 7.1 171 6.6171 -3.1247 20.996 - 1.2989
324 7.2842 6.7842 -6.688 19.5707 - 1.0283
325 7.3328 6.8328 -1.3157 18.9677 - 1 -0824
326 7.3 142 6.8 142 -1 -8091 20.83 16 - 1.6236
327 7.2685 6.7685 -0.5482 26.149 1 -0.9742
328 7.0542 6.5542 - 1.5898 2 1.6539 - 1.4072
329 7-0513 6.55 13 - 1.535 20.6 123 - 1.353
330 7.05 13 6.55 13 - I .8639 17.4876 -2.0025
33 1 6.7056 6.2056 -3 S633 24.8883 -0.9742
332 7.0328 6.5328 - 1.O964 2 1.4894 -0.81 18
333 6.9399 6.4399 - 1 -4253 22.0924 -0.7577
3 34 7.2 157 6.7 157 - 1.535 18.584 -0.433
3 35 7.46 14 6.96 14 - 1.535 18.3099 -1.1907
3 36 7.427 1 6.927 1 -2.2476 1 7.8 165 - 1 .O824
337 7.3328 6.8328 - 1 -6446 24.3949 -0.7577
338 7.2214 6.72 14 -2.138 17.049 -0.920 1
339 7.33 6.83 - 1.6446 18.2002 - 1 S695
340 7.2628 6.7628 -3.3988 17.2 135 - 1.5695
34 1 6.7256 6.2256 - 10.8543 25.7654 - 1 .O283
342 7.0242 6.5242 -8.3875 25.2 172 -0.8 1 18
343 7.297 1 6.797 1 - 1.2609 21.2153 -0.7036
344 7.2342 6.7342 -4.7693 2 1.4894 -0.7036
345 7.1871 6.687 1 -2.0283 2 1.7087 -0.5953
346 7.3028 6.8028 - 1.809 1 19.3514 -0.8659
347 7.68 14 7.1814 -1.3157 22.0924 -0.8659
348 7.587 1 7.087 1 - 1.3705 20.393 -0.920 1
349 7 -6028 7.1028 -1.480 1 16.1 171 -0.7036
350 7.7272 7.2272 -1 -535 18.584 -0.5953
Peak Displacement (withou Peak Displacement Min Head Max Head Min Shouldci
includinp gage)
- (in) (includingge) (in) Accei. (e's: Acccl. (g's: Acccl. (g's)
8.1486 7.6486 -10.361 42.3758 -2.706 1
9.6989 9.1989 -6.962 1 29.3287 - 1 -0283
10.3947 9.8947 -6.5784 24.7238 - 1 .O283
10.0289 9.5289 -7.62 26.4232 -0.920 1
1 1.9763 1 1.4763 -3.8374 27.1907 - 1.5695
1 1 .go63 1 1.4063 -4.221 1 28.1775 - 1.8932
12.1306 1 1.6306 -4.6597 30.6444 - 1.5695
1 1.932 1 1.432 -3.6181 25.6009 -2.0025
1 1.9863 1 1.4863 4.8242 22.42 14 - 1.353
12.1792 1 1.6792 -2.8506 22.7503 -0.7577
8.463 7.963 -9.9772 36.7294 - 1.353
9.4588 8.9588 -8.2778 28.3967 -1.1907
10.3289 9.8289 - 1.8639 21.8183 -2.1 1 O7
10.5218 10.0218 -6.688 19.2418 -1.3613
10.9576 10.4576 -5.70 13 19.79 -1.1907
I 1.3476 10.8476 -4.4952 21.Y01 -3.6261
1 1 S062 1 1 .O062 -4.879 22.7503 -3.139
1 1 .5748 1 1.0748 -3.4537 23.737 -2.4355
1 1.759 1 1 1.î59 1 4 - 1 1 I5 23.6274 -2.0566
12.0035 1 1S035 -3.4537 29.548 - I .2448
9.243 1 8.743 1 -6,7977 24.5593 -2.4355
9.3002 8.8002 -1 1.7315 26.86 18 -4.22 I5
9.596 9.096 -4.1663 1 5.6237 - 1.9484
10.0603 9.5603 -5.756 1 17.2683 -2.3272
9.9403 9.4403 -7.7296 20.064 1 -2.1 1O7
9.8274 9.3274 4.550 I 16.5556 - 1.786
10.6 175 10.1 175 -2.138 17-7617 -1.1907
10.2832 9.7832 -0.8223 14.8014 - 1.0283
10.6547 1O. 1547 -0.7675 1 7.652 -0.5312
11.149 10.649 -0.877 1 18.1354 - 1.0824
8.4873 7.9873 - 12.2248 33.5498 -2.4896
8.2 186 7.7 186 - 1.9735 26.149 1 -2.1648
8.2872 7.7872 -8.2778 23.408 1 - 1.6778
8.580 1 8.080 1 -8.223 2 1.4346 - 1 .%84
8.6 144 8.1 144 -1.3157 21.2153 -0.9742
