Professional Documents
Culture Documents
by
January 2011
CASE WESTERN RESERVE UNIVERSITY
DOCTOR OF PHILOSOPHY
candidate for the ______________________________degree *.
EDWARD BENZEL, MD
________________________________________________
________________________________________________
3 DECEMBER 2010
(date) _______________________
*We also certify that written approval has been obtained for any proprietary
material contained therein.
For Kendra and Elijah, Mom and Pop, and Ben, Noah and Jason -- this effort
was only possible through your years of love, patience and endless
encouragement
TABLE OF CONTENTS
I
METHODS...................................................................................... 88
RESULTS ....................................................................................... 96
DISCUSSION.................................................................................107
LIMITATIONS ................................................................................111
CONCLUSIONS ..............................................................................112
CHAPTER 6 – HELMET TESTING IMPACT SIMULATIONS ..........................115
INTRODUCTION.............................................................................115
BACKGROUND...............................................................................116
METHODS.....................................................................................121
RESULTS ......................................................................................129
DISCUSSION OF RESULTS ..............................................................135
LIMITATIONS ................................................................................139
CONCLUSIONS ..............................................................................139
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL
HELMET IMPACTS .............................................................................141
INTRODUCTION.............................................................................141
IMPACT PENDULUM........................................................................142
CADAVERIC IMPACTS .....................................................................153
HELMETED ATD IMPACTS................................................................167
LIMITATIONS ................................................................................177
CONCLUSIONS ..............................................................................178
SUMMARY........................................................................................181
APPENDIX........................................................................................186
BIBLIOGRAPHY.................................................................................191
II
LIST OF TABLES
Table 1-1. The 113 struck vehicle occupants with crash related diagnoses .. 9
Table 1-2. Reconstruction struck vehicle delta-V and mean acceleration
(n=105 crashes)................................................................................ 10
Table 2-1. Passenger vehicle crash tests used as inputs to MADYMO minor
rear crash simulation.......................................................................... 28
Table 2-2. Comparison of additional validation parameters ..................... 34
Table 2-3. Neck injury risk values as a percentage of threshold ............... 35
Table 2-4. Head injury risk values as percentage of threshold ................. 35
Table 2-5. BioRIDII output for pulse scaled by 40% and change from original
response .......................................................................................... 39
Table 3-1. Abbreviated Injury Scale (AIS) ............................................ 52
Table 3-2. Classification comparison for Nij (2000), Nusholtz et al. (2003)
and current study .............................................................................. 57
Table 3-3. 40% injury risk intercepts for the four parameters based on
regression curves .............................................................................. 58
Table 4-1. Test impact matrix............................................................. 68
Table 4-2. Coefficient of variation among all impact trial sets (Front =
Frontal, Obl = Oblique, Lat = Lateral) ................................................... 73
Table 5-1. Test impact matrix............................................................. 91
Table 5-2. Theoretical post-impact head velocity ................................... 94
Table 5-3. Coefficient of variation of each resultant impact parameter
corresponding to time of highest cg linear acceleration resultant .............. 99
Table 5-4. Low energy impact maxima (shaded = significant increase
p<0.05, italics = significant decrease p<0.05) ......................................104
Table 5-5. High energy impact maxima (shaded = significant increase
p<0.05, italics = significant decrease p<0.05) ......................................107
Table 6-1. Test impact matrix............................................................123
Table 6-2. Simulink simulation parameters for constant kinetic energy
comparison ......................................................................................128
Table 6-3. Simulink simulation parameters for constant momentum
comparison ......................................................................................129
Table 6-4. Additional validation parameter comparison.........................133
Table 7-1. Cadaver impact matrix ......................................................156
Table 7-2. Impact matrix for football helmet power analysis (*to avoid
facemask contact, frontal and oblique impacts used a 5° forward inclination)
......................................................................................................170
III
LIST OF FIGURES
IV
vs. Maximum Force Loading Rate. Volume intercepts drawn at 40% injury
risk, sphere radius equals AIS injury severity......................................... 56
Figure 4-1. Frontal, Oblique and Lateral impact setups ........................... 68
Figure 4-2. Resultant ATD head center of gravity (cg) linear acceleration for
each set of six (6) impact trials in the three impact directions at highest
impact energy ................................................................................... 69
Figure 4-3. Resultant ATD occipital force for each set of six (6) impact trials
in the three impact directions at highest impact energy........................... 70
Figure 4-4. Mean ATD head center of gravity (cg) linear acceleration
components from each set of six (6) impact trials in the three impact
directions at highest impact energy ...................................................... 71
Figure 4-5. Mean ATD occipital force components from each set of six (6)
impact trials in the three impact directions at highest impact energy......... 72
Figure 4-6. Resultant ATD head center of gravity (cg) linear acceleration for
the three impact directions compared with impact energy [NOTE: error bars
show +/- one standard deviation] ........................................................ 74
Figure 4-7. Resultant ATD head angular velocity ................................... 75
Figure 4-8. Resultant ATD head angular acceleration ............................. 76
Figure 4-9. Resultant ATD occipital bending moment ............................. 77
Figure 4-10. Resultant ATD occipital force ............................................ 78
Figure 5-1. Unpadded Impact Setup .................................................... 91
Figure 5-2. Bare Pendulum-Headgear .................................................. 92
Figure 5-3. MMA-Bare Head ............................................................... 92
Figure 5-4. Boxing-Bare Head............................................................. 92
Figure 5-5. Boxing-Headgear Impact Setup (High Energy Only)............... 93
Figure 5-6. Simplified elastic collision model ......................................... 94
Figure 5-7. Head cg resultant linear acceleration in high energy trials for the
four padding conditions tested ............................................................. 97
Figure 5-8. Occipital neck force resultant in high energy trials for the four
padding conditions tested ................................................................... 98
Figure 5-9. Head cg resultant linear acceleration in high and low energy
trials ...............................................................................................100
Figure 5-10. Head resultant angular velocity in high and low energy trials100
Figure 5-11. Head resultant angular acceleration in high and low energy
trials ...............................................................................................101
Figure 5-12. Resultant impact force in high and low energy trials............101
Figure 5-13. Sample of computational simulation results displaying Mises
stresses within brain from Box-Bare impacts using SIMon software..........102
Figure 6-1. Pendulum Impact setups for (A) Unhelmeted, (B) ‘Leatherhead’,
(C)-(E) NOCSAE Approved Helmets .....................................................123
Figure 6-2. Two DOF system dynamics model for SEA, Ltd. lateral pendulum
helmet impacts.................................................................................124
Figure 6-3. Unhelmeted lateral NOCSAE calibration drop test system
dynamics model ...............................................................................125
Figure 6-4. Helmeted NOCSAE lateral impact test system dynamics model126
Figure 6-5. Proposed NOCSAE linear impact with highly rigid neck
attachment ......................................................................................127
V
Figure 6-6. Linear impact used in prior NFL studies with flexible Hybrid III
neck ...............................................................................................127
Figure 6-7. Hybrid III head cg linear acceleration for the single (n=1)
pendulum head impacts.....................................................................130
Figure 6-8. Hybrid III head angular velocity for n=1 pendulum head impacts
......................................................................................................130
Figure 6-9. Hybrid III head angular acceleration for n=1 pendulum head
impacts ...........................................................................................131
Figure 6-10. Hybrid III head upper neck shear force for n=1 pendulum head
impacts ...........................................................................................131
Figure 6-11. Hybrid III head upper neck bending moment for n=1 pendulum
head impacts ...................................................................................132
Figure 6-12. Simulink simulation results compared with average response
from Hybrid III head cg linear acceleration for the three (3) helmeted
pendulum impacts ............................................................................133
Figure 6-13. Constant energy Simulink results for headform cg linear
acceleration .....................................................................................134
Figure 6-14. Constant momentum Simulink results for headform cg linear
acceleration .....................................................................................135
Figure 7-1. Examples of impact pendulums -- BioKinetics (Ottawa, Canada),
NOCSAE (Knoxville, TN), Johns Hopkins Applied Physics Lab (Baltimore, MD),
Disney Epcot Center (Orlando, FL), Texas Transportation Institute (College
Station, TX), Impact Jackets LLC (Silver Spring, MD) .............................143
Figure 7-2. SolidEdge Rendering and first version.................................144
Figure 7-3. Retrofit rendering and assembled final pendulum .................145
Figure 7-4. Cadaveric and ATD impact setups ......................................146
Figure 7-5. Electromagnets used in cadaver [500N] and ATD [1500N]
impacts ...........................................................................................147
Figure 7-6. Frame vibration analysis hammer strike locations.................148
Figure 7-7. FFT results for the four (4) frame hammer strikes. NOTE: Chart
areas enlarged from original scale of Frequency = 5000Hz, lY(f)l = 0.8 to
show locations of interest ..................................................................149
Figure 7-8. Fixture mass quantification for fastening directly to load cell and
fastening to fixture components sitting atop load cell .............................150
Figure 7-9. Fixture mass effects - headform resultant linear acceleration .151
Figure 7-10. Fixture mass effects - headform resultant angular velocity...152
Figure 7-11. Fixture mass effects - load cell resultant bending moment ...152
Figure 7-12. Cadaveric specimen preparation: Removing mandible,
Removing caudal structures and sawing through skull, Testing orientation,
Potting with cerroband liquid metal – sawbones replica ..........................155
Figure 7-13. Instrumentation and testing sequence: Affixing sensors to
mounting blocks, Initial cadaver position at 40-degrees forward tilt, National
Instruments DAQ system and Occipito-cervical disruption caused in specimen
001 after sixth impact .......................................................................157
Figure 7-14. Resultant linear acceleration results from three cadaver
specimens for the fifth impact (KE = 25.9J)..........................................158
Figure 7-15. Sagittal plane angular velocity results from three cadaver
specimens for the fifth impact (KE = 25.9J)..........................................159
VI
Figure 7-16. Sagittal plane angular acceleration results from three cadaver
specimens for the fifth impact (KE = 25.9J)..........................................160
Figure 7-17. Resultant force results from three cadaver specimens across
four impacts (KE = 4.7J, 10.3J, 14.5J, 25.9J) .......................................161
Figure 7-18. Resultant moment results from three cadaver specimens across
four impacts (KE = 4.7J, 10.3J, 14.5J, 25.9J) .......................................162
Figure 7-19. Resultant linear acceleration power analysis results from three
cadaver specimens and four instrumented spinal levels ..........................163
Figure 7-20. Angular velocity power analysis results from three cadaver
specimens and four instrumented spinal levels......................................164
Figure 7-21. Angular acceleration power analysis results from three cadaver
specimens and four instrumented spinal levels......................................165
Figure 7-22. Force and moment power analysis results from three cadaver
specimens and four instrumented spinal levels......................................166
Figure 7-23. Football helmet power analysis impact setup alignment, lateral
impact, oblique impact, frontal impact, helmet positioning, overall view ...169
Figure 7-24. ATD-measured linear acceleration, angular velocity, angular
acceleration and bending moment from frontal pendulum impacts at 84J
(5.0m/s) .........................................................................................171
Figure 7-25. ATD-measured linear acceleration, angular velocity, angular
acceleration and bending moment from oblique pendulum impacts at 84J
(5.0m/s) .........................................................................................172
Figure 7-26. ATD-measured linear acceleration, angular velocity, angular
acceleration and bending moment from lateral pendulum impacts at 84J
(5.0m/s) .........................................................................................173
Figure 7-27. Lateral football helmet impact power analysis (73/75 conditions
had P = 80% or higher).....................................................................174
Figure 7-28. Oblique football helmet impact power analysis (71/75
conditions had P = 80% or higher)......................................................174
Figure 7-29. Frontal football helmet impact power analysis (67/75 conditions
had P = 80% or higher).....................................................................175
Figure 7-30. NFL in-game reconstructed football helmet impact power
analysis (13/30 conditions had P = 80% or higher) ...............................176
VII
ACKNOWLEDGEMENTS
The assistance of Kay Kress, Sue Legg and Adam Ratliff in the gathering of
contribution of the crash test data from Gunter Siegmund and Bradley
Heinrichs of MEA Forensic Engineers and Scientists made the MADYMO model
and Chris Iams of SEA Limited deserve thanks for their roles in facilitating
during MADYMO model development and for providing the BioRIDII model at
no charge. The work performed in Chapter 3 would not have been possible
without heavy assistance from Guy Nusholtz from Chrysler who shared the
original piglet-ATD datasets and helped to vet some of the analysis methods
used. Also, Dr. Serkan Inceoglu was invaluable as a sounding board to help
Chapters 4, 5 and 6 would not have been possible without the help from SEA
Vincent Miele who provided human pilot impact data in order to help calibrate
the pendulum impacts. Special thanks go to Art Chou, Brian Hoying and
Greg Mathews and Rawlings sporting goods for providing their prototype
football helmet for the pilot impact tests. The work performed in Chapter 7
assisted greatly in preparing the kinematic sensors and data acquisition for
VIII
testing. Mike Oliver and Dave Halstead at NOCSAE provided the headforms
testing. Finally, a portion of this work was supported under the National
Grant AR050959.
IX
BIOMECHANICAL ENGINEERING ANALYSES OF HEAD AND SPINE
IMPACT INJURY RISK VIA EXPERIMENTATION AND COMPUTATIONAL
SIMULATION
ABSTRACT
by
Head and spine injuries, such as traumatic brain injury, skull fracture,
spine acts as the attachment point for the head, the boundary conditions
applied to the cervical spine influence the behavior of the head. Hence, in
analyzing injury risk for the head and the spine, each structure composes
risk for the head and the cervical spine might be more appropriately
head and cervical spine injury risk criteria developed from human, animal
studies separately analyzed head injury risk via linear acceleration, Head
Injury Criterion (HIC) or Gadd Severity Index (GSI), or cervical spine injury
X
risk via axial/shear forces, bending moments or the Neck Injury Criterion
Thus, designing for optimal head and cervical spine injury protection
engineering studies that comprised the chapters of this dissertation, the work
presented here broadly investigated head and cervical spine injury protection
low energy minor rear car crashes was analyzed. In Chapter 2, these minor
crash loading, Chapter 3 explored multivariate head and cervical spine injury
implications from direct head loading during frontal airbag inflation in high
test device. The low- and high-energy injury analysis methods developed in
XI
pendulum, pilot cadaveric injury response to direct head impact and analysis
related studies indicated that head and cervical spine injury risk were
XII
CHAPTER 1 – REAL WORLD MINOR REAR CRASHES
INTRODUCTION
In this Chapter, real world minor rear crashes in the United States and
the associated injury diagnoses are examined. These crashes, with struck
vehicle speed change of less than 15km/h, have long been implicated to cause
‘whiplash’-related head and neck injury and cost many billions of dollars yearly
in lost wages, insurance costs and medical bills. Yet, little is actually known
about the actual injuries seen in the field relative to vehicle kinematics. In this
BACKGROUND
significant source of head and spine injury related to inertial and/or direct
currently exists a paucity of real world minor rear aligned crash data for crashes
with struck vehicle delta-V less than or equal to 15 km/h. Because of this lack
of real world data, surprisingly little is actually known about the prevalence of
head and spine injury, including ‘whiplash’, in these crashes. And while previous
examined the relationship between minor rear crash struck vehicle delta-V and
risk of occupant ‘whiplash’, head and spine injury, no large scale studies exist
Eis et al. (2005) examined the German In-Depth Accident Study (GIDAS)
and found that for single rear impact crashes involving 1,724 struck vehicle
1
CHAPTER 1 – REAL WORLD MINOR REAR CRASHES
occupants, 59% of occupants were uninjured and 39% sustained a minor AIS13
injury. Of the AIS1 injured occupants, injury data for crashes with engineering-
available for 314 occupants. For these 314 occupants, 69% had AIS1 neck
injury, 16% had no neck injury and 15% had AIS1 neck injury in addition to
‘other’ injuries. Hell et al. (2002) analyzed crashes from the Gesamtverband der
Krafft et al. (2005) conducted a study of Swedish crashes pooled from more
than 60,000 vehicles equipped by Folksam Insurance with event data recorders
(EDR). From these vehicles, a total of 171 struck vehicle occupants were
involved in rear crashes with EDR recorded delta-V less than or equal to 15
km/h. Of these struck vehicle occupants claiming injury, 67% had no ‘whiplash’
injury, 26% had AIS1 ‘whiplash’ injury for less than one month post-crash and
7% had AIS1 ‘whiplash’ injury for longer than one month post-crash. Of
particular note to the current study was that in both of the minor rear crash
German studies and the Swedish study, the authors made no mention of any
There are few U.S. minor rear crash data to directly compare with the
collection currently exists in the U.S. Tencer et al. (2001) had a sizeable study
previously diagnosed with ‘whiplash’ attributed to rear crashes with peak delta-V
less than 11.3 km/h, were analyzed. All subjects were proceeding to litigation.
Immediate symptoms were considered but the specific number of days between
the crash and symptom onset was not reported. A total of 174 subjects had
AIS1 neck and upper back pain at mean delta-V of 8.0 km/h (s.d. = 2.1 km/h),
174 subjects had AIS1 neck and low back pain at mean delta-V of 7.7 km/h
(s.d. = 2.9 km/h) and 84 subjects had AIS1 neck pain and arm symptoms at
mean delta-V of 8.8 km/h (s.d. = 5.0 km/h). The authors concluded that pre-
existing lumbar degeneration was associated with (a) having arm symptoms or
(b) AIS1 low back pain with neck pain. While the entire study cohort reported
AIS1 neck pain, there were no reported head, thorax, abdominal or lower
extremity complaints.
While much can be learned from the prior studies, further U.S. minor rear
vehicle delta-V, and occupant ‘whiplash’, head and neck injury risk. The lack of
available U.S. data is particularly startling when considering that the National
806,000 occupants sustain ‘whiplash’ and minor head injuries in motor vehicle
crashes producing modern day economic and quality of life costs over $9 billion
yearly79. Over the past six decades, much has been done to improve the
treatment protocols, and vehicle bumper, seat and head restraint designs. Yet,
even with the voluminous amount of research in the engineering and medical
fields, ‘whiplash’ and minor head injuries in rear crashes have increased since
under the U.S. tort system have been cited as a contributor to current excess
3
CHAPTER 1 – REAL WORLD MINOR REAR CRASHES
engineering data set was analyzed. Crash data were sought from forensic
engineering sources since this is one of the few sources of large numbers of
minor rear crashes in the U.S. The study intended to be a first step in
‘whiplash’, spine and head injury in the U.S. Primarily, what are the delta-V and
minor rear crashes in the U.S.? Secondly, what is the complaint diagnosis
distribution in these crashes based on ICD-9-CM156 and AIS coding? Thirdly, can
currently known minor rear crash injury mechanisms explain the ICD-9-CM and
AIS coded diagnoses in these crashes? And lastly, does ICD-9-CM coding from
medical records differ when compared with AIS coding? Armed with answers to
these fundamental questions, safety engineers and clinicians alike will be able to
design vehicle systems to reduce ‘whiplash’, minor head injury and spine injury
insurance claims processing for legitimate minor rear crash injury and more
insurance claims.