8.5758 8.0758 - I .6U6 21 .I6O5 -0.2 165
8.5 158 8.0 158 -0.9868 19.35 14 -0.7036
8.2045 7.704 -0.8223 20.6 123 -0.7577
8.3644 7.8644 -1.1512 20.0093 -0.7577
8.1929 7.6929 -0.4934 2 1.6539 - 1.353
7.5928 7.0928 -2.3024 23.6274 -2.3272
7.66 7.16 -8.7 164 25.272 -1.4613
7.66 14 7.1614 - 1.8639 20.83 16 -1.4613
7-9057 7,4057 -6.7977 24.669 - 1 .O283
7.7414 7.24 14 -2.63 14 20.393 - 1 .O823
7.6486 7.1486 -8.223 26.423 2 -2.2 19
7.4843 6.9843 -2.63 14 21.2153 - 1.2989
7.477 I 6.977 1 -2.4669 2 1.8732 - 1.840 1
7.6486 7.1486 - 1 -5898 20.72 19 - 1 -353
7.4928 6.9928 -1.3157 18.2002 - 1.6236
Test # I ~ e î kDisplacernent (wiihoutl Peak Displacement Min Head Max Head Min Shouldcr
including gage) (in) - Accel. (g's) Accel. (g's)
(includinggage) (in)
40 1 6.85 13 6.35 13 -13.0471 29.822 1 - 1 -6236
402 7.2928 6.7928 -9.8 128 28.9449 - 1.5695
403 7.6229 7.1229 -7.07 18 22.8599 -0.487 1
404 7.6171 7.1 171 -1.4801 20.5575 -0.54 12
405 7.6886 7.1886 -6.4 1 39 22.202 1 -0.5953
406 7.5757 7.0757 -0.8223 18.7484 -0.8659
407 7.9029 7.4029 - I .6994 1 7.9809 -2.0566
408 7.79 14 7.29 14 - 1.6994 20.3382 - 1 -9484
409 7.8257 7.3257 - 1 -535 19.5159 - 1 -6236
410 7.6 1 7.1 1 -3.344 2 1 .O509 -1.7319
41 l 7.1814 6.68 14 -2.4669 25 .%6 1 -0.920 1
412 7.3285 6.8285 - 1 2.0056 33.495 -1.1907
413 7.5543 7.0543 -7.0 17 26.2588 -0.8 1 18
-114 7.62 14 7.1214 -7.29 1 1 26.0395 - 1.4072
415 7.7672 7.2672 -4.1 115 23 S726 - 1.2989
4 16 7.7886 7.2886 - 1.9735 2 1 5442 - 1.5695
317 7.9886 7.4886 - 1 -3705 27.0262 -0.920 1
3 18 8.1033 7.6043 -1.8091 25.4365 -0.7036
419 7.7886 7.2886 -2.2476 26.5329 - 1.2989
320 8 7.5 -3.8374 16.7749 -1.4613
Test # Max Shoulder Min Pelvis Max Pelvis Total Energy Peak Absolute
Acccl. (g's) Accel. (p's) Accel. (p's) Absorption (Ibf*in) Impact Load (IbO
1 16.561 1 - 1.83% 23.01 87 265 1.37034 1965.4235
2 13.0715 -4-6248 19.8655 1562.6294
3 1 3.3762 -4.7299 20.39 1 1525.I948
4 13.6385 -4.6773 20.969 1 1586.4 1 39
5 8.8759 -2.3649 10.616 83 1.798
6 8.93 -0.8934 9.4598 1027.342
7 8.8759 - 1.2087 9.7225 752.2827
8 8.7676 - 1.6292 9.9327 768.7049
9 7.0358 -0.2628 8.0408 828.55 19
10 8.1 182 -0.0526 8.25 1 803.4526
11 12.989 1 -3.73 13 18.394 1612.1286
12 1 1 -5278 4.6773 19.0246 lJ58.2808
13 12.3937 -3.8365 1 5.556 1 3 10.3677
I4 1 IS819 -3.7839 1 1.2992 1 199.004
I5 1 1 -0407 -3.3635 12.14 I 1 I7.6719
I6 10.9325 -3.73 1 3 12.1926 1093.4322
17 9.9042 -3.7839 11.0364 883-7005
IS 9.8501 -2.6803 I 1 -7721 798.6226
19 10.1207 -2.47 1 1 so94 8 1 1.7392
20 7.3605 -0.7883 9.5 123 802.2404
21 12.7726 -3.6262 17.1852 1346.48 1
22 16.8858 -6.41 16 23.1764 1423.2071
23 10.9866 -2.1022 13.61 15 1079.9937
24 9.9042 -1.2613 11.194 936.0938
25 9.5794 -1.7343 9.67 884. I99
26 1 1 -6361 -1.419 1 2.1926 912.4417
27 9.9583 -1.5241 11.194 936.5309
28 10.0 1 24 - 1 -7868 12.1926 950.6202
29 10.8242 - 1 -83% 13.2437 1051 395
30 9.6336 - 1.7868 13.191 1 1005.4855
31 1 5.9657 - 1.9445 17.0801 1408.9879
32 15.7493 -3.1007 1 7.2903 1397.81 1 1
33 12.989 1 - 1.9971 13.61 15 1274.629
34 12.069 - 1.5241 1 3.4539 1 307.8846
35 1 2.0 149 - 1.892 14.3999 1 330.5278
36 13.5303 -1.4715 14.7152 1403.1995
37 14.5586 -0.8409 16.2918 14 17.8733
38 1 3-4221 - 1 -3664 14,4524 1467.3229
39 12.935 -2.3649 16.975 1666.7131
40 1 3-6927 - 1 -3664 16.5546 1565.7387
41 15.2622 - 1.6292 17.5006 1748.3209
42 13.9091 - 1 -5766 15.3458 1689.7278
43 13.8009 - 1.9445 14.505 1625.3098
44 14.5586 -2.5226 16.975 1512.4019
45 1 3.5303 - 1.7343 13.8218 1575.5854
46 1 5.4245 -2.0496 16.2918 1614.8643
47 14.3962 -1.419 13.8743 1600.9336
48 15.3704 -1.1562 15.5035 1639-4236
49 14.4503 -0.8934 15.556 1568.2346
50 17.3 188 -5.7284 16.2393 1738.2302
- 197-
Max Shoulder Min Pelvis Max Pelvis Total Energy Peak Absolute
Acccl. ( g S ) Accel. (g's) Accel. (g's) Absorption (ibf*in) Impact Load (Ibn
15.3704 -2.1022 19.1823 2260.75225 1880.0298
15-9116 - 1.6292 17.0801 1698.5222
15.0457 -1.1562 16.0816 1650.4 133
13-6927 -0.9985 14.5575 1381 -7487
I4.342 1 -1.1036 1 7.3429 1408.8237
14.4504 - 1.3664 15.6086 1481 2024
1 3.6927 -1.2613 13.6641 1504.0769
13.9633 -1.3715 1 3.4539 1393-550.1
13.288 -4.2569 15.6086 1 503-4225
13.3679 -4.5722 14.3999 1375.6966
16.3987 - 1.9445 18.867 1850.6476
14.S292 -1.1562 16.7648 l8W.3827
14.0715 -1.051 1 14.3999 1602.526
14.288 - 1.2087 15.2407 1528.458 1
13.6927 -1.051 1 14.5575 1512.1832
12.9891 -1.051 1 12.4553 1389.1258
12.6 102 - 1.3664 12.7707 1360.8235
1 2.0 149 -4.1518 1 3.0334 1351.9517
11.4196 -2.2073 15.7 137 1334.9265
12-2314 -5.3605 12.6656 1341.2155
17.481 1 -3.1007 27.0 129 19 17.5595
14.9374 -5.308 22.0202 1677.8805
14.7751 -4.835 20.4436 161 1 .O574
1 5.3704 -4.5722 2 1.4947 1675.4 153
9.5794 -1.051 1 10.3532 1084.9 104
1 1.7443 -5.9386 18.394 1254.2382
7.4 146 -0.6307 7.4101 828.3779
7.4 146 -0.6307 7.4101 860.047 1
8.93 -0.3679 8.7766 855.1 176
8.6594 -0.7883 9.197 930.634
13.2339 -6.254 22.2304 1564.8372
15.0457 4.2569 19-34 1571.8258
1 2.8267 -3.994 1 19.6553 1461 -5058
13.9633 4.4146 20.864 1539.1996
12.3937 -3.7839 16.029 1 239.2342
1 1.3654 -3.6262 12.4028 1072.7068
10.1748 -1.4715 I I .O364 1068.6873
9.47 12 - 1.5766 11.194 943.900 1
8.7676 - 1.6292 10.93 1 3 9 15.836
8.497 -0.7358 9.6 174 8 12.2427
14.6668 -4.362 19.9706 1449.01 17
17.6976 -4.8875 23.5968 1531 -5588
12.502 -4.3094 16.29 18 1213.1 151
1 1.9067 -2.6277 13.2962 1 O83-8906
1 1.7443 -1.3139 1 1.4043 950.3488
1 1.8525 -2.2073 14.1371 1019.1419
13.0432 -2.2073 1 3.7692 979.6797
1 1.4737 - 1.9445 13.1386 1005.2IOI
12.7 185 - 1.7343 16.4495 1040.0666
10.9325 -2.1547 14.9254 1 060.7084
- 198-
Test Max Shouide Min Pelvis Max Pelvis Total Energy Peak Absolute
- Accel. (p's) Accel. (g's) m s ) Absorption (Ibf*in) Impact Load (Ibf)