METHODS
Real world minor U.S. rear crashes were sought via (I) a database search
and (II) a review and analysis of crash files from a forensic engineering
company. Inclusion criteria for the current study limited data to (a) U.S. real
4
CHAPTER 1 – REAL WORLD MINOR REAR CRASHES
world rear crashes, (b) crashes between two passenger vehicles with gross
weight less than 4,536 kg for each vehicle, (c) single contact between the front
of the striking vehicle and rear of the struck vehicle, (d) aligned contact between
the two vehicles, (e) full overlap between the two vehicles, (f) crash occurring
on a flat roadway, (g) reconstructed struck vehicle delta-V less than or equal to
chiropractor (D.C.).
A database search for U.S. real world minor rear crashes was conducted.
(SCI) and the National Trauma Databank (NTDB v6.2) were searched for all data
data from which to study minor rear crashes and resulting head and spine injury
risk, real world crash files from a forensic engineering company in the U.S. (SEA
Limited, Columbus, OH) were examined for crashes occurring from 1994 to
2006. In all crash files, occupants filed at least one property damage or injury
claim with a U.S. insurance company and were potential litigants in a U.S. court.
These files included available vehicle and scene inspections, police reports,
utilized for an engineering crash reconstruction. From these files, 105 minor rear
aligned crashes from 15 states met the inclusion criteria. These crashes
involved a total of 126 occupants in the striking vehicles and 151 occupants in
the struck vehicles. Front seat struck occupants totaled 137 persons. A total of
113 struck vehicle occupants were diagnosed with complaints attributed to the
For the 113 struck vehicle occupants, medical records were examined for
complaint diagnoses within five weeks post-crash. These records were obtained
(HIPAA) requirements158. Additionally, these records were redacted such that all
diagnosis was defined as any distinctive diagnosis given to the occupant during
the five week treatment period. Thus, the occupant could be diagnosed with a
cervical sprain repeatedly during visits to different clinicians, but this diagnosis
would only count for the first visit when it was originally diagnosed. Unique AIS
diagnoses differed slightly from ICD-9-CM in that AIS codes included only acute
injuries. Pain diagnoses were recorded as AIS1 injuries. Each crash was
individually reconstructed and the struck vehicle delta-V and acceleration were
Crash reconstruction
normal force restitution and kinematic equations of motion for the vehicles can
Impulse-momentum relationship:
m1v12 2 m2 v 22 2
m1v11 m2 v 21
2 2
w2 C2 m BEVi
2
2
2 Eq. 1-2; FCi dCdw i
2
Eq. 1-3
w 2 C 2 F dCdw
w1 C1
w1 C1 Ci
Kinematic equations of motion:
vi
vi 2 2a i xi Eq. 1-4; vi vi 2 vi1 Eq. 1-5; ai Eq. 1-6
t
Restitution:
v 22 v12
e Eq. 1-7
v11 v 21
Where:
mi vehicle ‘i’ mass (kg)
v1i vehicle ‘i’ pre-crash velocity (m/s, ‘i’ = 1,2)
vi 2 vehicle ‘i’ post-crash velocity (m/s, ‘i’ = 1,2)
ai vehicle ‘i’ average acceleration (m/s2)
xi vehicle ‘i’ linear displacement from impact area to final rest position (m)
vi vehicle ‘i’ velocity change (m/s)
t impact duration (s)
e restitution coefficient for crash contact
BEVi vehicle ‘i’ barrier equivalent velocity (m/s, ‘i’ = 1,2)
t2
F dt impulse integral of vehicle tire forces over contact duration for both vehicles (N-s)
t1
T
w2 C2
F dCdw vehicle ‘i’ energy expenditure during crash due to contact forces, crush depth
w1 C1
Ci
these parameters have been most often used in prior studies.39;77;138 Delta-V is
relating permanent crush deformation energy in a real world crash to the same
7
CHAPTER 1 – REAL WORLD MINOR REAR CRASHES
type of vehicle in a controlled barrier crash test. The BEV is equal to the barrier
impact speed and can be calculated as shown in Eq. 1-3. Given the same real
world and crash test vehicle with similar crush profiles, the real world vehicle
BEV and energy dissipation is approximately the same as the crash test vehicle.
Higher energy crash reconstruction methods are not utilized here because
there is minimal permanent crushing for both vehicles in the crash, the collisions
are treated as fairly elastic, restitution must be considered and tire force impulse
is a contributing factor.17;149
both vehicles. Thus, when crush is absent post-crash, one cannot determine
One possible solution for this problem has been to assume a small crush that is
not actually present and use this value with the vehicle stiffness to calculate a
algorithms based on barrier crash tests and require measurable crush in order to
determine vehicle delta-V. These programs are not validated for minor crashes
and inaccuracies of these reconstruction programs for minor crashes has been
shown previously.106
In minor crashes, one can still apply methods shown in Equations 1-1 through
1-6, but the vehicles are assumed to behave elastically. This means that the
8
CHAPTER 1 – REAL WORLD MINOR REAR CRASHES
crash data provided for vehicle weight, impact duration, restitution, crush
energy and bumper properties. These data, along with calculated and estimated
pre- and post-crash values were inserted into Equations 1-1 through 1-7 and
unknowns.
complaints and AIS injury diagnoses and crash dynamics were compiled and are
RESULTS
CRASH OCCUPANTS
mean age ± s.d. mean weight ± s.d. mean height ± s.d.
Male
(years) (kg) (m)
(n = 36)
39.3 ± 12.0 92.1 ± 12.9 1.76 ± 0.08
mean age ± s.d. mean weight ± s.d. mean height ± s.d.
Female (years) (kg) (m)
(n = 77)
40.3 ± 12.0 78.5 ± 22.8 1.62 ± 0.07
Table 1-1. The 113 struck vehicle occupants with crash related diagnoses
Each crash was reconstructed using the minor crash MER method
acceleration results are shown in Table 1-2, Figure 1-1 and Figure 1-2. For all
9
CHAPTER 1 – REAL WORLD MINOR REAR CRASHES
0.080 to 0.200 seconds with a mean of 0.136 seconds (s.d. = 0.024 seconds).
Restitution references varied from 0.10 to 0.70 with a mean restitution of 0.25
(s.d. = 0.11).
RECONSTRUCTION SUMMARY
Delta-V (km/h) Mean acceleration (g)
mean 6.3 mean 1.4
s.d. 2.1 s.d. 0.5
max 13.7 max 3.3
min 0.8 min 0.3
Table 1-2. Reconstruction struck vehicle delta-V and mean acceleration (n=105 crashes)
50
45
40
35
Frequency
30
25
20
15
10
5
0
< 4.0 4.0-6.0 6.1-8.0 8.1-10.0 10.1- 12.1-
12.0 15.0
Delta-V (km/h)
60
50
40
Frequency
30
20
10
0
< 1.0 1.0-1.4 1.5-1.9 2.0-2.4 2.5-3.5
Mean Acceleration (g)
10
CHAPTER 1 – REAL WORLD MINOR REAR CRASHES
vehicles, 24 occurred with light truck/ sport utility vehicles (SUV) striking a
truck/SUV and 12 occurred between two light truck/SUV. Mean striking and
struck vehicle weights were 1688 kg (s.d. = 506 kg) and 1509 kg (s.d. = 300
kg), respectively. Striking vehicle front bumper composition was rigid (n = 38),
piston (n = 17), polymer (n = 11), foam (n = 26) and box (n = 3). Struck
vehicle rear bumper composition was rigid (n = 23), piston (n = 38), polymer (n
Of all 210 vehicles involved, eight struck and two striking vehicles were
towed from the scene. This low percentage of tow-away crashes indicated that
the crash set studied here would not have been captured in U.S. towaway crash
databases. For the 75 struck vehicles with IIHS geometric head restraint
was ‘good’.
recorded belt restraint usage for 68 of the struck occupants seeking medical
69 reported three-point restraint usage, one reported lap belt usage and one
reported no restraint use. Police reports noted the posted speed limit in 54 of
the crashes, with 33 crashes occurring in speed zones less than or equal to 56.3
11
CHAPTER 1 – REAL WORLD MINOR REAR CRASHES
emergency room (ER), urgent care facilities or medical office locations as seen
in Figure 1-3.
Male Female
50
45
40
35
Frequency
30
25
20
15
10
5
0
Office ER/Urgent Care
Figure 1-3. Medical treatment facility for first visit post-crash (n=113 struck vehicle occupants)
The time between the crash and first medical treatment visit was
available for 111 of the struck occupants. One male occupant made an initial
treatment visit to both the ER and a medical office on the same day. Any visit
occurring the same day as the crash was given a value of zero days. Figure 1-4
details the first visit medical treatment distribution. The consulting clinician for
the first treatment was available for 85 occupants and is shown in Figure 1-5
Male Female
20
Days to First Visit Post-Crash
18
16
14
12
n = 13 n = 28
10
8
6 n = 23
4 n = 48
2
0
Office ER/Urgent Care
12
CHAPTER 1 – REAL WORLD MINOR REAR CRASHES
Male Female
40
35
30
Frequency
25
20
15
10
5
0
M.D. D.O. D.C.
The unique ICD-9-CM and AIS diagnoses for the first medical treatment
visit and all visits are shown in Figures 1-6 and 1-7, respectively.
50%
Percent of Diagnoses
ip
en
FS
l
r
c
ca
ra
de
ci
ea
th
ac
om
a
vi
N
ul
x/
H
O
or
er
r/S
ho
ck
bd
Th
C
er
Ba
/S
ba
/A
w
Ex
st
m
Lo
he
Lu
er
C
pp
U
50%
Percent of Diagnoses
ip
en
FS
l
r
c
ca
ra
de
ci
ea
th
ac
om
a
vi
N
ul
x/
H
O
or
er
r/S
ho
ck
bd
Th
C
er
Ba
/S
ba
t/A
w
Ex
m
Lo
s
he
Lu
er
C
pp
U
13
CHAPTER 1 – REAL WORLD MINOR REAR CRASHES
post-crash with at least one degenerative spine condition pre-existing the crash
These conditions were broken into five categories for the cervical, thoracic and
lumbar/sacral spine: (a) disc bulging, protrusion and herniation, (b) spinal
and myelopathic symptoms and (e) and abnormal spinal structure such as
occupants, 52, nine and 29 occupants were diagnosed with at least one cervical,
crash during the five week treatment period. The distribution of AIS1 injuries
for these occupants is shown in Figure 1-9 as compared with the 46 non-
14
CHAPTER 1 – REAL WORLD MINOR REAR CRASHES
120
DDD/DJD
(d)
100
80
Frequency
Spurring Abnormal
60 Disc structure
(c)
(a) (e)
40
Stenosis
(b)
20
50%
Percent of Diagnoses
ip
en
l
d
ic
ca
ra
de
ea
ac
m
ac
vi
ul
x/
H
or
o
er
r /S
ho
bd
h
C
er
T
/S
ba
/A
w
Ex
st
m
Lo
he
Lu
er
C
pp
U
‘whiplash’ in addition to ‘other’ initial diagnoses group (n=53) and a group with
‘other’ initial diagnoses only (n=25). Occupants who made their first treatment
15
CHAPTER 1 – REAL WORLD MINOR REAR CRASHES
visit beyond one week post-crash (n=15) were excluded in this analysis. For
these groups, all unique AIS1 injury diagnoses from the five week treatment
period were compiled and are shown in Figure 1-10. Non-specific AIS1 injury
Percentage of Diagnoses
'Whiplash' (n=50 dx.)
60%
'Whiplash' + 'Other' (n=208 dx.)
50% 'Other' (n=87 dx.)
40%
30%
20%
10%
0%
ip
ic
l
er
en
al
ca
ea
ac
/H
r
ld
om
ac
vi
ou
Ex
or
er
r/S
bd
Th
Sh
C
er
ba
/A
x/
w
m
Lo
st
Lu
he
er
pp
C
U
Figure 1-10. AIS1 diagnoses comparing ‘whiplash’ groups during five week treatment period post-
crash
degenerative cervical spine conditions was examined. The null hypothesis was
head and cervical spine injury rate over the five week treatment period defined
the mean of the population. A paired, two-tailed, t-test comparison was made
based on a p-value less than 0.05 being significant for a difference in the
between these groups is shown in Figure 1-11. Based on the p-value of 0.98 for
16
CHAPTER 1 – REAL WORLD MINOR REAR CRASHES
3.0
p = 0.98
2.5
1.5
1.0
0.5
0.0
occupants). The Group D diagnosed AIS1 lumbar/sacral injury rate over the five
week treatment period defined the mean of this population. Again, a paired,
two-tailed, t-test comparison was made based on a p-value less than 0.05 being
occupants. The results of this comparison are shown in Figure 1-12. The null
hypothesis for this analysis was accepted based on the resulting p-value of 0.30.
17
CHAPTER 1 – REAL WORLD MINOR REAR CRASHES
2.5
p = 0.30
Group D
1.5 Group C (n = 28)
(n = 16)
1.0
0.5
0.0
Figure 1-12. AIS1 lumbar/sacral injury rate in degenerative (Group C) and non-lumbar/sacral
degenerative occupants (Group D)
DISCUSSION OF RESULTS
The 105 minor crashes involved totals of 151 and 126 occupants in the
severity and resulting AIS1 diagnoses in these crashes than commonly used
weeks of the crash with complaints attributed to the crash. None of the striking
occupants were diagnosed with complaints attributed to the crash. The crash
Nearly ninety percent of struck vehicle occupants made their first medical
treatment visit within one week post-crash. Female and male occupants made
their first visit to emergency room and urgent care facilities approximately twice
as often as visiting a medical office location, and visited M.D. or D.O. more often
than chiropractors. This was likely because these emergency facilities were
staffed by M.D. and D.O. as opposed to chiropractors. For those occupants first
visiting a chiropractor, women had a higher percentage of visits than men. For
18
CHAPTER 1 – REAL WORLD MINOR REAR CRASHES
the occupants first visiting medical office locations, the time until first visit was
highly variable but these occupants averaged over eight days until their first
The first medical treatment visits resulted in a total of 410 unique ICD-9-
CM and 256 unique AIS1 diagnoses attributed to the crash. For the five week
post-crash diagnosis period, 761 unique ICD-9-CM diagnoses and 427 unique
AIS diagnoses were reported. Of these AIS diagnoses, two AIS2 injuries (lumbar
fracture, meniscus tear) were reported. It was questionable whether the crash
forces were of sufficient magnitude and direction to have caused the diagnosed
AIS2 injuries. ICD-9-CM had more total diagnoses than the AIS system due to
diagnoses with the five week diagnosis period, ICD-9-CM and AIS spine
diagnoses decreased by 2.6% and 9.0% of the total diagnoses during the five
week period, respectively. Over this same time period, head, chest and
Reasons for the decreasing number of spine related diagnoses could have been
due to the resolution of transient spine complaints after the first week of
lumbar/sacral spine, thoracic spine and extremities exceeded that of head and
cervical spine ‘whiplash’ related diagnoses. Since the most at-risk body parts in
these minor rear crashes were the head and cervical spine, and based on the
majority of all struck occupants with complaints would have ‘whiplash’ and head
19
CHAPTER 1 – REAL WORLD MINOR REAR CRASHES
The diagnosis frequency for the thoracic spine, lumbar/sacral spine and
extremities has not been reported in previous minor real world rear crash
research, has not been produced in human volunteer testing at similar or more
extreme crash severities and did not correlate with accepted minor rear crash
cited frontal crashes as a source of ‘whiplash’ and head injury,31;78 none of the
diagnoses in the current study and prior studies may have been due to the
unbiased cohort is needed to fully verify or refute the ICD-9-CM and AIS
diagnosis distributions and lack of frontal crash ‘whiplash’ in the current study.
When comparing the current AIS1 injury diagnoses to prior large minor
rear crash studies, discrepancies arose. While the struck vehicle delta-V range
in the current study, less than or equal to 15 km/h, was of similar magnitude or
39;62;77;152
below prior minor rear crash studies , the number of front seat struck
vehicle occupants who claimed injury was very different. In the Krafft et al.
(2005) study, it was reported that 33% of 171 front seat occupants reported an
AIS1 injury complaint while in the Eis et al. (2005) study, 41% of 1,724 front
seat occupants reported AIS1 or greater injury. These injury rates are much
lower than the 110 of 137 (80%) front seat occupants claiming AIS1 injury in
the current study. Secondly, in the Hell et al. (2002), Krafft et al. (2005) and
comprised 100% of injured occupants; in the Eis et al. (2005) study neck
20
CHAPTER 1 – REAL WORLD MINOR REAR CRASHES
not correlate with any data presented in these prior studies. The reasons for
‘whiplash’ complaints in the current minor rear crash cohort is not fully
degenerative spine condition during the five week post-crash visit period,
were in the majority. Similar proportions were found between disc bulging,
spurring and abnormal spine structure. When AIS1 diagnoses for these
the percentage of lumbar/sacral and thoracic spine AIS1 injury diagnoses were
high compared with cervical diagnoses since the cervical spine was the most at-
One of the more interesting findings of this study was that when
condition occupants were not statistically significant for different injury rates to
the cervical and lumbar/sacral spine, respectively. For the cervical comparison,
difference in the AIS1 ‘whiplash’ injury rate. This result bears resemblance to
for a different lumbar/sacral AIS1 injury rate. This comparison indicated that
normal occupants tended toward increased AIS1 injury rates over degenerative
While the current study examined many facets of minor rear aligned
crashes and ‘whiplash’ injury diagnoses, the study had weaknesses. Since there
were no publicly available minor rear crash U.S. databases, and the forensic
authors were not able to select an unbiased sample cohort for comparison. A
second weakness is that the struck vehicle delta-V values obtained from the
minor crash reconstruction method had some amount of error. While it would
have been desirable to conduct crash tests with exemplar vehicles and
occupants for all 105 crashes in order to validate the reconstruction results, the
scope and cost of this type of testing simply was prohibitive for this study.
Though the crash reconstructions were not validated by crash testing, the minor
crash reconstruction methodology used here has been shown to be within 10%
of the actual struck vehicle delta-V in modern vehicle crash tests. Thus, the
mean error present in the struck vehicle delta-V reconstruction estimates would
have been on the order of 0.6 km/h. This error would have had minimal effect
The current results indicate that minor rear ‘whiplash’ crashes might
However, this finding does not agree with any similar publications regarding real
39;62;77;117;134
world minor crashes
22
CHAPTER 1 – REAL WORLD MINOR REAR CRASHES
LIMITATIONS
sample size was limited to the 105 crashes studied here. Due to this small
study sample size and biased data toward litigants or potential litigants, a true
only be truly quantified via vehicle crash testing. Hence, while delta-V values
to 9% of their true values, the actual error rate for this study is unknown.
were not studied here that have been universally acknowledged to contribute to
‘whiplash’ symptoms and diagnoses. This study did not examine psychological
and litigation, these variables were not studied here. These interactions likely
played some type of a role in this dataset, and while they may have been
underscore a need for a larger, controlled, and unbiased U.S. minor crash study
is needed to examine occupant diagnoses due to minor rear crashes for delta-V
database modeled after NASS or CIREN for minor rear crashes would be
23
CHAPTER 1 – REAL WORLD MINOR REAR CRASHES
corresponding crash dynamics when struck vehicle delta-V is less than or equal
to 15 km/h.