101 17.1564 -2.1547 19.3925 3243.36589 1502.354 1
102 23-0556 -5.7284 25.0684 3393.97942
103 15.1539 - 1-9445 14.2947 3474.28 143
104 14.3421 - 1.892 15.3984 3 163.52266
105 I 3.4762 -1 .524 1 14.8728 3223.3993 I
106 1 3.6927 - I .7868 13.7692 3 176.65976
107 1 6.2905 -1.7343 17.3954 3102.20612
I08 14.1256 -1.6817 14.2947 2879.98663
I09 14.0174 -2.2598 17.0801 28 15.49558
110 14.3421 - 1.6292 17.448 2994.093 12
111 17.5894 -2.8379 20.0232 3059.04652
112 16.6693 -1.6817 18.0261 2962.67698
113 I5.0998 - I .892 16.3969 2987.72 13
114 14.6668 -2.4 175 15.7663 2990.54 I 1 3
115 16.994 - 1.8394 16.4495 2878.89763
1 I6 15.641 -2.3 124 16.0816 2959.787 14
117 15.!I657 -2.2598 15.8714 2750.8362 1
118 17.1023 -2.5752 17.3429 2865.34892
119 16.8858 -2.1022 16.8174 27 19.45451
1 20 18.5636 -5.308 17.448 275 1-68046
121 24. I38 -7.7255 30.534 2902.0 1 109
122 16.074 - 1.2087 18.6567 2985.1037 I
123 15.641 - 1.8394 18.0787 3022.44953
1 24 14.5045 -1.1036 16.5546 304 1.3665
1 25 15.8575 - 1.5241 I6.2918 3043.367 15
I26 14.6127 -1.051 1 15.2933 2990.08 149
I27 14.721 -1.1036 14.7677 3014.23383
I28 15.0998 -1.419 14.7 152 3007.750 I 1
1 29 15.3I63 -1.3715 16.3444 3043-08531
130 14.6127 -5.4656 15.6086 2957.388 14
131 18.347 1 -2.3649 20.7589 2936.20994
132 16.561 I - I .4 19 19.6027 2874.289 1
133 15.9657 - 1 -2087 16.9225 3083.6387
1 34 15.3I63 -1.1562 16.7648 3065.55422
125 14.5045 -1.3139 1 3-7692 3330.49369
I36 14.6127 -1.1562 14.8203 3258.85797
137 12.7726 -1.3139 12.8232 326 1 -00076
I38 1 3.4762 - 1.5766 14.I896 33 15.74776
139 1 2.4479 -1.1036 16.3969 3204.67079
140 12.6 102 -4.0467 12.7707 304 1.I8332
141 1 7.6976 -10.1955 32.058 1 402 1.I9734
133, 16.507 -2.4 I75 23.1764 4077.94897
143 16.0199 -4.835 19.8 129 4003.O8754
1 44 16.1281 -4.6773 2 1 -2845 3082.60263
145 15.8575 -2.0496 22.0202 3973.47079
I46 11.4196 - 1.9971 12.2977 4 150.20236
147 9.4171 -1.1036 10.3006 4 149.40209
148 8.3888 -0.9985 8-251 3096.05529
149 9.363 -0.7883 1 1 -7721 4027.7 161 1
150 8.0099 -0.3 153 8.3036 4274.2379 1
-199-
Max Shoulder Min Pelvis Max Pelvis Total Energy Peak Absolute
AcceI. (g's) Accel. (m Absorption (Ibfcin) Impact Load (Ibf)
1 8.9965 -2.2073 22.756 3460.34 17 21 15.0129
18.7259 - 1.8394 20.969 1
17.2647 -1.3139 18.2363
16.6152 - 1.2087 1 8.76 19
17.1023 -2.1022 18.0787
14.99 16 -2.0496 15.6086
1 3.2597 -1.3139 13-6641
13.3138 -1.051 1 15.5035
12.7726 -1.1562 13.2437
14.288 -3.4686 16.975
18.8883 - 10.4583 34.7909
17.481 1 -2.6277 26.5924
16.3987 4.7299 23.5443
1 6.1528 -5.0978 23.9 122
1 3.909 1 -6.7269 22.2304
14.5586 -6.359 1 21.2319
12.502 -4.1518 14.7 152
12.1231 -2.1022 12.3502
12.069 - 1.3664 12.9809
1 1.7443 -0.946 12.9283
15-7493 -6.62 18 25.226
15.5328 -3.3 109 22.0202
15.5328 -5.99 12 15.9765
15.3163 -5.2554 2 1.7574
14-6127 -5.99 12 22.0202
15.5328 -6. I488 23.0 187
14.6668 -4.5 197 18.1838