CONCLUSIONS
In the current study, while searching available crash databases and the
published literature, it was found that few data exist to study actual ‘whiplash’
and head injury complaints in real world U.S. minor rear aligned crashes at
delta-V less than or equal to 15 km/h. This required gathering real world
‘whiplash’, head and spine injury diagnoses. In this forensic crash data set,
From 105 real world minor U.S. crash reconstructions, the mean delta-V was
found to be 6.3 km/h and mean acceleration was 1.4 g. The majority of struck
threshold ranges for AIS1 injuries found in volunteer and real-world studies.
Even with these crashes being of minor severity, a majority of struck occupants,
113 out of 151, had complaints attributed to the crash that were diagnosed by a
M.D., D.O. or chiropractor within a five week treatment period post-crash. Many
spine and extremities, has never been found in any prior study investigating real
prospective U.S. cohort minor rear crash study and merits extremely cautious
Based on the injury exposure risk of the head and cervical spine due to
direct loading from the head restraint and inertial loading from the head-neck
24
CHAPTER 1 – REAL WORLD MINOR REAR CRASHES
region, and the findings of prior minor rear crash studies, the authors expected
diagnosed complaints to body locations other than the head and cervical spine
seen here has not been reported in real world studies performed in other
complaints attributed to the crash as ‘whiplash’ and head injuries has also been
spine injury in the present data set. It is a distinct possibility that the injury
complaint differences seen here when compared with prior studies could have
prospectively collect unbiased minor crash and occupant diagnosis data in the
complaints related to real world minor rear crash severity, the U.S. has yet to
follow suit.
25
CHAPTER 2 –MINOR REAR CRASH MADYMO SIMULATION
INTRODUCTION
development, validation and results from a minor crash simulation using the
MADYMO numerical simulation software. Real world minor rear crash test
simulated vehicle model. Hybrid III and BioRIDII crash test dummies were
placed in the vehicle, subjected to the minor rear crash acceleration pulse,
BACKGROUND
In Chapter 1 real world minor rear crashes, where the struck vehicle
a high frequency to regions other than the head and spine. To further
understand injury risk to the head, spine and these other regions in this
crashes are evaluated through the use of the computational crash model in
26
CHAPTER 2 –MINOR REAR CRASH MADYMO SIMULATION
used in motor vehicle crash safety design and analysis. Hence, development
and validation of a minor rear crash passenger vehicle MADYMO model based
on the real-world crashes studied in Chapter 1 allows for accurate and cost-
occupants in real world minor rear crashes without the need to undertake
rear crash MADYMO model using the mean of impact data similar to the
minor rear crash tests in Chapter 1 with an instrumented BioRIDII ATD, (2)
use this validated MADYMO model to observe and compare BioRIDII and
Hybrid III 50th percentile minor rear crash responses and (3) evaluate model
paid to head and spine loads, with comparison to existing injury thresholds
and criteria
The simulated minor rear crash acceleration pulse was taken from the
mean of six (6) late model passenger vehicles weighing less than 4,536 kg
involved in rear barrier minor crash tests with an instrumented BioRIDII ATD
in the driver position. These crash tests involved the same types of vehicles
27
CHAPTER 2 –MINOR REAR CRASH MADYMO SIMULATION
summary of the crash tests is shown in Table 2-1 and the specifics of these
Table 2-1. Passenger vehicle crash tests used as inputs to MADYMO minor rear crash
simulation
From these six minor rear crash pulses, the mean vehicle x-axis, or
longitudinal, acceleration pulse seen in Figure 2-1 was calculated for use in
the MADYMO simulation for the first 225 milliseconds following barrier
contact. Also shown is the mean longitudinal velocity change of the crash
tests over this pulse duration. The mean crash test delta-V was 12.1 km/h.
Figure 2-1. MADYMO longitudinal vehicle crash pulse and delta-V from minor rear crash tests
The mean rear crash pulse was applied to a generic late model vehicle
interior model in MADYMO V6.4 with a BioRIDII ATD seated in the driver and
28
CHAPTER 2 –MINOR REAR CRASH MADYMO SIMULATION
passenger seats as seen in Figure 2-2. Utilizing two BioRIDII ATD allowed
was positioned five millimeters above the seat cushion and in front of the
acceleration field was applied to the BioRIDII for the first 500 milliseconds of
this time period to fully settle any minor contact loading discontinuities due
29
CHAPTER 2 –MINOR REAR CRASH MADYMO SIMULATION
to initial position. The crash test pulse acceleration field was applied to the
vehicle in the forward direction during the next 225 milliseconds. Because
during the forward rebound phase, after the vehicle and BioRIDII achieved a
common velocity, were small or insignificant after the 225 millisecond rear
cushion and head restraint rigidity, seat back and head restraint joint
stiffness and seat back and head restraint orientation, and less vital
the BioRIDII driver and passenger. The belts were subsequently removed
outputs, involving pelvis, thorax and head x-axis linear accelerations and
head y-axis angular acceleration traces fell generally within corridors defined
by +/- one standard deviation of the real world crash test BioRIDII mean
and z-axis forces, y-axis bending moment, pelvis and head z-axis
crash neck injury criterion, Nij, potential rear crash neck injury criterion #1,
30
CHAPTER 2 –MINOR REAR CRASH MADYMO SIMULATION
Nkm, and potential rear crash neck injury criterion #2, NIC, were compared
with the crash test maxima and provided further confirmation that the model
rear crash model was validated for the 12.1 km/h delta-V minor rear crashes,
a 50th percentile HybridIII ATD replaced the BioRIDII in the passenger seat
and the simulation was re-run. The HybridIII initial position was similar to
the BioRIDII initial position. This ATD response comparison was performed
since the HybridIII is a more widely used ATD in vehicle crash testing but has
been shown to have a different response than the BioRIDII in rear crashes.
(3) REDUCED INPUT PULSE - Finally, the 12.1 km/h delta-V crash test
pulse was reduced by 40% and applied to the model for the BioRIDII to
delta-V minor rear crash. This delta-V was chosen since it was of similar
RESULTS
time response, were compared with BioRIDII crash test corridors for the 225
Figures 2-3 through 2-6. The shaded areas on each figure define the
corridors surrounding +/- one standard deviation of the mean crash test
31
CHAPTER 2 –MINOR REAR CRASH MADYMO SIMULATION
generally fell within the crash test corridors and showed good correlation to
32
CHAPTER 2 –MINOR REAR CRASH MADYMO SIMULATION
The peak simulation upper neck x- and z-axis forces, y-axis bending
moment, pelvis and head z-axis accelerations and head and thorax y-axis
angular velocities were compared with the crash test ranges in Table 2-2.
The crash test ranges encompass the average maxima, plus or minus one
33
CHAPTER 2 –MINOR REAR CRASH MADYMO SIMULATION
generally fell within the crash test ranges, with no major differences that had
Based on the results of Table 2-2, the upper neck forces and moments
were analyzed for potential injury risk. When comparing crash test and
simulation upper neck forces and moments to accepted tolerance values, the
upper neck shear force was at least 95% below the threshold, the axial
tension force was at least 80% below the threshold and the extension
respectively.144
To further compare the crash tests and simulation outputs, neck injury
risk criteria, Nij, Nkm and NIC, were calculated as seen in Table 2-3 and
that the Nij has only been approved for use with the HybridIII ATD in frontal
crash tests100 and therefore the values shown here may not be especially
meaningful. The Nkm130 and NIC10 are both proposed minor rear crash injury
metrics that can be calculated using outputs from the HybridIII or BioRIDII.
34
CHAPTER 2 –MINOR REAR CRASH MADYMO SIMULATION
These results showed that the MADYMO simulation over predicted neck
injury risk when using these three criteria as compared with the crash tests.
Interestingly in the simulation, the Nij and Nkm were well below their
injurious value.
Finally, head injury risk was investigated for the validated model and
compared with the crash test results. Head injury criterion (HIC), cumulative
due to vacuum pressure (DDM) and relative brain motion damage measure
(RMDM) were calculated for both the mean crash test results and simulation
results via SIMon v3.051 Beta (SIMulated Injury Monitor, National Highway
All of the head injury results show risk well below threshold or
completely absent. But for the CSDM, DDM and ASDH results, SIMon
35
CHAPTER 2 –MINOR REAR CRASH MADYMO SIMULATION
currently allows calculation of these head injury metrics only for the HybridIII
ATD family, so the calculations for the BioRIDII ATD, while appearing to be
injury risk might seem trivial and unnecessary considering the minor nature
note that head injury risk in both the crash tests and MADYMO simulation
BioRIDII model, a HybridIII 50th ATD was positioned in the passenger seat
alongside the BioRIDII driver. BioRIDII and HybridIII pelvis, head and
millisecond crash pulse time period and are shown in Figures 2-7, 2-8 and 2-
BioRIDII but then started deviating near the peak value. The maximum
HybridIII pelvis x-acceleration was 6.6g as compared with 5.4g for the
acceleration for the HybridIII and BioRIDII it can be seen in Figure 8 that the
HybridIII head acceleration onset was faster by six milliseconds and followed
approximately the same slope to the maximum value. The HybridIII peak
was 14.8g versus 11.4g for the BioRIDII. During deceleration, the HybridIII
declined at approximately the same rate as the BioRIDII. The thorax x-axis
36
CHAPTER 2 –MINOR REAR CRASH MADYMO SIMULATION
MADYMO responses. The HybridIII had a similar peak value of 4.5g versus
4.7g for the BioRIDII but these peaks were separated by fifty milliseconds.
37
CHAPTER 2 –MINOR REAR CRASH MADYMO SIMULATION
The differences in the BioRIDII and HybridIII ATD crash test response
spine anthropometry, since the BioRIDII has a full spine with articulating
segments and the HybridIII has simplified spine segments only between the
38
CHAPTER 2 –MINOR REAR CRASH MADYMO SIMULATION
(3) REDUCED INPUT PULSE – The crash test pulse was reduced by
40% of its original value for the 225 millisecond crash pulse time period in
approximating the mean crash from the Chapter 1 real world data set.
BioRIDII kinetics were quantified in response to this reduced pulse and are
MADYMO output.
Table 2-5. BioRIDII output for pulse scaled by 40% and change from original response
When viewing the reduced pulse simulation, the BioRIDII was seen to
have reduced motion which directly resulted in lower kinetic and kinematic
outputs for all parameters. Injury criteria were not calculated for the
reduced pulse since they would have all been reduced from their already
minimal levels. The average percent change for the twelve parameters
presented in Table 2-5 was -51%, with the largest reductions in important
39
CHAPTER 2 –MINOR REAR CRASH MADYMO SIMULATION
DISCUSSION
ATD has been shown to be valid when compared with six crash tests
involving passenger vehicles. While the model did not exactly fall within all
parameters of the crash tests, the model closely estimated the crash test
were several times larger than those in the vertical (z-axis) direction.
The BioRIDII MADYMO response could have been affected by the lack
longitudinal (x-axis) pulse. In the rear crash tests, the vehicle sustained
both longitudinal and vertical acceleration. And while it was obvious that
likely provided results that compared even more favorably with the real world
tests with similar delta-V, and by creating BioRIDII response corridors for
40
CHAPTER 2 –MINOR REAR CRASH MADYMO SIMULATION
these tests, the current study developed a generic MADYMO rear crash model
differences would exist between individual seats and the MADYMO seat
a particular crash pulse, rather than any one of the hundreds of different
BioRIDII responses, head and neck injury risk proved to be well below any
accepted, or proposed, tolerance values. Except for the case of the proposed
NIC rear crash neck injury risk, which is still a topic of debate, the injury risk
values for the minor rear crash tests and MADYMO model were expected to
and NIC injury prediction accuracy have been questioned45 and more
BioRIDII response to the same input crash pulse. While this finding was not
itself very surprising, what was interesting to note was that the HybridIII
anthropometry between the two ATDs would account for much of this loading
rear crash simulations that utilize the HybridIII ATD overestimate occupant
loading when compared with a more biofidelic ATD such as the BioRIDII.
41
CHAPTER 2 –MINOR REAR CRASH MADYMO SIMULATION
model predicted reduced accelerations, velocities, forces and moments for all
lessened head restraint contact were the main reasons for reductions in
head, upper neck and thorax loading parameters. Head and neck injury risk
was not calculated for the reduced pulse simulations as these risks would
have been trivial since they had to be lower than the aforementioned
LIMITATIONS
This study had many limitations. The main limitation was a result of
stiffness, floor mat friction, etc. the solution presented here is validated for
only one set of these parameters. There would likely be many other possible
BioRIDII and Hybrid III kinematics. But since the occupant loading was
major deficit to the current generalized model is that the validation stability
was not quantified. Thus, it was unknown exactly how stable the solution
fine-tuning the head-restraint timing and had little influence on pelvis, thorax
42
CHAPTER 2 –MINOR REAR CRASH MADYMO SIMULATION
performance alongside the BioRIDII. Because time and resources did not
permit for full-scale vehicle crash testing, the Hybrid III results should be
CONCLUSIONS
occupant kinetics and kinematics in these minor crashes were quantified via
BioRIDII ATD. This numerical simulation approach was used over cost-
prohibitive crash testing and results validated based on response from six (6)
previously performed minor rear crash tests with mean 12.1 km/h delta-V.
In the crash tests and validated MADYMO model, head and neck
loading were minimal and sufficient to cause only minimal AIS1 level injury
loads due to a stiffer spine structure. Thus, minor rear crash MADYMO
occupant head and spine loading when compared with the more biofidelic
BioRIDII for the same crash pulse. When the crash pulse was reduced by
40% to represent a common real world minor rear crash delta-V of 7.3 km/h,
43
CHAPTER 2 –MINOR REAR CRASH MADYMO SIMULATION
BioRIDII loading was reduced by 11% to 81% for all twelve parameters
studied. This resulted in reduced head and neck injury risk from the already
minimal risk found in the BioRIDII response to the original pulse. This
head and spine kinetics and kinematics for passenger vehicle minor rear
testing.
44
CHAPTER 3 – NECK INJURY RISK CLASSIFICATION BASED ON ORIGINAL Nij DATASETS
INTRODUCTION
significant cause of economic and healthcare costs in the United States. And
the first neck injury criterion, also known as Nij. While the Nij serves only to
starting point from which to define the injury risk boundary conditions to the
Chapters 1 and 2 for low energy rear crashes are expanded upon here with
more theoretically-based graphical and empirical neck injury risk criteria for
BACKGROUND
In an effort to reduce fatal spine and head injuries due to high energy
airbag loading in frontal crashes, the upper spine injury criterion, also known
Vehicle Safety Standard 208 crash testing in 2000.100 Nusholtz et al. (2003)
45
CHAPTER 3 – NECK INJURY RISK CLASSIFICATION BASED ON ORIGINAL Nij DATASETS
predictor than Nij and posited that Nij injury prediction inaccuracies arose
from the use of prismatic beam bending theory, which treated the cervical
relationships have been validated when analyzing tibia injury risk,100 this
composition of joints, bones, ligaments, tendons and muscles has not been
validated.113 And several authors have shown that the Hybrid III ATD neck
during impact testing,1;40;74;113 the use of which in the Nij can lead to
Taking cues from the prior work of Nusholtz et al.,113 and recognizing
force, tensile force loading rate, tensile force impulse, tensile force ‘binning’
and AIS2 injury severity. This method of re-examining injury data to derive
new risk measures has precedent, as it has been done previously for head,
relevant, accurate and useful spine injury risk metrics existed within the
original piglet-ATD cervical spine tensile force data. Spine injury prediction
46
CHAPTER 3 – NECK INJURY RISK CLASSIFICATION BASED ON ORIGINAL Nij DATASETS
tests. The Nij,100 as well as peak axial spine tension113 from the paired piglet-
ATD tests, served as benchmarks for these new spine injury metrics.
NIJ DEVELOPMENT
The Nij was developed from a series of fifty-eight (58) paired airbag
the lower, middle or upper thorax. The airbag was activated in response to
crash test sled pulses ranging from 4.2g to 14.9g and 28.6 km/h to 57.0
time was varied during the testing. Piglet injury was rated according to
AIS1980.2 While there were varying injury severities observed, ranging from
AIS0 “no injury” to AIS5 “critical injury”, for comparison the piglets were
divided into a ‘no/low injury’ group (AIS<3) and an ‘injured’ group (AIS3+).
Following the piglet tests, the representative three-year-old child ATD was
impacted to the anterior cervical spine, jaw and upper thorax under similar
initial seating, crash pulse and airbag firing conditions. ATD upper cervical
spine axial force, FZ, and sagittal plane bending moment about the occipital
condyles, MY, were two of the measured outputs. Finally, piglet injury scores
were compared to forces and moments measured from each paired test.
combination was developed to relate upper cervical spine injury risk and ATD
47
CHAPTER 3 – NECK INJURY RISK CLASSIFICATION BASED ON ORIGINAL Nij DATASETS
FZ M
N ij Y 1.0 Eq. 3-1
F Zc MYc
STUDY METHODS
et al. (1982) and Prasad and Daniel (1984) studies, we used the thirty-four
tests with continuous axial cervical spine force data with minimum 100
millisecond duration. The Nij results for these ‘no/low injury’ (AIS<3, n=20)
and ‘injured’ (AIS3+, n=14) tests used in this study are shown in Figure 3-1.
The kite-shaped boundary defines theoretical critical loading limits for Nij
over the duration of the test for the three-year-old ATD. The axial cervical
spine force-time histories from the same ‘no/low injury’ and ‘injured’ tests
Figure 3-1. Nij v. time for ‘no/low injury’ (A) and ‘injured’ (B) piglet-ATD tests
48
CHAPTER 3 – NECK INJURY RISK CLASSIFICATION BASED ON ORIGINAL Nij DATASETS
Figure 3-2. ATD axial force (Fz) for paired tests with ‘no/low injury’ (A) and ‘injured’ (B)
piglets
From these 34 piglet-ATD axial spine force data sets, the following
For each test, the axial tensile force was normalized against the
mechanisms in the piglet tests were all tensile in nature.94;127 The maximum
normalized tension force for each test was calculated as shown in Eq. 3-2.