15.a1 -2.2073 14.8203
12.5561 - 1 -9971 15.1356
10-3371 - 1.5766 12.4028
23.3262 - 12.5079 27.696 1
2 1.432 -7.5 153 26.5399
13.3138 -3.94 16 16.1341
16.1281 -5.5 182 22.5458
15.7493 -4.6248 19.0772
13.288 - I .8394 16.6597
14.6127 - I .7343 14.8203
14.1256 -2.7854 15.1356
13.0174 -2.4 175 15.6086
14.1256 -1.6817 1 5.2407
23.5427 -3.73 1 3 23.0713
23.8674 -5.9386 25.0 158
17.3188 - 1.9445 18.0261
16.6IS2 -2.1022 16.9225
17.481 1 -2.1022 19.0772
1 5.3704 -1 -5241 15.0305
15.2622 - 1-5766 15.1356
16.9399 - 1.9971 19.1297
1 5.2622 - 1.892 17.2378
1 5.4787 - 1.892 18.6042
-20 1-
Max Shouldei Min Pelvis Mair Pelvis Total E n e r a Peak Absolutc
Accel. (g's) Accel. (g's) Absorption (Ibf*in) Impact Load ( I b n
l9.7oOl -2.943 21.8626 4377.04986 2ûû3-2595
25.924 -4.0467 25.0 158 4262.52738 1967.3855
17.5353 -2.47 16.975 4259.1 1426 1983.5147
17.2647 -2.5752 17.1852 4322.65 183 1 942.9 192
18.1306 -3.1533 19.1297 3977.28878 2 162.1363
18.4553 -2.943 18.9721 4063.73573 2 1oO.4825
20.6202 -2.9956 19.7604 3905.0408 1 21 33.5969
17.3729 -2.6277 16.4495 3955.44524 2009.9322
19.5919 -3.1533 19.6553 3750.37293 2147.0619
22.785 -5.1503 26.4348 3866.88604 220 1 -0782
19.7542 -2.5226 23.49 I7 4 1 17.08635 22 17.392
22.785 -3.416 26.3 822 4056.85608 2227.979
18.6177 - 1.5766 19.3925 4169.41461 2090.2 197
2 1.3779 -3.6788 26.277 1 3879.46494 2301.4954
17.481 1 -1.4715 18.2889 4 108.20098 204 1 S969
16.507 -1.5241 17.2378 4 170.19777 1839.2986
17.7517 -1.4715 17.6057 4233.49597 1 848.06 14
16-8317 - 1.2087 16.7648 4026.55079 I 777.9697
17.2105 -2.3 1 24 19.1823 4 1 13.53476 1927.0875
17.481 1 4.940 1 18.867 3954.63292 1869.6346
19.4836 - 1.892 22.4932 4252.33066 2 180.3375
19.213 -2.1022 2 1 -4947 4228.50688 2 197-7887
2 1.3779 -3.6788 26.277 1 3879.46494 230 1.4954
18.7259 - 1.7343 21.1268 41 lO.9938I 2167.1407
17-481 1 -1 -6817 18.5516 4 185.83059 2096.934 1
17.1O23 -2.1022 18.9 195 4 182.70308 21 13.8106
14.8833 - 1.7343 15.2933 4343.40844 1761.3135
14.3421 -1.419 15.6086 4484.13436 1722.5266
14.0174 - 1 .?O87 15.7663 4433.55677 1692.7162
13.5303 -1 -4715 14.6626 4352.10605 1 634.2052
18,7259 -3.O48 1 31.2172 5650.349 15 1957.2858
17.5353 -5.0978 28.4844 5636.6 1056 1909.5272
18.347 1 -3.2584 3 1.6902 5725.13142 1 9 12.4079
17.6435 -4.835 22.5983 5397.840 17 2000.3483
17.7517 -8.67 14 24.7005 5096.805 18 2 149.5362
16.7234 -7.2525 25.5939 53 10.6 1942 1901 -9238
19.4836 -7.305 27.2756 5074.5 1303 1913.984
14.9374 -6,3065 2 1.7574 5236.1428 1 1648.9367
16.6693 -5.8861 24.1224 4998.07 18 1867.7797
2 1 S402 -7.7255 34.2654 5039.99709 1 763.5 106
17.9682 - 12.0349 35.3 164 5862.98658 1856.0794
16.8858 -4.9927 25.1209 5496.47444 181 1.7126
17-481 1 -7.1999 19.1823 559 1.66405 1884.3782
18.7259 -5.308 24.753 5370.90834 181 1.1392
17.9682 -6.5693 2 1S472 51 14.99178 1768.0084
17.5894 -4.467 I 24.0698 4845.21012 161 1.5854
15.8034 -5.7284 20.5487 5342.27 1 12 1 799.9278
16.561 1 -5.1503 20.4436 5349.73427 1629.90 1 2
15.7493 4.0467 18.1312 5457.23297 15392887
25.5993 -5.6233 28.8523 4666.3 1 264 1427.7675
-202-
Test r Max Shoulder Min Pelvis Max Pelvis Peak Absolute
= Acce!g's) Absorption (Ibf*in) Impact Laad (Ibn
30 1 23-651 - 1 1 .O889 25.96 18 5796.62892 1626.2837
302 23.1639 -9.302 1 25.699 1682.5928
303 16.1822 4.9927 18.1838 1662.9183
303 18.8342 -6.62 18 1 8.6567 1560.9905
305 17.9141 -6.0437 19.9 18 1532.4584
306 16.2364 -2.9956 18.34 14 1416.5009
307 16.4528 -2.7854 18.6567 1182.5915
308 14.775 1 -2.6277 15.1356 1 143.O462
309 16.2364 -2.4 175 16.1 867 1226.1737
3 10 15.3 163 - 1 -9971 19.6553 1265.3052
31 1 25 .O58 1 -6.4 1 16 24.4377 1776.1669
312 25.5452 -3.73 13 25 .226 1679.9141
313 19.5377 -3.3 109 17.8159 1772.5554
3 13 23.1639 -3.2584 22.4932 1592.5686
315 17.38 1 1 -2.4 175 1 8.76 19 1762.8564
3 16 16-8317 - 1.892 16.1867 1882.2705
3 17 16.7776 -1.6817 15.8714 1845.1633
318 17.427 - 1.8394 15.6086 1897.4192
319 17.6976 - 1.5766 17.7108 2057 -0009
320 19. IO48 -2.5226 15.8714 1924.9659
321 2 1.7026 -3.2584 23.3866 2 IO8.9706
322 2 1.7567 -4.ISI8 23.1764 2202.0 1O8
323 19.7001 -2.9956 21.2319 2086.889
324 17.9682 -3.521 l 19.1823 22 14.655
325 17.5894 -3.1533 20.7589 2 160.954
326 19.8625 -3.2584 22.3355 2 175.3268
327 22.1356 -2.3649 21.0217 2 1 58.59€4
328 22.4603 -3.1533 22.9662 2342.7668
329 20.9449 -3.7854 2 1 .O742 2288.9877
330 20.2954 -3.3 IO9 19.7604 2 180.7502
33 1 23.0556 -2.3649 26.5924 2370.084
332 19.7001 -2.1022 22.283 2357.6449
333 19.8083 - 1.9445 2 1.9677 2320.6533
334 19.3213 -2.1547 20.1808 2224.739
335 20.5 1 19 -3.1533 2 1 -5998 2 195.6756
336 19.213 -2.47 19.4976 2101.76
337 19.3836 -2.3649 19.2874 2009.13 19
338 18.347 1 -2.1547 19.5502 2043.2623
339 18.5636 -2.943 20.3385 2024.76 14
340 17.5353 -6.7795 19.8139 I88l.1439
34 1 20.133 1 -2.47 24.6379 2300.3832
342 20.9449 - 1.892 25.1209 23 13.0528
343 20.4578 - 1.6292 2 1.7049 2299.O246
343 20.6202 - 1.5766 23 -4917 2358.1424
345 20.7825 - 1.6292 24.5954 2353.745
346 19.32 13 -3.73 13 2 1 S998 23 10.7723
337 17.86 -2.2073 19.918 2036.3434
348 16.7776 -2.943 18.394 1899.9097
349 14-731 -1.4715 18.7619 1813.1177
350 15.208 -3.7839 21.2319 1776.7367
-203-
Max Shouidei Min Pelvis Max Pelvis Total Energy Pcak Absolute
Accel. (g's) Accel. (g's) -b(Ibr in) Impact Load (Ibo
20.999 -1 1 -667 37.9967 6633.1378 2264.407
18.78 -3.4686 3 1.O595 6408.0955 2234.8829
19.213 4.7824 28.1691 6415.83719 2358.66 15
19.8083 -5.413 1 33.1617 608 1.62305 2262.8792
20.2954 -9.302 1 27.7486 5555.27 181 2468.0563
22.352 -8.829 1 29.3778 5379.59 159 2502.7 1
2 1.7567 -8.0408 33.1092 5396.866 1 1 2328.6074
20.133 1 -7.6729 30.7968 5579.05 156 2 1 17.2976
19.9707 -6.254 29.3778 5442,13044 22 15.8764