F
100
Z
Maximum Normalized Tensile Force max Eq. 3-2
1130N 0
factor related to spine injury risk because the spine is comprised of both
viscous and elastic constituents.53 This hypothesis was reinforced by the fact
that loading rate has been used as an injury prediction tool in cadaveric
49
CHAPTER 3 – NECK INJURY RISK CLASSIFICATION BASED ON ORIGINAL Nij DATASETS
impact testing.129 Thus, for each paired piglet-ATD test the maximum tensile
loading rate over the first 100 milliseconds was calculated using Eq. 3-3.
dF 100
Maximum Tensile Loading Rate max Z Eq. 3-3
dt 0
reasonable risk measure since crash-related injury has been long known to
depend upon the force accumulated over the impact duration and the piglet
impulse accumulation over longer time durations that has been previously
cumulative tensile impulse was analyzed over the 100 millisecond time
100
Cumulative Tensile Impulse
F
0
Z dt Eq. 3-4
time spent at each loading level. The normalized tensile force data were
each test during the 100 millisecond time period. The time spent in each
discretized bin increment was summed and the percent time spent in each
bin for the 100 millisecond time period was computed. An example of force
50
CHAPTER 3 – NECK INJURY RISK CLASSIFICATION BASED ON ORIGINAL Nij DATASETS
binning can be seen in Figure 3-3, where the time spent in each bin is
represented by the portion of the pulse dwelling in each bin. In our study,
Bin0.9, with the total percent time spent above the bin with normalized force
equal to 0.9, was identified as the bin level showing the best delineation
Figure 3-3. Binning method applied to normalized force data from Piglet Test #6368
injury severity and threat-to-life. Table 3-1 illustrates the complete AIS
scoring system. Considering injury severity via AIS allowed for a dramatic
51
CHAPTER 3 – NECK INJURY RISK CLASSIFICATION BASED ON ORIGINAL Nij DATASETS
visual depiction of the various risk measures. When the simplified binary
were used to examine the piglet-ATD data and determine which method(s)
best separated ‘injured’ and ‘no low injury’ tests based on the
AIS while the statistical methods grouped injury according to the ‘no/low
injury’ (AIS<3) and ‘injured’ (AIS3+) and combinations of the injury metric
regression plots to compare the doubly censored tests and injury risk
52
CHAPTER 3 – NECK INJURY RISK CLASSIFICATION BASED ON ORIGINAL Nij DATASETS
The ‘no/low injury’ group D-statistic was 0.2181 and the critical value at
=0.05 was 0.2941. The ‘injured’ group D-statistic was 0.2054 and the
critical value was 0.3489. Based on the normality test, both groups were
treated as normal distributions. For each risk parameter, the binary logistic
regression fit was calculated based on binary ‘injured’ (AIS3+) and ‘no/low
severity. Injury risk equal to 40% was used as a threshold for all parameters
because it resided between the injury risk values used for head injury and
neck injury analyses.40;63;93 Figure 3-4 shows the regression curves for the
53
CHAPTER 3 – NECK INJURY RISK CLASSIFICATION BASED ON ORIGINAL Nij DATASETS
Figure 3-4. Logistic regression curves for four parameters: (a) Maximum Normalized Load,
(b) Maximum Loading Rate, (c) Cumulative Impulse, (d) %Time Above Bin=0.9. Intercepts
drawn at 40% injury risk
in a two-dimensional space. Data were plotted for the six (6) combinations
of two risk parameters with the marker radius defined by the AIS magnitude
defined the marker radius, injury classification accuracy was solely based on
54
CHAPTER 3 – NECK INJURY RISK CLASSIFICATION BASED ON ORIGINAL Nij DATASETS
Figure 3-5. Three parameter bubble scatterplot for six (6) injury parameter combinations:
(a) Maximum Loading Rate vs. Maximum Normalized Force, (b) Cumulative Impulse vs.
Maximum Normalized Force, (c) Cumulative Impulse vs. Loading Rate, (d) Maximum Loading
Rate
radius of the resulting spherical markers but again, the injury classification
accuracy depended on marker center location, and not marker radius. The
55
CHAPTER 3 – NECK INJURY RISK CLASSIFICATION BASED ON ORIGINAL Nij DATASETS
Figure 3-6. Four parameter spheroidal scatterplot of four (4) injury parameter combinations:
(a) Maximum Normalized Force vs. %Bin Time vs. Cumulative Impulse, (b) Maximum
Normalized Force vs. %Bin Time vs. Maximum Loading Rate, (c) Maximum Normalized Force
vs. Cumulative Impulse vs. Maximum Loading Rate, (d) %Bin Time vs. Cumulative Impulse
vs. Maximum Force Loading Rate. Volume intercepts drawn at 40% injury risk, sphere radius
equals AIS injury severity.
RESULTS
The results for each injury risk metric are summarized in Table 3-2
and compared with Nij and peak axial tension.113 Nij and peak axial tension
correctly classified ‘no/low injury’ tests at 80% and 88%, respectively. For
‘injured’ tests, Nij and peak axial tension classification accuracy was 65% and
80%, respectively.
56
CHAPTER 3 – NECK INJURY RISK CLASSIFICATION BASED ON ORIGINAL Nij DATASETS
% Correctly %
Injury Risk classified Correctly
Risk Measure
Method 'no/low classified
injury' 'injured'
Logistic Regression
As seen in Figure 3-4, the regression curves for the four parameters
show discrimination between ‘injured’ and ‘no/low injury’ data at a 40% risk
of injury. As seen in Table 3-2, the best ‘no/low injury’ accuracy was 100%
for both normalized force and binning. The highest ‘injured’ prediction
accuracy was 95% for normalized force, binning and cumulative impulse.
57
CHAPTER 3 – NECK INJURY RISK CLASSIFICATION BASED ON ORIGINAL Nij DATASETS
The AIS marker radius in each plot appeared fairly chaotic in nature, with no
clear cut trend for AIS severity and parameter magnitude other than
40% Intercepts
Normalized Force (N/N) 1.10
Maximum Loading Rate (N/s) 1.10E+05
Cumulative Impulse (N-s) 30.1
%Bin Time above 0.9 9.85%
Table 3-3. 40% injury risk intercepts for the four parameters based on regression curves
Bubble Scatterplot
40% failure envelope is shown for the bubble scatterplots in Figure 3-5.
Based on the results in Table 3-2 and considering the 40% failure envelopes,
the most accurate correct classification rate was 95% for ‘no/low injury’ tests
rate vs. binning, correctly classified 100% of tests, respectively. Again, the
magnitude.
58
CHAPTER 3 – NECK INJURY RISK CLASSIFICATION BASED ON ORIGINAL Nij DATASETS
Spheroidal Scatterplot
combination classified ‘no/low injury’ tests based on the theoretical 3-D 40%
injury risk volume at a 95% accuracy rate. For ‘injured’ tests, all four
mentioned in the regression and bubble plots, the AIS marker radius
DISCUSSION
between AIS<3 ‘no/low injury’ tests and AIS3+ ‘injured’ tests at rates as
injury risk threshold, the four risk metrics correctly classified ‘no/low injury’
tests at 90% to 95% and ‘injured’ tests at 79% to 100%, respectively. The
bubble scatterplots had classification accuracy between 90% and 95% for
‘no/low injury’ tests and for the ‘injured’ tests, classification accuracy ranged
accuracy of 90% to 95% for ‘no/low injury’ tests and 100% for all ‘injured’
tests for marker centroids residing within the 3-D failure volume. Normalized
force, cumulative impulse and binning classified ‘no/low injury’ and ‘injured’
tests at the highest rates while maximum loading rate had the lowest
59
CHAPTER 3 – NECK INJURY RISK CLASSIFICATION BASED ON ORIGINAL Nij DATASETS
rate higher than the 65% rate for Nij and thirteen of fourteen metrics
classified ‘injured’ tests at a rate higher than 80% for maximum axial
tension. For ‘no/low injury’ tests, all risk measures classified injury at a
higher rate than the 80% seen in Nij or the 88% rate based on axial force
LIMITATIONS
subjects, this study had several limitations. The main limitation was that the
paired piglet-ATD tests may not have been representative of the human
injury response to frontal airbag deployment. Since the Nij frontal impact
the piglet surrogate, young human cadavers simply were not available to
conduct this type of testing. Nonetheless, the piglets had an internal and
external anthropometry vastly different from humans and in these tests the
Numerous cervical spine injury impact tests have been done with human
cadavers since the Nij piglet-ATD frontal airbag tests in the early
require further rigorous validation with these or future human cadaver tests.
60
CHAPTER 3 – NECK INJURY RISK CLASSIFICATION BASED ON ORIGINAL Nij DATASETS
aliasing effects were not quantifiable. The 40% risk threshold was chosen
range of similar risk thresholds for HIC and Nij, the exact threshold to use
limitation in the current study was that in the paired piglet-ATD tests the
were vastly different. In test series where ATD impacts were repeated after
a single piglet impact of the same magnitude, the peak axial tensile force
N-m, respectively.94;127 This intratest variability may have been one major
additional limitation was that the test sample sizes used here, n=14 ‘injured’
and n=20 ‘no/low injury’, were fairly small. It would be desirable to have a
larger sample size for both injury classifications in order to fully validate the
novel injury measures proposed here. A final limitation was that while
sagittal spine moment was not considered in the spine injury criteria
moment for spine injury risk. While Nusholtz et al. (2003) have proposed
that the entire cervical spine cannot be treated as a prismatic beam as was
done with Nij, bending moment has been shown to be a reliable indicator of
intervertebral segment injury and thus may serve a crucial role in segmental
61
CHAPTER 3 – NECK INJURY RISK CLASSIFICATION BASED ON ORIGINAL Nij DATASETS
Deserving special note in our study is that the piglet-ATD Nij tests94;127
involved head-thorax airbag impacts that resulted many severe piglet brain
been possible that spine injury risk in these tests was a function of head
injury risk and vice versa. Particularly interesting is that piglets sustained
brain injuries due to both linear and rotational motion but never sustained
essential to predict injury. And the effect of combined head linear and
the Nij datasets. Therefore, while we did not specifically consider a direct
linkage between head injury risk and spine injury risk in this study, the
62
CHAPTER 3 – NECK INJURY RISK CLASSIFICATION BASED ON ORIGINAL Nij DATASETS
CONCLUSIONS
public, accounting for 37% of the most critical motor vehicle injuries seen in
US trauma wards. And while the frontal crash Nij has been part of US crash
testing regulations since 2000, as first posited by Nusholtz et al. (2003), its
currently using Nij as a basis for injury prediction in rear automobile crashes,
frontal train crashes, lateral airplane fuselage impacts, boating and military
prediction accuracy in the Nij, these other risk metrics may suffer in their
ability to discriminate between safe and unsafe loading levels in the human
spine.
To this end, a total of fourteen (14) novel spine injury risk graphical
paired piglet-ATD Nij test database and evaluated against Nij and peak
tension force prediction accuracy. The risk metrics were based on cervical
normalized force, loading rate, impulse, force ‘binning’ and injury severity.
The injury risk metrics were found to correctly classify ‘injured’ tests up to
100% accuracy versus 68% for the Nij, and uninjured subjects up to 95%
accuracy versus 78% for the Nij, respectively. These new and accurate spine
injury risk metrics may prove useful in the future development of spine injury
criteria for non-frontal crash vectors where the predominant failure mode is
63
CHAPTER 3 – NECK INJURY RISK CLASSIFICATION BASED ON ORIGINAL Nij DATASETS
64
CHAPTER 4 – ATD HEAD IMPACTS AND HEAD-NECK DYNAMICS
INTRODUCTION
In this Chapter, head and spine injury analysis transitions from low
Chapters 2 and 3, in vitro study of high energy head and spine injury is best
done with anthropomorphic surrogates. But the influence of the Hybrid III
energy athletic impacts has not been quantified. Therefore, in this Chapter
dummy in order to quantify its frontal, oblique and lateral response to direct
experiments lays the foundation for the analyses of padded and unpadded
BACKGROUND
pose threats to athletes engaged in contact sports. In this study the Hybrid
III 50th percentile male anthropomorphic test device (ATD) was used to
might occur in concussive athletic impacts. The Hybrid III ATD was
Standard (FMVSS) 208 frontal crash testing.100 The ATD head-cervical spine
kinetics and kinematic response was developed and validated in the sagittal
65
CHAPTER 4 – ATD HEAD IMPACTS AND HEAD-NECK DYNAMICS
While the Hybrid III ATD was originally validated and is primarily used
in FMVSS 208 crash testing, it has also been widely used in athletic head
head-cervical spine biofidelity was not specifically studied and still remains
one potential confounding factor that may influence athletic head impact
particular interest because of the potential that in vitro ATD head impact
dynamics during concussive impacts. Furthermore, the Hybrid III ATD head-
cervical spine was validated in the sagittal plane without direct head
spine impacts,1;74 and the hollow ATD skull and absence of a deformable
spine complex. Thus, in this study the ATD head-cervical spine response to
direct head impact loading in the frontal, oblique and lateral directions in
METHODS
A Hybrid III 50th Percentile ATD (Denton ATD, Rochester, MI) was
66
CHAPTER 4 – ATD HEAD IMPACTS AND HEAD-NECK DYNAMICS
y-axis shear forces by the 0.01778m moment arm. All data were collected at
angular velocity signal after determining proper signal content and pre-
directions and energy levels tested, individual x-, y- and z-axis components
comprise voluminous data sets such that only the resultant dynamic
A 3.6kg steel sphere impact pendulum was used in all impacts. This
mass was chosen to deliver similar impacts as the effective mass involved in
padded boxing punches to the head.168 The mass was hung from a 6.4mm
steel braided cable that was free to swing from a ceiling-mounted carabineer
above the ATD head. The ATD was seated in a chair and secured with tie-
down straps. Prior to testing, the lower extremities were removed. In order
to minimize inertial responses below the cervical spine, the test stand was
weighed down with approximately 3500N of sandbags. For each impact trial,
the pendulum was released from the desired swing height by cutting a nylon
string attached to the steel sphere. The swing heights were selected to be
67
CHAPTER 4 – ATD HEAD IMPACTS AND HEAD-NECK DYNAMICS
risk of concussive traumatic brain injury and/or cervical spine injury. In this
study, a total of fifty-four (54) head impact pendulum trials were performed,
in sets of six (6) trials for each impact direction (frontal, oblique, lateral) and
impact energy level (lowest, medium, highest) as shown in Figure 4-1 and
Table 4-1. Medium and highest energy impacts were limited by the available
swing height within the testing space. The frontal and oblique impacts
RESULTS
The dynamic results are plotted against impact energy and are shown
in Figures 4-2 through 4-10 and Table 4-2, respectively. The same dynamic
data were also compared with impact momentum and showed the same
trends as the comparison with impact energy. Since impact energy was a
68
CHAPTER 4 – ATD HEAD IMPACTS AND HEAD-NECK DYNAMICS
Figure 4-2. Resultant ATD head center of gravity (cg) linear acceleration for each set of six
(6) impact trials in the three impact directions at highest impact energy
69
CHAPTER 4 – ATD HEAD IMPACTS AND HEAD-NECK DYNAMICS
Figure 4-3. Resultant ATD occipital force for each set of six (6) impact trials in the
three impact directions at highest impact energy
component-based plots were also produced. Figures 4-4 and 4-5 show
representative plots of x-, y- and z-axis linear acceleration and occipital force
for the mean values from the high energy impact sets.
70
CHAPTER 4 – ATD HEAD IMPACTS AND HEAD-NECK DYNAMICS
Figure 4-4. Mean ATD head center of gravity (cg) linear acceleration components from each
set of six (6) impact trials in the three impact directions at highest impact energy
71
CHAPTER 4 – ATD HEAD IMPACTS AND HEAD-NECK DYNAMICS
Figure 4-5. Mean ATD occipital force components from each set of six (6) impact trials
in the three impact directions at highest impact energy
impacts for each impact energy and direction was quantified and is displayed
in Table 4-2. This was done to ensure each set of six impact trials was
repeatable and that using the mean value for each dynamic parameter was
justified. The COV was calculated by dividing the standard deviation from
each trial set of six impacts by the trial set mean value. The highest mean
COV was seen in occipital force with 7.7% variation among all trials and all
angular acceleration had the lowest mean COV, 4.6% and 4.7%,
respectively. Because the COV for all dynamic parameters was low, the trials
72
CHAPTER 4 – ATD HEAD IMPACTS AND HEAD-NECK DYNAMICS
were determined to be very repeatable and use of the mean value from the
Occipital Force 8.3 5.7 5.9 18.4 7.7 8.3 5.0 3.3 6.4
Table 4-2. Coefficient of variation among all impact trial sets (Front = Frontal, Obl = Oblique,
Lat = Lateral)
acceleration generally trended linearly upward, but with three impact energy
levels plotted for each impact location, any nonlinear response was not
73
CHAPTER 4 – ATD HEAD IMPACTS AND HEAD-NECK DYNAMICS
150
100
50
0
0 10 20 30 40 50 60 70 80 90 100
Energy(J)
Figure 4-6. Resultant ATD head center of gravity (cg) linear acceleration for the three impact
directions compared with impact energy [NOTE: error bars show +/- one standard deviation]
For the resultant head angular velocity shown in Figure 4-7, at the
lowest energy impact level, all three directions produced similar resultant
energy level. At the highest energy impact level, frontal impacts had the
respectively.
74
CHAPTER 4 – ATD HEAD IMPACTS AND HEAD-NECK DYNAMICS
20
15
10
0
0 10 20 30 40 50 60 70 80 90 100
Energy(J)
similar trend to the angular velocity results. At the lowest energy level, all
crossover was seen in the medium impact energy level as lateral magnitude
dropped below the oblique magnitude. At the highest impact energy level,
frontal angular acceleration was the largest, followed by oblique and lateral
impacts.
75
CHAPTER 4 – ATD HEAD IMPACTS AND HEAD-NECK DYNAMICS
7000
6000
5000
4000
3000
2000
1000
0
0 10 20 30 40 50 60 70 80 90 100
Energy(J)
In Figure 4-9, the resultant occipital moment had a clear trend for all
three impact energy levels studied. The frontal impact moment was always
76
CHAPTER 4 – ATD HEAD IMPACTS AND HEAD-NECK DYNAMICS
100
Moment (N-m)
50
0
0 10 20 30 40 50 60 70 80 90 100
Energy(J)
Finally, in Figure 4-10, the resultant occipital force was compared with
impact energy. The frontal impacts always had the highest occipital force
across all impact energy levels. The oblique and lateral impacts had very
similar results with the largest difference between these two directions
77
CHAPTER 4 – ATD HEAD IMPACTS AND HEAD-NECK DYNAMICS
Frontal
2500
Oblique
Lateral
2000
Force (N)
1500
1000
500
0
0 10 20 30 40 50 60 70 80 90 100
Energy(J)
DISCUSSION
A key finding in this Chapter was that the data from the ATD impact
tests showed that impact direction played a heavy role in resulting ATD head
magnitudes, the directional sensitivity to impact was lost. Hence, while data
showed, for example in Figure 4-7, that lateral and oblique medium energy
loading the head and neck were very different. This finding has implications
beyond ATD testing, and could be useful in analyzing head and neck
78
CHAPTER 4 – ATD HEAD IMPACTS AND HEAD-NECK DYNAMICS
magnitude as the only measure of injury risk as is done with HIC or GSI
The results also indicated that ATD head-cervical spine stiffness played
resultant ATD cg linear acceleration was always highest for lateral impacts.