17.481 1 4.0467 25.1735 5703.19063 1917.3 192
19.213 - 12.4028 37.2084 6592.44925 2039.9338
18.5636 -10.143 3 1 -48 6176.15831 203 1.34
22.1356 -5.6759 24.0698 5949.69855 2 loO.2477
19.8625 -6.1488 2 1 -4947 6376.9 1779 21 55.0521
2 1 6485 -6.8846 2 I .4947 6 120.4 1573 2217.8109
19.8083 -5.2029 23.702 5903-69509 2154.57I l
18.0765 -6.0437 23.9647 585 1.99748 2 109.2547
18.1847 -5.9386 24.0 173 58 19.25799 2062.90 15
18.5094 -5.6759 25.226 5816.68 14 1996.4038
20.7284 -4.0992 27.7486 5770.14628 2049.8728
24.0839 -8.3561 28.1691 6565.15035 1862.7242
24.625 1 -8.4087 30.06 1 6154.01 125 1790.0928
17.5894 -6.20 14 20.3385 6 164.03535 1775.4964
20.6202 -6.359 1 1 8.499 1 6244.24508 1779.8342
19.646 -6.4642 20.2859 6005.02649 1685.6188
17.1564 -5.5 182 19.O246 5739.86087 1627.3893
18.5093 -6.1488 18.8144 6225.86205 1468.7 149
17.6976 -2.8905 18.026 1 6007.9983 1 12%. 1291
16.83I7 -2.7854 17.9735 6527.67796 121 1.6364
14.5045 -3.1007 16.3444 6434.11094 1305.0228
32.0939 - 1 3.2962 34.1602 6656.0905 1746.3423
26.8982 -10.5108 29.5355 6424.16053 1778.628
27.6559 4.5722 26.9603 5962.0799 1761.305
25.0581 4-5722 22.1253 61 15.29318 1751 -9477
19.1038 -3.4686 1 9.4976 6 194.77497 1884.0168
17.86 -2.7854 16.8 174 6065.4 1697 2044.6096
17.86 -2.3649 16.1341 6032.72805 1967-3297
18.9965 -2.3 1 24 18.2889 621 1.74085 1973.8452
18.3471 - 1 -9445 18.1838 6039.65539 2047.7345
19.5919 -2.6803 18.2889 6344.00009 2 107.9009
24.1922 -3.5737 25.226 6263.75624 2139.1798
24.0839 -4.362 23.8596 5779.74 158 2 199-6347
20.2954 -2.8905 20.7589 58 12.2928 2253.484
2 1 2696 -2.9956 22.7034 596 I .78 1 2244.3379
19.213 -2.47 22.1779 5702.54242 2247.73 13
24.0839 -4.6773 22.0728 56 13.56725 2239.359 1
20.566 -2.3649 20.60 1 3 5524.20498 2288.6978
22.2979 -2.8379 22.5458 5475.6635 1 2449.5936
2 1.2696 - 1.997 1 24.753 5988.59 23 1 5.2725
22.8933 -2.4 175 30.2 187 5590.85486 2448.8971
-204-
Tcst i Max Shoulder Min Pelvis Max Pelvis Peak Absolutc
Accel. (gTs) Acccl. (p's) Absorption (lbf'in) Impact Load (1 b t)
40 1 26.6276 -6.7269 28.4844 599 1.O2202 2338.187
402 23.8 133 -5.2029 27.7486
403 2 1 -3237 - 1.8394 22.5458
403 20.8367 -2.3 124 2 1-4947
405 19.7001 -3.1533 2 1.6523
306 19.8625 -1.997 1 2 1 S998
407 19.646 -2.9956 17.5006
408 19.2671 -2.6803 19.9706
409 20.24 1 3 -2.8905 22.8085
410 19.5919 -6.5693 26.750 1
cll l 22.4603 -2.0496 24.1749
412 24.4628 -2.47 28.43 18
cl1 3 23.00 15 -2.4 175 26.5924
4 14 22.785 - 1.997 1 27.223 1
415 22.0273 - 1.8394 25.54 1 3
416 2 1.865 -4.940 l 24.2275
417 19.3754 -4.5722 26.5399
4 18 20.24 13 -5.6233 25.0 158
4 19 1 9 .O507 -5.5707 27.59 1
420 16.3446 -5.1 503 17.3954
Potential Energy Energy ai Impact E.A.F. based on E.A.F. based on Average E.A.F.
Potential Enerpy Impact Energy
0.7 197 0.6990
Potential Energy Energy at Impact E.A.F. based on E.A.F. based on Avcnge E.A.F.