This might have been due to the lateral impact location, which was closer to
the head cg as opposed the oblique or frontal impacts. This response may
also be attributed to the ATD head-cervical spine complex being stiffer in the
lateral direction, causing less head motion during impact and allowing for the
linear acceleration to increase. The resultant ATD head angular velocity and
angular acceleration results were very similar. At the lowest impact energy
related to the ATD head-cervical spine being stiffer in the lateral direction
investigating the resultant occipital moment results, it was likely that ATD
frontal and oblique impacts had a faster increase in occipital moment than
the lateral impacts. In the lateral direction, the smaller bending moment
79
CHAPTER 4 – ATD HEAD IMPACTS AND HEAD-NECK DYNAMICS
compared with the oblique/frontal impacts might have indicated that the
spine sustains more overall bending as opposed to bending mainly about the
the frontal impacts always had the highest resultant force. The oblique and
lateral impacts had similar behavior at the lowest energy while oblique
impacts had a larger force magnitude at the highest energy impact level.
While no prior study known had specifically investigated Hybrid III ATD
compression, some comparisons were made with prior human and ATD
80
CHAPTER 4 – ATD HEAD IMPACTS AND HEAD-NECK DYNAMICS
extension and torsional motion, respectively. The ATD lateral rotation axis in
this study was the stiffest, which tended to mimic the lateral motion
tolerance finding, and the ATD had a stiffer response to oblique loading
rather than extension loading, which differed from cadaveric and volunteer
data. However, it was important to note that the Hybrid III ATD head-
cervical spine was validated solely in the sagittal plane for flexion-extension
axial torsion, extension and lateral bending, respectively, with the lateral
cervical bending moment was approximately 2-3 times higher in the oblique
volunteer inertial sled tests showed that lateral tests had lower cervical
bending moments and total rotation when compared with oblique tests.170
Dynamic human volunteer direct bending tests showed the tolerable lateral
cervical spine force to be almost twice that of the tolerable oblique force.121
frontal bone had occipital accelerations 50% higher than those cadavers
impacting the temporal region.55 All of these studies, with the exception of
81
CHAPTER 4 – ATD HEAD IMPACTS AND HEAD-NECK DYNAMICS
respectively. These findings were mirrored in the ATD impact data based on
occipital bending moment. When examining the Got et al. (1978) study, the
cadaveric head acceleration behavior was reversed from these results with
the ATD in that in this study frontal head acceleration was approximately
50% of the lateral acceleration. In summary, while the Hybrid III ATD head-
cervical spine junction was created solely as a tool to evaluate sagittal plane
motion and injury risk in frontal crashes, the ATD appeared to have
LIMITATIONS
noted again that impact location on the ATD head could have played a role in
the response. While frontal and oblique impacts struck the same elevation on
ATD head, the lateral impact struck approximately three inches superior,
Unfortunately, the ATD could not have been struck in the same transverse
plane as the frontal and oblique impacts because this would have involved
direct cervical spine impact. The second drawback was that the total head-
82
CHAPTER 4 – ATD HEAD IMPACTS AND HEAD-NECK DYNAMICS
could only monitor head rotation via the angular rate sensor. Thus, a true
moment and segmental rotation, was not possible. Finally, the ATD was
limitations.
CONCLUSIONS
experiments, the in vitro head-cervical spine response of the Hybrid III 50th
pendulum tests at impact energy ranging from 27J to 89J and momentum
ranging from 14N-s to 25N-s, it was determined Hybrid III ATD head-cervical
compression. For all impacts, Hybrid III ATD resultant head center of gravity
linear acceleration was 52g to 247g, resultant head angular velocity was
83
CHAPTER 4 – ATD HEAD IMPACTS AND HEAD-NECK DYNAMICS
When comparing head impacts at similar energy levels but different impact
all impact energy levels ATD occipito-cervical bending moment was highest in
Resultant ATD occipital force was similar in the lateral and oblique impact
energies.
did not appear coupled to impact direction at lower impact energy levels
levels and when comparing the current ATD loading behavior to the majority
These were the first known data to specifically study Hybrid III ATD
without applied axial compression. This new understanding of Hybrid III ATD
84
CHAPTER 4 – ATD HEAD IMPACTS AND HEAD-NECK DYNAMICS
85
CHAPTER 5 – ATHLETIC PADDING INFLUENCE ON ATD HEAD-NECK IMPACT DYNAMICS
IMPACT DYNAMICS
INTRODUCTION
head impact was quantified. In this Chapter, the isolated Hybrid III response
is built upon via in vitro Hybrid III athletic padding impacts. Particular
to include specific skull, brain and neck injury measures. Finally, theories on
metrics are explored via potential deficits in current athletic testing standards
that utilize only the Gadd Severity Index. These results help to build the
BACKGROUND
measure athletic head injury risk. But linear and rotational dynamics are
injury. Current head injury metrics, namely the Gadd Severity Index (GSI)46
and Head Injury Criterion (HIC)163, are based on resultant linear head center
junction. These injury metrics were developed from human cadaveric head
86
CHAPTER 5 – ATHLETIC PADDING INFLUENCE ON ATD HEAD-NECK IMPACT DYNAMICS
impacts and have long been correlated to skull fracture risk but not
and impact duration are the most widely used and only currently-mandated
injury metrics for athletic head padding impact testing,4;101;142 in spite of the
responsible for concussion and spine injury, athletic helmets and head
padding are still certified solely by the GSI linear kinematic criterion. More
troubling is the fact that while it is commonly believed that padding reduces
athletes’ concussion and spine injury risk, there is little evidence that
sports like boxing and mixed martial arts (MMA). Therefore, while athletic
padding may be capable of reducing linear dynamic quantities like GSI, HIC,
concussion and risk for spine injury are unclear. In this study, athletic type
quantify concussion and spine injury risk for different padding conditions for
LITERATURE SURVEY
rugby and equestrian events. In these sports, padding has been investigated
87
CHAPTER 5 – ATHLETIC PADDING INFLUENCE ON ATD HEAD-NECK IMPACT DYNAMICS
exception of one American Football helmet study166 and one study of jockey
many studies have reported variable linear responses, with some padding
football helmet study showed that linear and rotational dynamics were no
importantly, several head padding studies reported that none of the models
METHODS
This study, which was part of a larger investigation into head impact
dynamics, concussion and spine injury risk, investigated the linear and
rotational dynamics of boxing and mixed martial arts (MMA) padding during a
series of fifty-four (54) pendulum impacts at low (27-29J) and high (54-58J)
88
CHAPTER 5 – ATHLETIC PADDING INFLUENCE ON ATD HEAD-NECK IMPACT DYNAMICS
Test Device (ATD.) A total of five (5) padding combinations were examined:
including fifteen (15) ‘linear’, five (5) ‘rotational’ and nine (9) combined
parameters.
A Hybrid III 50th Percentile ATD was instrumented with a triaxial linear
NM) mounted at the head cg. A set of three triaxial linear accelerometers
were mounted externally on the head but this data will be presented in future
work. A six-channel upper neck load cell (Model 1716A, Denton ATD,
arm. A triaxial linear accelerometer was mounted in the ATD chest but was
later found to have minimal magnitude and will not be presented here. All
and sign convention adhered to SAE J1733145. Linear head jerk, angular jerk
Linear head velocity was calculated via filtered signal integration. The ATD
head weighed 5.1kg with instrumentation, the cervical spine weighed 1.6kg
89
CHAPTER 5 – ATHLETIC PADDING INFLUENCE ON ATD HEAD-NECK IMPACT DYNAMICS
impactor mass. The mass was chosen to approximate the effective upper
study.168 A steel eye hook was threaded into the flat machined face and the
impactor was balanced such that the mass moment of inertia was translated
only in the vertical direction. The same steel sphere was used in all trials.
above the ATD head cg. The ATD was secured to a test stand with tie-down
straps and lower extremities removed at the femur. Inertial responses below
the neck were minimized by securing the test stand with 3500N of sandbags.
The pendulum mass was raised into position via nylon fishing line with 222N
rupture strength, tied to the eye hook and routed through a ceiling-mounted
In each head impact trial, the pendulum mass was raised to 0.76m or
1.52m and the nylon line was cut with a pair of scissors. The swing heights
were selected to re-create low and high energy impacts on par with boxing
punches to the lateral ATD head. These punch data will be presented in a
future publication, but produced peak ATD linear head cg accelerations of 24-
impact dynamics were on par with prior boxing studies124;141;148;165 and in the
90
CHAPTER 5 – ATHLETIC PADDING INFLUENCE ON ATD HEAD-NECK IMPACT DYNAMICS
training glove (0.27kg) and UFC Official MMA glove (0.18kg) were tested.
The boxing and MMA gloves were firmly affixed to the pendulum via a
combination of double sided tape on the mass and gaffers tape on the
exterior. The headgear was secured with double-sided tape on the head and
by cinching taut the chinstrap and posterior lacing. The impact conditions
are described in Table 5-1 and shown in Figure 5-1 through Figure 5-5.
Impact
Impact
Impact Setup Momentum
Energy (J)
(N-s)
Low High Low High
Unpadded 26.8 53.7 13.9 19.7
Bare Pendulum-Headgear 26.8 53.7 14.6 19.7
MMA-Bare Head 28.3 56.5 15.0 20.7
Boxing-Bare Head 28.9 57.9 13.9 21.2
Boxing-Headgear - 57.9 - 21.2
91
CHAPTER 5 – ATHLETIC PADDING INFLUENCE ON ATD HEAD-NECK IMPACT DYNAMICS
92
CHAPTER 5 – ATHLETIC PADDING INFLUENCE ON ATD HEAD-NECK IMPACT DYNAMICS
A total of six (6) trials were conducted for the five impact setups and
the two impact energy levels for a total of 54 impacts. The boxing-headgear
setup was omitted in the low energy condition. Similarly, the MMA glove was
not used in headgear impacts because the boxing glove delivered a more
severe impact under similar test conditions. The multiple trials at each swing
rotation. The coefficient of variation (COV) between setup trials was 2.1% to
Because the swing height was kept constant between setups, the
impact energy and momentum varied by 5.0% to 7.5% between the low and
high energy setups due to the changing pendulum mass. These variations
were considered when analyzing the two-tailed t-test results for the dynamic
Further test variation was found due to the headgear mass, which
93
CHAPTER 5 – ATHLETIC PADDING INFLUENCE ON ATD HEAD-NECK IMPACT DYNAMICS
Theoretical
Post-Impact Vh
Impact Setup
Low High
Energy Energy
Unpadded 3.21 4.54
Bare Pendulum-Headgear 3.30 4.67
MMA-Bare Head 3.34 4.73
Boxing-Bare Head 2.99 4.23
Boxing-Headgear - 4.42
2.9% to 4.3%, adding padding to only the head would theoretically decrease
94
CHAPTER 5 – ATHLETIC PADDING INFLUENCE ON ATD HEAD-NECK IMPACT DYNAMICS
post-impact head velocity by 6.8% and adding padding to both the impactor
Parametric Analysis
For each setup, twenty-nine (29) relevant concussion and spine injury
are presented in the Appendix. The parameters were grouped into ‘linear’,
minimum resultant impact force loading rate, head kinetic energy, head
power, impact duration, Gadd Severity Index (GSI)46, Head Injury Criterion
resultant upper neck moment and resultant head angular momentum. The
95
CHAPTER 5 – ATHLETIC PADDING INFLUENCE ON ATD HEAD-NECK IMPACT DYNAMICS
(RMDM).151 For each parameter, the maximum value, from the mean of the
six trials in each setup, was used in a two-tailed paired t-test comparison
against the unpadded impacts at either low or high energy. Significance was
RESULTS
data were plotted for each of the six trials in each padding condition at high
energy and are shown in Figure 5-7 and Figure 5-8, respectively. Since all of
the impacts were laterally directed, the resultant values were analyzed and
individual x-, y- and z-axis components were not examined separately since
linear and rotational dynamics were largest in the directions of largest motion
and loading.
96
CHAPTER 5 – ATHLETIC PADDING INFLUENCE ON ATD HEAD-NECK IMPACT DYNAMICS
Figure 5-7. Head cg resultant linear acceleration in high energy trials for the four padding
conditions tested
97
CHAPTER 5 – ATHLETIC PADDING INFLUENCE ON ATD HEAD-NECK IMPACT DYNAMICS
Figure 5-8. Occipital neck force resultant in high energy trials for the four padding conditions
tested
As seen in Figures 5-7 and 5-8, there was some degree of variability
variability in the data collected, Table 5-3 collates the respective coefficient
of variation (COV) of each impact condition for the set of six (6) repeated
impact trials. The COV was calculated by dividing the standard deviation
from each set of six impacts by the mean value at the time of highest linear
cg acceleration resultant.
98
CHAPTER 5 – ATHLETIC PADDING INFLUENCE ON ATD HEAD-NECK IMPACT DYNAMICS
Low Energy
Bare-
Quantity Control UFC-Bare Box-Bare
Headgear
Linear CG Acceleration 4.7 6.9 11.8 12.4
Angular Velocity 10.3 24.6 10.0 19.9
Angular Acceleration 10.2 22.1 21.1 13.2
Occipital Condyle Force 7.4 21.1 22.0 16.9
Occipital Condyle Moment 15.1 14.0 27.4 16.9
High Energy
Bare- UFC- Box- Box-
Quantity Control
Headgear Bare Bare Headgear
Linear CG Acceleration 2.1 4.3 13.5 9.9 37.7
Angular Velocity 12.7 29.0 10.2 9.3 33.3
Angular Acceleration 8.4 25.5 19.4 10.6 20.1
Occipital Condyle Force 3.5 5.9 10.8 9.3 23.3
Occipital Condyle Moment 7.8 10.5 20.2 7.8 9.9
Table 5-3. Coefficient of variation of each resultant impact parameter corresponding to time
of highest cg linear acceleration resultant
Thus, even though the conditions were nearly identical for each
padded impact trial set, the COV shown in Table 5-3 was seen to vary from
head angular acceleration and impact force from the five impact setups are
99
CHAPTER 5 – ATHLETIC PADDING INFLUENCE ON ATD HEAD-NECK IMPACT DYNAMICS
Figure 5-9. Head cg resultant linear acceleration in high and low energy trials
Figure 5-10. Head resultant angular velocity in high and low energy trials
100
CHAPTER 5 – ATHLETIC PADDING INFLUENCE ON ATD HEAD-NECK IMPACT DYNAMICS
Figure 5-11. Head resultant angular acceleration in high and low energy trials
Figure 5-12. Resultant impact force in high and low energy trials
101
CHAPTER 5 – ATHLETIC PADDING INFLUENCE ON ATD HEAD-NECK IMPACT DYNAMICS
Figure 5-13. Sample of computational simulation results displaying Mises stresses within brain
from Box-Bare impacts using SIMon software
ATD, the peak resultant linear head acceleration decreased at both energy
levels. Resultant head angular velocity increased for the bare-headgear and
slightly in the bare-headgear setup at high energy; in low energy impacts, all
with the control impact having the highest resultant impact force and the
The twenty-nine (29) dynamic head and neck impact parameters were
measured or computed for each setup by taking the mean of the maximum
102
CHAPTER 5 – ATHLETIC PADDING INFLUENCE ON ATD HEAD-NECK IMPACT DYNAMICS
values from the six trials. These parameters were then compared, via a two-
tailed t-test, to the unpadded control impacts for the low and high energy
setups. The results are summarized in Table 5-4 and Table 5-5. The
maximum value for each of the parameters is shown with the percentage
decrease from the unpadded control setup. Based on this t-test comparison,
For the linear dynamics t-test results, impact duration and HIC
duration increased significantly from the control, while all other parameters
significantly decreased by 33% to 86% across all setups. While the linear cg
acceleration was significantly reduced in all three setups by 59% to 64%, the
setups with the exception of angular velocity, occipital neck moment and
setup was nearly significant (p=0.078.) All of the setups with impactor
padding.
using padding except for the Nij and CSDM05 in the boxing-bare setup,
103
CHAPTER 5 – ATHLETIC PADDING INFLUENCE ON ATD HEAD-NECK IMPACT DYNAMICS
(p=0.066.) Because the DDM was zero for all setups, this parameter was not
In a similar fashion to the low energy t-test results, the high energy
parameters were analyzed for significant differences that might have been
due more to the mass effect as opposed to padding properties. Again, head
104
CHAPTER 5 – ATHLETIC PADDING INFLUENCE ON ATD HEAD-NECK IMPACT DYNAMICS
delta-V, linear momentum and kinetic energy may have been potentially
duration and HIC duration significantly increased across all setups; the longer
impact duration allowed for head delta-V to remain relatively unaffected even
variation than the low energy results. These tests had largest variation in
41% and angular momentum by 18% to 30% while having a varying effect
effect on angular acceleration and was nearly significant for impact moment
that GAMBIT, wPCS, HIP and DDM all significantly decreased for all setups
for the MMA-bare setup. The three CSDM results varied across setups; the
105
CHAPTER 5 – ATHLETIC PADDING INFLUENCE ON ATD HEAD-NECK IMPACT DYNAMICS
MMA-bare setup consistently had higher CSDM values, ranging from 22% to
102% higher than control, with the CSDM05 level being significantly higher.
29% while the 60% (p=0.062) and 65% (p=0.061) reduction in CSDM10 and
risk the Nij significantly decreased for the MMA-bare and boxing-headgear
boxing-headgear setups.
106
CHAPTER 5 – ATHLETIC PADDING INFLUENCE ON ATD HEAD-NECK IMPACT DYNAMICS
High energy
Bare- Box-
Parameter max value Unpadded UFC-Bare Box-Bare
Headgear Headgear
LINEAR PARAMETERS
Impact Duration (s) 0.00350 0.0114 0.0143 0.00730 0.0182
Delta-V (m/s) 2.99 3.04 3.01 2.85 2.78
Head Accel (g) 232 129 117 144 65.0
Jerk (m/s3) 2.74E+06 9.54E+05 8.48E+05 1.27E+06 3.10E+05
Neck Force (N) 1250 746 836 868 629
Momentum (N-s) 15.2 15.5 15.3 16.3 15.9
Impact Force (N) 12800 7160 6630 8900 4240
Max Force Load Rate
1.14E+07 3.96E+06 3.36E+06 5.47E+06 1.26E+06
(N/s)
Min Force Load Rate
-1.22E+07 -4.64E+06 -4.21E+06 -6.71E+06 -1.77E+06
(N/s)
KE head (J) 22.7 23.7 23.1 23.3 22.1
Power head (W) 2.14E+04 1.29E+04 1.19E+04 1.53E+04 6.76E+03
GSI 768 265 219 311 79.0
HIC 676 229 188 271 66.0
HIC Duration (s) 0.0012 0.0020 0.0022 0.0020 0.0038
SFC 200 127 112 124 60.0
ROTATIONAL PARAMETERS
Angular Velocity
18.5 22.6 26.2 16.3 18.3
(rad/s)
Angular Accel (rad/s2) 5260 5550 5240 3800 1740
Angular Jerk (rad/s3) 2.57E+06 2.32E+06 2.08E+06 1.60E+06 6.30E+05
Occipital Neck Moment
48.8 34.0 47.7 38.2 42.4
(N-m)
Angular Momentum
0.345 0.406 0.449 0.321 0.364
(N-m-s)
COMBINED PARAMETERS
Nij 0.39 0.28 0.38 0.31 0.33
GAMBIT 0.936 0.541 0.491 0.583 0.264
wPCS 173 80.9 71.3 87.8 35.8
HIP 20500 12700 11600 12700 5430
CSDM05 63.6 77.4 81.9 45.2 0.210
CSDM10 12.5 21.8 24.5 5.00 0.40
CSDM15 2.4 4.8 5.3 0.8 0.0
DDM 0.8 0.0 0.0 0.0 0.0
RMDM 2.35 2.48 2.17 2.03 1.08
Table 5-5. High energy impact maxima (shaded = significant increase p<0.05, italics =
significant decrease p<0.05)
DISCUSSION
Gadd Severity Index (GSI), in both low and high energy impacts, rotational
107
CHAPTER 5 – ATHLETIC PADDING INFLUENCE ON ATD HEAD-NECK IMPACT DYNAMICS
were unaffected across the range of setups tested in this study. Prior athlete
head padding literature showed variable linear and rotational head and spine
In particular, a recent study examining NFL football helmets found that of the
four new helmet models tested against an existing helmet model, none of the
new helmets consistently reduced linear or rotational head and spine injury
that this prior study conducted descriptive statistical analyses and could not
study.
It was seen in the low energy tests that padding always reduced linear
and combined dynamic parameters with the exception of the Nij and the
variability in Nij and occipital neck moment may have altered concussion and
without head padding. With the exception of Nij and CSDM05, a consistent
decrease in combined dynamic head injury parameters for all low energy
108
CHAPTER 5 – ATHLETIC PADDING INFLUENCE ON ATD HEAD-NECK IMPACT DYNAMICS
setups was seen. The RMDM results, which were a correlate for brain
bridging vein rupture, were cause for concern. For low energy setups, the
RMDM exceeded the tolerance limit of 1.0, indicating striking sport athletes
may be at risk for brain injury even with the use of the hand or head padding
tested here.
In the high energy tests, the results were more variable than observed
in the low energy tests. These tests were conducted at the higher range of
human punch magnitude, while still being below impact magnitudes found in
NFL collisions.38;165;168 Again, mass effects may have been present in head
delta-V, momentum and kinetic energy; the significance of these results may
angular velocity for the MMA-bare and boxing-bare setups may have implied
the control impact; this could have implied that this combination of
109
CHAPTER 5 – ATHLETIC PADDING INFLUENCE ON ATD HEAD-NECK IMPACT DYNAMICS
endured by the cervical spine even though both gloved conditions indicating
increased brain injury risk. This variation in occipital moment between the
setups but could have been also been due to known stiffness issues with the
subjected to repeated blows while using these padding combinations may not
HIP and DDM were significantly reduced for all padding setups when
reduced for MMA-bare and bare-headgear but not the boxing-bare or boxing-
headgear setups. This result might have indicated that the interactions
between the boxing glove or MMA glove and head were important
considerations to minimize spine injury risk, but it must be noted that all Nij
results were well below the injury risk threshold of 1.0. The CSDM was only
correlate to diffuse axonal injury (DAI) and concussion, this result was
more effective at reducing DAI and concussion risk than using padded
gloves. Finally, the RMDM results indicated concussion all padding setups
exceeded the threshold of 1.0 for brain bridging vein rupture. This result
110
CHAPTER 5 – ATHLETIC PADDING INFLUENCE ON ATD HEAD-NECK IMPACT DYNAMICS
may be of concern for athletes who use head and hand padding tested here
but are subjected to repeated blows near these high energy impact
magnitudes.
LIMITATIONS
the contact characteristics between the ATD head, padding and pendulum in
these tests may not have been an exact replica of the friction and
even though impact velocity was held constant, the changing pendulum
including delta-V, momentum and kinetic energy, may not have been as
may not have necessarily been solely a function of padding. Similarly, the
was hypothesized that real-world striking sport punches may involve variable
impact mass due to muscle force and inertial effects. Thus, the constant
3.6kg pendulum effective arm mass may not have been a proper
111
CHAPTER 5 – ATHLETIC PADDING INFLUENCE ON ATD HEAD-NECK IMPACT DYNAMICS
uppercut or oblique punches have yet to be studied. Finally, because the aim
construction characteristics may have all played a role in the presence or lack
CONCLUSIONS
The Gadd Severity Index (GSI) has been the gold standard by which to
quantify athletes’ head injury risk for nearly five decades. However, GSI
accounts only for linear acceleration of the skull while ignoring rotational
injury in athletes. And the cervical spine, which defines the loading boundary
conditions for the head, may play a role in concussion as well. In spite of the
responsible for concussion as well as spine injury, athletic helmets and head
padding are still certified solely by the GSI linear kinematic criterion without
testing for brain injury or spine injury risk. More troubling is the fact that
while it is commonly believed that head and/or fist padding reduces athletes’
concussion and spine injury risk, there is little evidence that confirms this
spine injury risk in striking sports like boxing and mixed martial arts (MMA).
112
CHAPTER 5 – ATHLETIC PADDING INFLUENCE ON ATD HEAD-NECK IMPACT DYNAMICS
This test series of fifty-four (54) low (27-29J) and high (54-58J)
test device (ATD) showed that measured Hybrid III linear head acceleration
(55 to 144 g), angular acceleration (1741 to 5552 rad/s2), angular velocity
(16 to 26 rad/s), neck moment (19 to 49 N-m) and impact force (3083 to
8901 N) were within ranges of theoretical concussion and spine injury limits
in spite of GSI (42 to 311) being below the accepted skull fracture limit of
1200. These results indicated that head injury risk due to linear head cg
force, resultant impact force loading rate, head power, HIC, GSI, SFC,
GAMBIT, wPCS and HIP in striking sports was significantly (p<0.05) reduced
via head and hand boxing and MMA padding at both energy levels. Spine
injury risk predicted by Nij was below the accepted threshold in all setups but
was significantly decreased in only four of the seven setups tested. Padding
energy levels. And in high energy impacts head angular velocity, CSDM and
injury (DAI) and bridging vein rupture regardless of the head and hand
padding used.
indicated that while padding protects against head injury from linear dynamic
113
CHAPTER 5 – ATHLETIC PADDING INFLUENCE ON ATD HEAD-NECK IMPACT DYNAMICS
padding.
114
CHAPTER 6 – HELMET TESTING IMPACT SIMULATIONS
INTRODUCTION
The results from the Hybrid III athletic padding impacts in Chapter 5
indicated that padding inserted between the impactor and ATD head
influenced head, neck and spine injury risk – in some cases, padding
increased head and neck injury risk. These results are particularly
interesting when considering the multiple impact test methodologies that are
used to test athletic padding – drop tests, pendulum impacts and linear
head, neck and spine injury risk results, the purpose of this Chapter is to (i)
collect pilot football helmet impact data and (ii) use this data to develop and
injury risk results from five (5) impact test setups. Thus, the Chapter 5
and rotational kinematics (future testing standards). It was found that the
impactor method used by the NFL with a flexible Hybrid III ATD neck.
Furthermore, drop tests and horizontal impact tests with a rigid head-neck
115
CHAPTER 6 – HELMET TESTING IMPACT SIMULATIONS
coupling currently used by NOCSAE may induce high head accelerations while
BACKGROUND
the gold standard by which to measure athletic head injury risk, with
authors have attempted to unlock the relationship between inertial and direct
contact loading and understand the linear and rotational injury mechanisms.
The rationale for using linear acceleration as the metric to predict head
injury was first postulated in the 1940’s by Gurdjian and Lissner et al.58-60;81
al.116 concluded that inertial rotation alone could not produce injury levels
caused by direct head impact. This was later revised by Ommaya and
approximately half of brain injury potential, with injury due to direct impact
comprising the other half. Gennarelli et al.48;50 showed that linear translation
116
CHAPTER 6 – HELMET TESTING IMPACT SIMULATIONS
hematomas while rotational effects were likely responsible for diffuse axonal
injuries. This study also suggested that linear motion likely induced brain
cavitation, which was thought responsible for blood vessel rupture due to
and skull. Finally, Ono et al.118 conducted a series of primate studies and
impact.
criteria, including the Gadd Severity Index (GSI),46 Head Injury Criterion
correlate well with skull fracture risk, it has been widely thought that
as well.60;75;115
rotational acceleration via inertial and/or direct head loading could cause
tensile strains within the brain tissue. In a study of bridging vein rupture
117
CHAPTER 6 – HELMET TESTING IMPACT SIMULATIONS
diffuse axonal tearing and acute subdural hematomas via induction of intra-
crashes that zero cases of brain injury occurred in absence of direct head
impact and that brain injury was not generally associated with neck injury.
Element brain injury package called SIMon (Simulated Injury Monitor). This
returned risk values for skull fracture, diffuse axonal injury, dilatational
damage due to vacuum and acute subdural hematoma due to tissue tearing.
A finite-element based study by Zhang et al. (2004) showed that shear strain
and shear strain rate due to rotationally-induced shearing within the brain
Viano et al.165 studied concussive impact case studies in the National Football
League with video reconstructions, helmeted ATD impact tests and the
Wayne State Human Model Finite Element simulation. These case studies
showed that brain shear strain and shear strain rate were highest in impacts
118
CHAPTER 6 – HELMET TESTING IMPACT SIMULATIONS
definitive relationship was defined for head and neck dynamics as a function
linear kinematic studies, head injury criteria with rotational and linear
(ULP).86
During the second half of the 20th Century, interest mounted in athletic
helmet testing as players were sustaining severe and fatal head and neck
developing the HIC and GSI criteria, they also created the Wayne State
impact surrogates, like the Hybrid II or Hybrid III automotive crash testing
in drop testing with athletic helmets in the 1975.64 A simple test, wherein
119
CHAPTER 6 – HELMET TESTING IMPACT SIMULATIONS
lowered to 1200 in 1998 – was selected for Abbreviated Injury Scale (AIS)
level 3 head injury based on data from earlier cadaver testing and skull
fracture risk.46;98;163
day, with NOCSAE being the only testing entity and all helmets being
or spine injury.
Review Summary
Hence, though much research has been done on head, neck and spine
injury risk for direct and inertial loading, injury parameters like the Gadd
Severity Index (GSI) and Head Injury Criterion (HIC) developed in the 1960’s
remain the only widely used head injury risk metrics. And GSI is the
exclusive metric by which NOCSAE tests all athletic helmets in headform drop
tests. But GSI and HIC were developed from a series of cadaveric impact
tests and correlate to skull fracture risk; as such, GSI and HIC have no
injury.
concussion, neck and spine injury risk, in this Chapter pilot football helmet
impact experiments were conducted, the data from these experiments were
Simulink and this validated model was then compared to existing NOCSAE
120
CHAPTER 6 – HELMET TESTING IMPACT SIMULATIONS
drop test models as well as proposed pneumatic linear impact models with
METHODS
helmets. Lateral impacts were conducted with the impactor striking a seated
Hybrid III 50th percentile male ATD (Denton ATD, Rochester, MI) in the head.
The ATD head was instrumented with a triaxial linear accelerometer (Model
Denton ATD, Rochester, MI) measured spine forces and moments in three
shear force by the 0.01778m moment arm. All data were collected at
angular velocity signal after determining proper signal content and pre-
150J and impact momentum of 45N-s. The impact energy and momentum
121
CHAPTER 6 – HELMET TESTING IMPACT SIMULATIONS
impactor used in the proposed NOCSAE linear impact standard. The striking
mass was uninstrumented. The impactor was hung from four (4) 6.4mm
carabineers above the ATD head. The ATD was seated in a custom-built
direction and the torso secured with tie-down straps. In order to minimize
inertial responses below the cervical spine, the ATD below the waist was
weighed down with approximately 2500N of sandbags. For each impact trial,
the impactor cg was position such that the initial impact location was in the
coronal plane +/- 2cm from the head cg. The impactor was released from a
consistent swing height of 2.18m above ATD head cg by pulling the nylon
energy and momentum for NOCSAE drop tests, this swing height was
total of five (5) head impact trials were performed as shown in Figure 6-1
122
CHAPTER 6 – HELMET TESTING IMPACT SIMULATIONS
(A) (B)
(D) (E)
(C)
Figure 6-1. Pendulum Impact setups for (A) Unhelmeted, (B) ‘Leatherhead’, (C)-(E) NOCSAE
Approved Helmets
Impact
Impactor Drop Height Impact
ATD Pendulum Impacts Momentum
Mass (kg) (m) Energy (J)
(N-s)
Lateral 7 2.18 45 150
was built in Simulink and validated against the response of the three
behavior of the helmet shell, padding and facemask. The model is shown in
Figure 6-2.
123
CHAPTER 6 – HELMET TESTING IMPACT SIMULATIONS
Figure 6-2. Two DOF system dynamics model for SEA, Ltd. lateral pendulum helmet impacts
validation parameter in the Simulink model against the lateral impacts, with
the peak angular velocity ( N ) being a secondary validation check. For this
model, mass, stiffness and damping parameters were all taken from values
found in literature.9;32;36;123;150;164;176
calibration drop test, (b) NOCSAE helmeted headform certification drop test,
(c) horizontal pneumatic impact with helmeted headform and rigid neck and
(d) horizontal pneumatic impact with helmeted headform and flexible Hybrid
124
CHAPTER 6 – HELMET TESTING IMPACT SIMULATIONS
III neck. These models are illustrated in Figures 6-3, 6-4, 6-5 and 6-6,
respectively.
Figure 6-3. Unhelmeted lateral NOCSAE calibration drop test system dynamics model
125
CHAPTER 6 – HELMET TESTING IMPACT SIMULATIONS
Figure 6-4. Helmeted NOCSAE lateral impact test system dynamics model
126
CHAPTER 6 – HELMET TESTING IMPACT SIMULATIONS
Figure 6-5. Proposed NOCSAE linear impact with highly rigid neck attachment
Figure 6-6. Linear impact used in prior NFL studies with flexible Hybrid III neck
127
(b)
CHAPTER 6 – HELMET TESTING IMPACT SIMULATIONS
conditions. The current NOCSAE drop test calibration protocol served as the
lower bound for initial energy (105J) and momentum (38N-s). The lateral
impact tests conducted at SEA, Ltd. were the next highest initial energy
those conducted in prior NFL head trauma research.21 The inputs for each
model under these conditions are shown in Tables 6-2 and 6-3.
NOCSAE-
7.0 5.5 6.5 7.7 10.9
CALIBRATION
NOCSAE-HELMETED
9.0 4.8 5.8 6.8 9.6
DROP
NOCSAE-LINEAR
14.5 3.8 4.5 5.4 7.6
IMPACT RIGID NECK
NFL LINEAR IMPACT
14.5 3.8 4.5 5.4 7.6
FLEXIBLE NECK
SEA PENDULUM
7.0 5.5 6.5 7.7 10.9
IMPACTS
Table 6-2. Simulink simulation parameters for constant kinetic energy comparison
128
CHAPTER 6 – HELMET TESTING IMPACT SIMULATIONS
NOCSAE-
7.0 5.5 6.4 10.9 21.9
CALIBRATION
NOCSAE-HELMETED
9.0 4.2 5.0 8.5 17.0
DROP
NOCSAE-LINEAR
14.5 2.6 3.1 5.3 10.6
IMPACT RIGID NECK
NFL LINEAR IMPACT
14.5 2.6 3.1 5.3 10.6
FLEXIBLE NECK
SEA PENDULUM
7.0 5.5 6.4 10.9 21.9
IMPACTS
RESULTS
SEA, Ltd. for unhelmeted, ‘leatherhead’ and helmeted Hybrid III are shown in
129
CHAPTER 6 – HELMET TESTING IMPACT SIMULATIONS
20.0
10.0
0.0
0.490 0.500 0.510 0.520 0.530 0.540
-10.0
Linear Acceleration (g)
-20.0
-30.0
-40.0
-50.0
-60.0
-70.0
-80.0
-90.0
-100.0
time (sec)
Figure 6-7. Hybrid III head cg linear acceleration for the single (n=1) pendulum head impacts
0.0
0.490 0.500 0.510 0.520 0.530 0.540
Angular Velocity (rad/sec)
-5.0
-10.0
-15.0
-20.0
-25.0
-30.0
-35.0
time (sec)
Figure 6-8. Hybrid III head angular velocity for n=1 pendulum head impacts
130
CHAPTER 6 – HELMET TESTING IMPACT SIMULATIONS
2000
1000
0
0.490 0.500 0.510 0.520 0.530 0.540
-1000
-2000
-3000
-4000
-5000
time (sec)
Figure 6-9. Hybrid III head angular acceleration for n=1 pendulum head impacts
700
600
500
Neck Force (N)
400
300
200
100
0
0.490 0.500 0.510 0.520 0.530 0.540
-100
time (sec)
Figure 6-10. Hybrid III head upper neck shear force for n=1 pendulum head impacts
131
CHAPTER 6 – HELMET TESTING IMPACT SIMULATIONS
40
30
Neck Moment (N-m)
20
10
0
0.490 0.500 0.510 0.520 0.530 0.540
-10
-20
-30
time (sec)
Figure 6-11. Hybrid III head upper neck bending moment for n=1 pendulum head impacts
with the average head cg linear acceleration response from the three lateral
132
CHAPTER 6 – HELMET TESTING IMPACT SIMULATIONS
30.0
20.0
10.0
0.0
acceleration (g)
-20.0
-30.0
-40.0
-50.0
-60.0
-70.0
time (sec)
Figure 6-12. Simulink simulation results compared with average response from Hybrid III
head cg linear acceleration for the three (3) helmeted pendulum impacts
Maximum
Angular
ATD Head Impact Direction
Velocity
(rad/s)
Lateral Impact Test
-26.2
Average
Simulation -30.5
After the lateral impact simulation was validated against the data
test, NOCSAE horizontal impact test with rigid neck, NFL horizontal impact
test with flexible neck and SEA pendulum impact tests) was exercised in
133
CHAPTER 6 – HELMET TESTING IMPACT SIMULATIONS
Figure 6-13. Constant energy Simulink results for headform cg linear acceleration
134
CHAPTER 6 – HELMET TESTING IMPACT SIMULATIONS
Figure 6-14. Constant momentum Simulink results for headform cg linear acceleration
DISCUSSION OF RESULTS
vitro baseline impact data for the subsequent Simulink validation. Though
the resulting data should be viewed with caution, as only a single impact was
conducted for each of the unhelmeted, ‘leatherhead’ and three (3) helmeted
135
CHAPTER 6 – HELMET TESTING IMPACT SIMULATIONS
acceleration (-62.3g) that was lower than either the unhelmeted (-92.1g) or
against skull fracture and the NOCSAE drop test measurement of resultant
velocity results indicated that in this limited lateral impact dataset that the
examining the ATD head angular acceleration for these pilot impacts, the
risk has been linked to both linear and angular head kinematics,75 the fact
protect against rotational concussion injury better than the ‘leatherhead’ was
helmets have been optimized to reduce linear acceleration but have not yet
the helmeted impacts (526N) reduced peak lateral shear force better than
136
CHAPTER 6 – HELMET TESTING IMPACT SIMULATIONS
(642N). One possible explanation for this reduction could have been that the
contributed to lower dynamic force acting at the skull base. The mean
helmeted neck moment in lateral bending for the helmeted impacts was
rotational kinematic results and may indicate that helmets have been
optimized to reduce linear kinetics but have yet to reduce rotational kinetics.