(IbPin) Potential Energy Impact Energy
b
Material Properties and Sample input Data for the Aluminum Crosshead
MATERIAL PROPERTIES AND SAMPLE INPUT DATA FOR THE
ALUMINUM CROSSHEAD
Material Properties and Sample Input Data for the Wooden Backing Plate
MATERIAL PROPERTlElS AND SAMPLE INPUT DATA FOR THE WOODEN
BACKING PLATE
Material Property
i
Value (with corresponding units)
I
I Density
--
1 - -
1244 kg/m3
-
1
Young's Modulus 1 1 GPa
i
Material Propenies and Sample input Data for the Low Density Polyurethane Foam
MATERIAL PROPERTIES AND SAMPLE INPUT DATA FOR THE LOW
DENSITY POLYURETHANE FOAM
Material Properties and Sample Input Data for the Steering Wheel Aluminum Alloy
MATERIAL PROPERTIES AND SAMPLE INPUT DATA FOR THE STEERING
WHEEL ALUMINUM ALLOY
'3EF :SE-CEVE
S TKTS IS T E 2 30AD C U R E FOR THE S T R E S S V S . EFFECTim PLASTIC S T E W I N
5 FO?. IXZ ALLXIN[R.I M A T S R I A L ( U N I T S M E BASSD UPON S A S E SET-RG,MM.SZC)
S iC 13 SIDR SCLA SCLO OFFA OFF0
L O
S STRAIN STRESS
0.3~COOC000000 70333,06712
0.003924830310 78986.52068
0 . OOe752652348 89018.03038
O . 013564565110 98380.72751
0.ù?Z361082510 107143.5089
0.023140387900 115369.8456
0.327902336490 123117.6128
0.032046457660 130438.9184
0.Û37372357360 137379.9328
C.242079720430 143980.7i81
O. 0 4 6 7 6 8 3 1 3 0 3 0 150275.0576
0.051137984900 156290.2847
0.056086671800 162047.1125
0.060720397820 167559-4633
0.065233277740 172834.2974
O . O69927519390 177871.4429
O . O74503426020 182663-4247
0.079061398580 187195.2943
0.083601938140 191444-4583
C.066125648230 195380 - 5084
0.092033237130 198965.0507
0.997125520270 202151. 5346
O.iOLOG3422600 204885.0822
0.106067980700 207102-3183
Q.i1052û345600 208731.1986
O. 7 1 2 7 4 2 3 4 7 0 0 0 209300.0793
APPENDIX G
Sample Data From the LS-DYNA Input File Regardinp Hourglass Control
SAMPLE DATA FROM THE LS-DYNA INPUT FILE REGARDING
HOURGLASS CONTROL
c--------:----+----2----+----3----+----4----+----5----+---- 6---------7--------- a
s ADDITIONAL LINESS ADDED a~ RILL ALTENHOF (HOURGLASS C O E ~ R O L )
C--------1---------2----~----~-7----+----4----+----5----+----~----+----7----+---- 8
5 TEE SGUZGLAÇS CONTROL USED I N ALL S I M U L A T I O N S IS THE
5 FLAtVAGAii-SELYTSCHKO S T I F L N E S S FORM WITH EXACT VOLUME ZNTEGPATION (TYPE 5 )
S ?'XE HOURGLASS C O E F F I C I E N T IS SET A S T H E DEFACnT QM=O.lO
S U O U R G G S S CONTROL IS ONLY IMPLEMENTED FOR THE BODYFORM PART d l
'3 G U G G Z S . S
S 1UQ QM 1EQ Q 1. Q2 QB QW
5 0-1 L
$--------:----c----2---------j-----4----+----s----v----6----+----7----+-- --8
APPENDIX H
Sample Data From the LS-DYNA input File Regarding Contact and Mass Scaling
SAMPLE DATA FROM THE LS-DYNA INPUT FILE REGARDING CONTACT
AND MASS SCALING
Sample Input Data from the LS-DYNA input file (contact algorithm):
Sample Input Data from the LS-DYNA input file (timestep and mass scaling):
S VASS SCASING I S INVOKED 9Y INCORPORATING THZ A P P R O P R I A T E VALUE UNDER 9TMS
S 20.. T?S!SIENT ANALYSES DT'MS MWST 51 N E G A T I V F !
S INITIAL S I . % L A T I O N HAVE PROVEN TFAT MASS S C A L I N G TO THE EXTEXT C O X P L E T E D
S IPi T E S E S I . W L A T I O N S D I D NOT AFFECT T H E RESULTS.
S FOR THE ?UR?OSES O F T H I S STUDY, !4ASS S C X I N G WAS ONLY PERFOF.YE9 O N THE
.
S S-==TiI:;G
--+ I.'EEZL, NO A D D I T I O N A L MASS WAS ADDED TC T H E DROPPIPIG ASSEXBL'L'
S C O ? i S I S T I ? : G O F T'SE CROSSHEAD, WOODEN BACKING P L A T E , AND DECORMASLE D m Y .
C---+----'----+----2----+----3----+----4----*----5----+----6---------7----*---- 6
'COIK'FZOL-T IMEST EP
S X'I?lIL SCFT ISW TSLIMT DTMS LEM ERODE MS1ST
.O 0 0 -900 O -3. G E - 0 7
~--------:----+----2----~----3---+----4----+----~----+----5----*----7----+---- 8
VITA AUCTORIS