After analyzing the pilot pendulum impact test data, a simplified two
against the linear cg acceleration and showed very good correlation for the
initial impact loading rate, peak acceleration and unloading rate. The
simulation angular velocity (-30.5rad/s) was slightly higher than the helmet
After validating the lateral pendulum tests at 150J and 45N-s, the
and the NFL linear impact-rigid neck model at constant energy (105J, 150J,
210J and 420J) and constant momentum (38N-s, 45N-s, 76N-s, 154N-s).
For the constant energy simulations, the headform linear acceleration was
137
CHAPTER 6 – HELMET TESTING IMPACT SIMULATIONS
followed by the NOCSAE helmeted drop test; the NOCSAE and NFL linear
impact tests, as well as the SEA pendulum tests were fairly tightly grouped
calibration test also had the lowest impact duration, due to the hard contact
between headform and MEP pad. The linear impacts (NOCSAE, NFL and SEA)
all had similar impact durations and all were longer than the helmeted drop
test. Of note was that the linear impacts with deformable neck always had
longer impact durations and slightly lower linear acceleration peak. Thus,
when considering constant energy impacts, the varying mass for each of the
calibration and helmeted NOCSAE drop tests always had highest linear
acceleration. However, it was seen that for the higher momentum that the
helmeted NOCSAE drop test had longer contact duration than the three linear
impact conditions studied. This result may point to some deficiency in the
elastomeric MEP pad model and likely requires further investigation. Again,
the linear impacts with hybrid III neck model showed slightly lower pulse
width but in the cases of higher momentum showed some difference in peak
acceleration when comparing the NFL to the SEA models. Physical testing
was needed to bear out these results but the constant momentum results
indicate that the impact condition and impactor mass had heavy influence on
138
CHAPTER 6 – HELMET TESTING IMPACT SIMULATIONS
LIMITATIONS
the models analyzed here were very simplified in comparison to variable real-
world impact testing. The simple 1DOF (NOCSAE drop test, NFL linear
impactor) and 2DOF (NFL with flexible neck, Cleveland Clinic) models studied
here lay the foundation for more thorough simulation analysis and parametric
study. Hence, while the helmeted parameters were verified via ATD-helmet
validated against complex real-world data for head and neck dynamics.
and show some resonance in the simulation results. Because this resonance
was not seen in the lateral ATD impact data, system damping requires more
investigation.
CONCLUSIONS
The results from the physical helmeted impact testing indicated that
velocities and with different end conditions like constrained neck, whole body
139
CHAPTER 6 – HELMET TESTING IMPACT SIMULATIONS
The Simulink model validated for the pilot helmet impacts was
compared with four other impact conditions. These models showed that the
impact tests have been to make the impacts as realistic as possible, the
neck did not produce a real-world impact scenario when compared with the
NFL linear impact simulation modeling boundary conditions at the base of the
kinematics since the impact conditions, including impactor mass, energy and
While much more physical testing would be needed to validate all the
Simulink results, the results presented in this Chapter indicated that impact
test method, impactor mass, impact energy and impact momentum all
head and neck injury response. Ultimately, any horizontal impact method
Therefore, based on the results from Chapter 6, and for ease of portability,
out of extruded aluminum sections was developed for use in cadaver and
140
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL HELMET IMPACTS
INTRODUCTION
drop testing setup was less desirable. In this Chapter, the design,
and validated with the express goal of enabling future highly controlled
cadaveric and ATD impact tests that allow analysis of a full range of head
conditions. The second portion describes use of the pendulum impactor for
helmeted ATD head impacts are conducted and similar statistical analyses
are discussed in support of anticipated future helmet impact tests and the
relationship between head, neck and spine injury risk during direct head
141
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL HELMET IMPACTS
ATD head, neck and spine injury in the future. Furthermore, this Chapter
with HIC or GSI -- might omit crucial information regarding x-, y- and z-axis
helmeted impacts conducted in this Chapter also take into consideration the
components contributing to each resultant value for head and neck injury.
IMPACT PENDULUM
DESIGN
human surrogate test devices. Chosen in the current study over other well-
known impact systems like the linear pneumatic impactor or drop test
system, the current impact pendulum design was driven by a desire for
pendulum devices currently in use around the world. Figure 7-1 shows some
examples.
142
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL HELMET IMPACTS
m/s in order to replicate lower energy impacts seen in youth head impacts as
143
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL HELMET IMPACTS
was chosen in order to make the pendulum construction simple and allow for
height be less than ten feet tall, and the width and depth were chosen to
were created in SolidEdge V2.0 (Siemens PLM Software, Plano, TX) and
Columbia City, IN) for production. After the aluminum sections were
2.
After initial assembly, the A-frame was found to have less rigidity than
desired. The frame was retrofit with lateral and longitudinal stabilizers and
144
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL HELMET IMPACTS
After finalizing the assembly of the pendulum, the impact fixtures were
similarly designed and ordered. Images of the cadaveric and ATD impact
145
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL HELMET IMPACTS
and components. The fixtures allow for 360-degree rotation, zero to ninety
degree tilt and submillimeter indexing in the x- and y-shear force directions.
pulling force was used to raise and hold the impactor mass steady prior to
146
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL HELMET IMPACTS
Figure 7-5. Electromagnets used in cadaver [500N] and ATD [1500N] impacts
For the cadaveric pig spine impacts, a two-cable system was used to
prevent impactor rotation and for ATD impacts, a three-cable system was
alignment.
VIBRATION ISOLATION
frame itself might influence impact test results, a series of hammer strikes
against the frame were conducted to attempt to isolate these inputs. The
frame was struck in a number of locations as shown in Figure 7-6 and the
six-channel load cell (ATI Omega IP65, ATI Industrial Automation, Apex, NC)
collected data for the vibration duration. A control data set, with load cell
147
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL HELMET IMPACTS
After collecting the hammer strike frame vibration data, a fast Fourier
(Mathworks, Natick, MA) using the measured load cell vertical force strain
gage. Before conducting the FFT, the data were filtered according to impact
data collection standard SAE J211/1.143 The FFT results are shown in Figure
7-7, with each hammer strike compared against the FFT control data.
148
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL HELMET IMPACTS
Single-Sided Amplitude Spectrum of Load Cell Strain Gage2(t) Single-Sided Amplitude Spectrum of Load Cell Strain Gage2(t)
0.05 0.05
Top Hammer Strike Left Top Hammer Strike
0.045 Baseline 0.045 Baseline
0.04 0.04
0.035 0.035
0.03 0.03
|Y(f)|
|Y(f)|
0.025 0.025
0.02 0.02
0.015 0.015
0.01 0.01
0.005 0.005
0 0
0 50 100 150 200 250 300 350 400 450 500 0 50 100 150 200 250 300 350 400 450 500
Frequency (Hz) Frequency (Hz)
Single-Sided Amplitude Spectrum of Load Cell Strain Gage2(t) Single-Sided Amplitude Spectrum of Load Cell Strain Gage2(t)
0.05 0.05
Left Middle Hammer Strike Left Lower Hammer Strike
0.045 Baseline 0.045 Baseline
0.04 0.04
0.035 0.035
0.03 0.03
|Y(f)|
|Y(f)|
0.025 0.025
0.02 0.02
0.015 0.015
0.01 0.01
0.005 0.005
0 0
0 50 100 150 200 250 300 350 400 450 500 0 50 100 150 200 250 300 350 400 450 500
Frequency (Hz) Frequency (Hz)
Figure 7-7. FFT results for the four (4) frame hammer strikes. NOTE: Chart areas enlarged
from original scale of Frequency = 5000Hz, lY(f)l = 0.8 to show locations of interest
vibratory response when compared with the baseline control data. The
highest magnitude vibrations were seen for the hammer strike to the left
side-bottom of the frame at 23Hz, 120Hz and 205Hz, but these magnitudes
from the pendulum support location (top hammer strike) showed signal
amplitude on par with the baseline control data. With the FFT analysis
complete, confidence was established that collected impact data was isolated
149
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL HELMET IMPACTS
FIXTURE MASS
The fixture mass affixed to the top of the load cell, in between the load
cell and base of cadaveric spine, had a inertial mass effect on load cell
Figure 7-8, load cell data was collected under two conditions: (1) with a
surrogate mass fastened to the load cell and (2) with the same surrogate
mass attached to the fixture sitting atop the load cell. Pendulum impacts
were conducted with the same energy and headform impact location for both
conditions. The load cell forces and moments were compared for these two
Figure 7-8. Fixture mass quantification for fastening directly to load cell and fastening to
fixture components sitting atop load cell
A total of six (6) impacts were conducted with the 3.6kg impactor and
150
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL HELMET IMPACTS
12.3N-s) for the two mounting conditions. In addition to the six-channel load
cell, the ATD headform was instrumented with three (3) uniaxial 500g linear
uniaxial 210 rad/s angular velocity sensors (DTS ARS12K, Seal Beach, CA).
The mean ATD response was analyzed for each data channel. For purposes
resultant angular velocity and resultant bending moment load cell data is
presented here.
140
Mounted to Chuck
Mounted Directly
120
100
Linear acceleration (g)
80
60
40
20
0
0.01 0.015 0.02 0.025 0.03 0.035 0.04
Time (s)
151
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL HELMET IMPACTS
20
Mounted to Chuck
18 Mounted Directly
16
12
10
0
0.01 0.015 0.02 0.025 0.03 0.035 0.04
Time (s)
60
Mounted to Chuck
Mounted Directly
50
Bending Moment (N-m)
40
30
20
10
0
0.01 0.015 0.02 0.025 0.03 0.035 0.04
Time (s)
Figure 7-11. Fixture mass effects - load cell resultant bending moment
152
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL HELMET IMPACTS
From Figures 7-9 and 7-10 it can be seen that the headform
But the resultant bending moment displayed in Figure 7-11 showed that the
mounting to the chuck produced higher bending moments for the impact
duration. Because of the additional 13.1cm moment arm between the chuck-
mounting and load cell, the shear force contributions were responsible for
this increased bending moment. This offset was then qualitatively taken into
BACKGROUND
component basis (head or neck), but not a systems basis (connected head-
neck). But the head and neck are a mechanically linked osteoligamentous
pressure) or crash injury (Neck Injury Criteria-Nij), have utilized head and
spine injury risk criteria developed from impact testing related to the
GSI) or neck injury risk (Nij) separately, relatively few studies have taken a
153
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL HELMET IMPACTS
were conducted with the express intent of (i) developing a reliable method to
study head and osteoligamentous cervical spine linear and rotational dynamic
(ii) collect dynamic impact data from at least four (4) instrumented spinal
column levels in at least three (3) cadaveric specimens, (iii) analyze these
increase in spine injury criterion (neck forces, neck moments, Nij) during the
shear force), and (v) gather sufficient data to conduct a sampling power
analysis to determine the necessary number and type of specimens for future
experiments.
Clinic and were tested in vitro. The specimens were stripped of soft tissue
by removing the mandible and sawing through the skull along a line
154
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL HELMET IMPACTS
comprising the occipital protuberance dorsally and the hyoid bone ventrally.
Any remaining brain and spinal cord tissue was removed from the superior
spinal column and an aluminum striking plate was affixed to the skull based
spinal segments from the third thoracic vertebrate inferiorly. The specimen
temperature. The specimen was kept moist with saline solution and at room
155
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL HELMET IMPACTS
The skull base, C1, C3 and C5 vertebral bodies were prepared for
made from a dice blank. For each mounting block, three (3) 500g uniaxial
linear accelerometers were attached with the exception of C5 which only had
rotation. The ATI six-axis load cell measured reaction forces and moments
at the base of the cervical spine. All data were collected via National
impacted the aluminum interface plate with impact velocity, energy and
momentum detailed in Table 7-1 for a total of six (6) impacts per specimen.
156
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL HELMET IMPACTS
The sixth impact caused injury in each of the three specimens tested.
In between impacts, the specimen was inspected visually and also subjected
Figure 7-13. Instrumentation and testing sequence: Affixing sensors to mounting blocks,
Initial cadaver position at 40-degrees forward tilt, National Instruments DAQ system and
Occipito-cervical disruption caused in specimen 001 after sixth impact
acceleration, angular velocity, force and bending moment test results are
and angular velocity data are displayed from the four instrumented spinal
levels during the 25.9J impact. The force and moment data are shown for
157
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL HELMET IMPACTS
the 4.7J, 10.3J, 14.5J and 25.9J impacts, respectively. Data from the 37.5J
impact are not shown here due to the resulting macroscopic injury in all
Skull Base C1
400 400
Linear Acceleration (g)
200 200
100 100
0 0
0 0.05 0.1 0 0.05 0.1
time (s) time (s)
C3 C5
400 400
Linear Acceleration (g)
200 200
100 100
0 0
0 0.05 0.1 0 0.05 0.1
time (s) time (s)
Figure 7-14. Resultant linear acceleration results from three cadaver specimens for the fifth
impact (KE = 25.9J)
158
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL HELMET IMPACTS
Skull Base C1
200 200
0 0
CAD01 CAD01
CAD02 CAD02
CAD03 CAD03
-100 -100
0 0.05 0.1 0 0.05 0.1
time (s) time (s)
C3 C5
200 200
Angular Velocity (rad/s)
0 0
CAD01 CAD01
CAD02 CAD02
CAD03 CAD03
-100 -100
0 0.05 0.1 0 0.05 0.1
time (s) time (s)
Figure 7-15. Sagittal plane angular velocity results from three cadaver specimens for
the fifth impact (KE = 25.9J)
159
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL HELMET IMPACTS
5 5
x 10 Skull Base x 10 C1
0.5 0.5
0 0
-0.5 CAD01
-0.5 CAD01
CAD02 CAD02
CAD03 CAD03
-1 -1
0 0.05 0.1 0 0.05 0.1
time (s) time (s)
5 5
x 10 C3 x 10 C5
Angular Acceleration (rad/s 2)
0.5 0.5
0 0
-0.5 CAD01
-0.5 CAD01
CAD02 CAD02
CAD03 CAD03
-1 -1
0 0.05 0.1 0 0.05 0.1
time (s) time (s)
Figure 7-16. Sagittal plane angular acceleration results from three cadaver specimens for the
fifth impact (KE = 25.9J)
160
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL HELMET IMPACTS
4.7J 10.3J
500 500
CAD01 CAD01
Force (N)
Force (N)
300 300
200 200
100 100
0 0
0 0.05 0.1 0 0.05 0.1
time (s) time (s)
14.5J 25.9J
500 500
CAD01 CAD01
Force (N)
300 300
200 200
100 100
0 0
0 0.05 0.1 0 0.05 0.1
time (s) time (s)
Figure 7-17. Resultant force results from three cadaver specimens across four impacts (KE =
4.7J, 10.3J, 14.5J, 25.9J)
161
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL HELMET IMPACTS
4.7J 10.3J
CAD01 CAD01
Moment (N-m)
Moment (N-m)
CAD03 CAD03
50 50
0 0
0 0.05 0.1 0 0.05 0.1
time (s) time (s)
14.5J 25.9J
CAD01 CAD01
Moment (N-m)
CAD03 CAD03
50 50
0 0
0 0.05 0.1 0 0.05 0.1
time (s) time (s)
Figure 7-18. Resultant moment results from three cadaver specimens across four impacts (KE
= 4.7J, 10.3J, 14.5J, 25.9J)
Using these cadaveric impact tests as pilot data, a power analysis was
between the porcine specimens tested here and the desired future testing in
a human cadaver head-neck, it was hoped that this power analysis would
variables were measured, the maximum of the resultant value from each of
162
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL HELMET IMPACTS
bending moment were analyzed. The null hypothesis (Ho) was that for each
that of the mean of maximum values. The alternative hypothesis (Ha) was
that deviation of at least 10% from the mean of maximum sample values
Type II error is 20% or less. The power analysis plots are show in bar
Acceleration
Necessary samples for Power = 0.80
100
50
37.5
25.9 C5
14.5 C3
10.3
C1
4.7
Occiput
Energy (J) 1.2
Figure 7-19. Resultant linear acceleration power analysis results from three cadaver
specimens and four instrumented spinal levels
163
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL HELMET IMPACTS
Angular Velocity
150
100
50
37.5
25.9 C5
14.5 C3
10.3
C1
4.7
Occiput
Energy (J) 1.2
Figure 7-20. Angular velocity power analysis results from three cadaver specimens and four
instrumented spinal levels
164
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL HELMET IMPACTS
Angular Acceleration
150
100
50
37.5
25.9 C5
14.5 C3
10.3
C1
4.7
Occiput
Energy (J) 1.2
Figure 7-21. Angular acceleration power analysis results from three cadaver specimens and
four instrumented spinal levels
165
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL HELMET IMPACTS
Force/Moment
150
100
50
37.5
25.9
14.5
10.3
Moment
4.7
Energy (J) Force
1.2
Figure 7-22. Force and moment power analysis results from three cadaver specimens and
four instrumented spinal levels
As seen in the previous figures, for rotational data like angular velocity
and angular acceleration, the required samples were generally higher than
linear acceleration or the measured force or moment from the load cell.
Based on this pilot set of cadaveric impacts, and because of the variability in
80% power across all parameters was n > 62. It is likely infeasible to
166
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL HELMET IMPACTS
After conducting the cadaveric impacts and pilot test power analysis, a
similar set of experiments were conducted with helmeted ATD. This was
done because there is high need for tightly controlled ATD direct head impact
to a Hybrid III 50th percentile ATD neck. The base of the neck was bolted
total of three (3) linear accelerometers like those used in the cadaveric study
were attached to the mounting location near the headform center of gravity
and a custom mount was crafted to enable attachment of three (3) angular
was implemented to transform the measured angular velocity into the linear
the global reference frame and was stationary throughout impact. A second
NOCSAE medium headform was placed inside a Riddell VSR-4 football helmet
with polycarbonate shell, vinyl nitrile foam and steel facemask. The total
striking mass, including headform and helmet, was 6.7kg. This helmet was
positioned such that the center of gravity of the second headform was level
with the stationary headform at the moment of impact. The headform and
helmet were hung in an axial ‘spearing’ position from three light steel cables
167
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL HELMET IMPACTS
attached to the impact frame cross members so that at impact the crown of
the headform inside the striking helmet was oriented normal to the struck
impact, the swing height was confirmed and the struck and striking helmets
were returned to their initial position via use of an Optotrak Certus three-
dimensional motion monitoring system. Video was taken at 30fps for each
impact and each helmet was impacted six (6) times consecutively during
168
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL HELMET IMPACTS
Figure 7-23. Football helmet power analysis impact setup alignment, lateral impact, oblique
impact, frontal impact, helmet positioning, overall view
The impact velocity was selected to mimic lower velocity youth football
impacts, well below that of prior NFL testing at 6-12m/s.165 Three impact
169
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL HELMET IMPACTS
injury from in-game analysis.122 While there were eleven (11) NOCSAE-
helmets was chosen due to time constraints. Hence, one modern helmet was
chosen from each of the four current helmet manufacturers and an authentic
included based on pilot testing results found in Chapter 6. The impact matrix
Swing Impact
Energy Momentum
HELMET MODELS TESTED ORIENTATION* Height Velocity
(J) (N-s)
(m) (m/s)
RIDDELL VSR-4, SCHUTT AIR
LATERAL,
ADVANTAGE, XENITH X1,
OBLIQUE, 0.20 2.0 13 13
ADAMS A2000,
FRONTAL
LEATHERHEAD
RIDDELL VSR-4, SCHUTT AIR
LATERAL,
ADVANTAGE, XENITH X1,
OBLIQUE, 0.62 3.5 41 23
ADAMS A2000,
FRONTAL
LEATHERHEAD
RIDDELL VSR-4, SCHUTT AIR
LATERAL,
ADVANTAGE, XENITH X1,
OBLIQUE, 1.28 5.0 84 33
ADAMS A2000,
FRONTAL
LEATHERHEAD
Table 7-2. Impact matrix for football helmet power analysis (*to avoid facemask contact,
frontal and oblique impacts used a 5° forward inclination)
170
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL HELMET IMPACTS
70 25
60
20
50
15
40
30
10
20
5
10
0 0
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
Time (s) Time (s)
120 7000
6000
100
80
40
2000
20
1000
0 0
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
Time (s) Time (s)
Figure 7-24. ATD-measured linear acceleration, angular velocity, angular acceleration and
bending moment from frontal pendulum impacts at 84J (5.0m/s)
171
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL HELMET IMPACTS
70 25
60
20
50
15
40
30
10
20
5
10
0 0
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
Time (s) Time (s)
120 7000
6000
100
80
40
2000
20
1000
0 0
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
Time (s) Time (s)
Figure 7-25. ATD-measured linear acceleration, angular velocity, angular acceleration and
bending moment from oblique pendulum impacts at 84J (5.0m/s)
172
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL HELMET IMPACTS
70 25
60
20
50
15
40
30
10
20
5
10
0 0
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
Time (s) Time (s)
120 7000
6000
100
80
LATERAL: 5.0 m/s
4000
MODERN HELMETS
60
LEATHERHEAD
3000
40
2000
20
1000
0 0
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
Time (s) Time (s)
Figure 7-26. ATD-measured linear acceleration, angular velocity, angular acceleration and
bending moment from lateral pendulum impacts at 84J (5.0m/s)
After collecting and collating the data from the 270 individual impact, a
power analysis was done for each helmet during each of the 45 impact
angular acceleration, force and moment from the individual struck NOCSAE
headform impacts were compared to the mean of each set of six impacts for
each condition. The null hypothesis (Ho = mean of the maximum values
from each set of six impacts), alternate hypothesis (Ha deviates 10% from
Ho), significance level (a = 0.05) and power (P = 80%; Type II error risk =
20%) remained the same as was used in the cadaveric analysis. The football
173
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL HELMET IMPACTS
Calculated Power
Calculated Power
0.6
0.6
0.4
0.4
0.2
Moment 0.2
0 Moment
0
Force
01 Force
01
02 Ang Accel
02 Ang Accel
09 Ang Vel
09 Ang Vel
16
Lin Accel 16
Helmet LB Lin Accel
Helmet LB
Calculated Power
0.6
0.4
0.2
Moment
0
Force
01
02 Ang Accel
09 Ang Vel
5.0 m/s 16
Lin Accel
Helmet LB
Figure 7-27. Lateral football helmet impact power analysis (73/75 conditions had P = 80% or
higher)
Calculated Power
0.6 0.6
0.4 0.4
0.2 0.2
Moment Moment
0 0
Force Force
01 01
02 Ang Accel 02 Ang Accel
0.6
0.4
0.2
Moment
0
Force
01
02 Ang Accel
09
5.0 m/s 16
Ang Vel
Lin Accel
Helmet LB
Figure 7-28. Oblique football helmet impact power analysis (71/75 conditions had P = 80% or
higher)
174
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL HELMET IMPACTS
Calculated Power
Calculated Power
0.6 0.6
0.4 0.4
0.2 0.2
Moment Moment
0 0
Force Force
01 01
02 Ang Accel 02 Ang Accel
Calculated Power
0.6
0.4
0.2
Moment
0
Force
01
02 Ang Accel
09
5.0 m/s 16
Ang Vel
Lin Accel
Helmet LB
Figure 7-29. Frontal football helmet impact power analysis (67/75 conditions had P = 80% or
higher)
Thus, across all of the 45 impact conditions, and the five parameters
analyzed for this power analysis, a total of 210 out of 225 (93%) analysis
As a comparison, the NFL test power analysis was done and is shown
in Figure 7-30.
175
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL HELMET IMPACTS
0.6
0.4
0.2
6.3
8.1
8.8
9.1
9.4
9.5
9.6 Lin Accel
10.3 Ang Accel
Impact Velocity (m/s) 10.7 GSI
10.8
Figure 7-30. NFL in-game reconstructed football helmet impact power analysis (13/30
conditions had P = 80% or higher)
Thus, the higher energy NFL tests achieved 80% power in 13 out of 30
achieve 80% power was n=7 for shell-to-shell and n=13 for facemask
impacts. This was contrasted with the lower energy groups of n=6 shell-to-
shell impacts studied here; power was 80% or higher for 5m/s impacts in 71
out of 75 (95%) cases. This indicated it was very much possible to conduct
sample size was gathered. Furthermore, this analysis might have provided a
rationale that some of the NFL impacts lacked sufficient power and required
176
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL HELMET IMPACTS
LIMITATIONS
There were several limitations in both the pig spine and helmeted ATD
impacts. For the pig spine testing, controlling specimen orientation proved to
precariously before the next impact. Oftentimes, the specimen fell out of
balance and tilted to one side which required aborting the impact and re-
positioning the occiput. Another issue with pig spine testing was that
have been desirable to develop a jig that would hold the pig spine in place
challenging as well, and although the specimens were all prepared similarly,
Finally, in the pig spine testing, there was no means to arrest the impactor
impact.
For the helmeted ATD impacts the main limitation was that only a
single helmet of each type was tested. The intra-helmet impact attenuation
effects remain unquantified. Hence, the results shown here could have had
some artifact related to the specific helmet tested and that the true helmet
177
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL HELMET IMPACTS
make and model were tested in a similar manner. Also, the striking helmet
was always the same Riddell VSR-4, and the potential performance
for the impacting interface lifespan should be developed in order to avoid any
CONCLUSIONS
Impactor Design
structure. The test fixtures mounted inside the structure were designed for
Finally, the impactor vibration response and inertial effect of the fixture mass
on the load cell were studied. Frame vibration during impact contributed
negligible artifact to the sensor signals. The added fixture mass on top of the
load cell induced higher measured bending moments due to a larger moment
arm between load cell sensing face and perpendicular shear force. This
result played an important role in quantifying end condition loading for future
178
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL HELMET IMPACTS
thoracic vertebrae, was developed. While each specimen was injured at the
highest energy impact, little else can be ascertained from these experiments
analysis indicated that for measuring linear and angular kinematics, a future
necessary statistical power. Hence, future work may benefit more from a
focus on functional spine unit testing in order to isolate the many variables
uncontrolled here. But these pilot experiments still did much to support the
rotational head and neck kinematics increased. Qualitatively, the peak head
increased at roughly the same rate as peak head angular velocity. When
examining the base of the spine, the shear force and bending moments
same rates. Thus, these data gave initial qualitative insight as to the benefit
whole. Thus, head and spine injury risk may be dependent upon end
conditions applied at the base of the neck as well as inputs to the head.
179
CHAPTER 7 – IMPACT PENDULUM DESIGN, PIG SPINE TESTING, FOOTBALL HELMET IMPACTS
condition and the power was calculated based on the collected data. From all
impacts and all data metrics analyzed, it was found that for these lower
energy impacts that 93% of the metrics achieved power of at least 80%.
This was contrasted with the high energy NFL impacts wherein 44% of
metrics achieved the necessary power. And while not specifically analyzed
180
SUMMARY
8 SUMMARY
In the first Chapter, inertial loading to the head and spine was
were that nearly 80% of struck occupants visited a medical professional post-
crash while none of the striking occupants did the same, crash Delta-V for
struck vehicles was generally 4km/h to 8km/h, on par with sitting in a chair
dummy in the software MADYMO. The model was validated with BioRIDII
against real world data and the full body loading was quantified via virtual
sensors within the model. Injury risks for all body regions were all less than
10% of the appropriate injury tolerance data with the exception of the
proposed ‘whiplash’ neck protection criterion (Nkm). When these injury data
were compared with the diagnosed injuries in Chapter 1, it was found that
the dynamic loading in these crashes was insufficient to cause AIS2+ injury.
The model was further exercised with HybridIII 50th Percentile ATD and it
was found generally that the Hybrid III loading was slightly higher than
181
SUMMARY
crash tests or simulations that use Hybrid III to consider that the loading
than that measured by HybridIII. Finally, the simulation showed low impact
head contact with the head restraint. As a means to study head-neck injury
further, direct contact injury was studied in Chapter 3 via re-analysis of the
low energy inertial and direct impact loading to high energy frontal airbag
impacts was performed. The Nij, or neck injury criterion, was established in
quantify neck injury risk due to airbag contact. In this chapter, novel neck
normalized force along with industry standards for injury rating through the
with accuracy as high 95% in uninjured specimens and 100% for injured
specimens. This was contrasted with Nij accuracy of 78% and 65% for
The head-neck injury analysis from Chapter 3 was carried into Chapter
4 via extensive use of anthropomorphic test device (ATD) tests. And while
182
SUMMARY
quantify the Hybrid III ATD response to frontal, oblique and lateral impacts.
athletics, but the ATD response had never been quantified to direct impact in
absence of compression. It was found that the lateral direction provided the
This was an important result as the Hybrid III was originally only validated
Chapters 1-3 into studying direct head impacts in athletics. In this Chapter,
pugilistic head strikes were recreated using a weighted pendulum and several
protective padding conditions were studied. Based on the results from this
chapter, it was seen that certain head and neck kinematic and kinetic impact
and ATD. But heavier padding on the impactor, such as using a boxing glove
padding. These results were also analyzed via the SIMon finite element brain
model and some brain injury risk parameters when using padding actually
Essentially, head and neck injury protection due to direct head impact was
183
SUMMARY
system dynamics model versus real-world pilot helmeted head impact data.
Many impacting systems were analyzed with the express requirements of the
and ATD), cost effective and able to conduct experiments with high
systems. The results were that a simple gravity-fed pendulum could provide
the NFL, but cost less and provided more fidelity at lower impact velocities.
This result was crucial in the development of a gravity fed head-neck impact
pendulum in Chapter 7.
validated and put through cadaveric and crash dummy testing. The
these data also hinted that the boundary conditions at the base of the neck,
as well as direct inputs to the head, influence head-neck loading. The helmet
tests and power analysis indicated that n=6 impacts done under controlled
184
SUMMARY
sufficient power 93% of the time. Of note was that prior high energy NFL
The findings from this dissertation could prove useful in aiding design
dynamics responsible for head and cervical spine injury related to direct head
impact or inertial loading as well as the creation of novel injury metrics for
185
APPENDIX
9 APPENDIX
Impact duration was selected based on the length of the loading and unloading pulse after
~
aR t ~aCGX t 2 ~aCGY t 2 ~aCGZ t 2
where
~
aCGX t measured head cg acceleration along x - axis
~
aCGY t measured head cg acceleration along y - axis
~
aCGZ t measured head cg acceleration along z - axis
~
jR t
d ~
dt
aR t
A-5 Resultant Occipital Force:
~ ~ 2 2 2
FoccR t F t F~occY t F~occZ t
occX
where
~
FoccX t measured occipital force along x - axis
~
FoccY t measured occipital force along y - axis
~
FoccZ t measured occipital force along z - axis
~
pR t mheadDelta - VR t
where
mhead 5.08kg
186
APPENDIX
~ ~ 2 2 2
FIR t F t F~IY t F~IZ t
IX
where
~ ~ ~
FIX t mheadaCGX t FoccX t
~ ~ ~
FIY t mheadaCGY t FoccY t
~ ~ ~
FIZ t mheadaCGZ t FoccZ t
d ~
Maximum Resultant Impact Force Loading Rate max F t
IR
dt
d ~
Minimum Resultant Impact Force Loading Rate min F t
IR
dt
1
KEhead t mhead Delta - VR t
2
2
d
P t KEhead t
dt
T
~
a t
2.5
GSI R dt
0
where
~
aR t in g
0 : T essential impact duration, chosen to be 15 msec
A-13 Head Injury Criterion (HIC):
2.5
t2
1 ~
HIC t2 t1 aR t dt
t2 t1
t1
where
~
aR t in g
t1 : t2 time period, in msec, where HIC is maximized
A-14 HIC Duration:
HIC Duration t2 t1
187
APPENDIX
maxDelta - VR t 1
SFC
HIC Duration g
m
g 9.81 2
s
~ t
R ~X t 2 ~Y t 2 ~Z t 2
where
~ t measured head angular velocity about x - axis
X
~ t measured head angular velocity about y - axis
Y
~ t measured head angular velocity about z - axis
Z
~ t
R ~X t 2 ~Y t 2 ~Z t 2
where
~ t measured head angular acceleration about x - axis
X
~ t measured head angular acceleration about y - axis
Y
~ t measured head angular acceleration about z - axis
Z
R t
d ~
dt
R t
A-19 Resultant Occipital Moment:
~ ~ 2 ~ 2 2
MoccR t t M ~
occY t MoccZ t
M
occX
where
~
MoccX t measured occipital moment about x - axis
~
MoccY t measured occipital moment about y - axis
~
MoccZ t measured occipital moment about z - axis
R t
~ I X ~X t 2 IY ~Y t 2 IZ~Z t 2
where
I X head mass moment of inertia about x - axis 0.0204kg - m2
IY head mass moment of inertia about y - axis 0.0211kg - m2
I Z head mass moment of inertia about z - axis 0.0143kg - m2
The combined group comprised the remaining eight parameters: Neck Injury Criterion (Nij)40,
Generalized Acceleration Model for Brain Injury Threshold (GAMBIT)103, weighted Principal
Component Score (wPCS)57, Head Impact Power (HIP)105, Cumulative Strain Damage Measure
188
APPENDIX
(CSDM-0.05, 0.10 and 0.15 levels)151, Dilatation Damage Measure (DDM)151 and Relative
~
F
Nij max occZ
t F~occY t d M
~
occX
FZcrit M Xcrit
where
FZcrit critical occipital axial force 6806N
M Xcrit critical occipital moment 135N m
d 0.01778m
1
~ 2.5 ~ t 2.5 2.5
aR t
GAMBIT t ~ ~R
aC C
where
~
aC critical head linear acceleration 250g
~
critical head angular acceleration 25,000rad/s2
C
~ ~ ~ ~ ~
HIP mhead aCGX t aCGX t dt aCGY t
a t dt a
CGY CGZ t ~
aCGZ t dt
~ t
~ t I
~ t
~ t I
~ t
t
~
I X X X Y Y Y Z Z Z
CSDM 05 % Brain volume exceeding 5% principal strain as determined by SIMon Finite Element Model
CSDM10 % Brain volume exceeding 10% principal strain as determined by SIMon Finite Element Model
189
APPENDIX
t
CSDM15 % Brain volume exceeding 15% principal strain as determined by SIMon Finite Element Model
DDM % Brain volume experiencing negative pressure less than - 101.4 kPa as determined by SIMon Finite Elemen
t
RMDM
F t , t
where
t bridging vein strain as determined by SIMon Finite Element Model
F t , t bridging vein failure strain at given strain rate as determined by SIMon Finite Element Model
190
BIBLIOGRAPHY
10 BIBLIOGRAPHY
injury scale 1980. 1980. Barrington, Illinois, USA, Association for the
5. Bailey MN, Wong BC, Lawrence JM. Data and methods for estimating
2008;52:117-28.
191
BIBLIOGRAPHY
2003;13:66-72.
11. Brault JR, Siegmund GP, Wheeler JB. Cervical muscle response during
12. Broglio SP, Ju YY, Broglio MD et al. The Efficacy of Soccer Headgear.
J.Athl.Train. 2003;38:220-4.
13. Broglio SP, Ju YY, Broglio MD et al. The Efficacy of Soccer Headgear.
J.Athl.Train. 2003;38:220-4.
192
BIBLIOGRAPHY
14. Broglio SP, Ju YY, Broglio MD et al. The Efficacy of Soccer Headgear.
J.Athl.Train. 2003;38:220-4.
2005;14:84-9.
neck model validated for near-vertex head impact. Stapp Car Crash
17. Campbell KL. Energy basis for collision severity. Society of Automotive
19. Carter JW, Ku GS, Nuckley DJ et al. Tolerance of the cervical spine to
20. Cassidy JD, Carroll LJ, Cote P et al. Effect of eliminating compensation
for pain and suffering on the outcome of insurance claims for whiplash
193
BIBLIOGRAPHY
21. Casson IR, Pellman EJ, Viano DC. Concussion in the National Football
2009;20:195-214, x.
22. Casson IR, Pellman EJ, Viano DC. Concussion in the National Football
2009;20:195-214, x.
23. Caswell SV, Deivert RG. Lacrosse Helmet Designs and the Effects of
J.Orthop.Res. 1992;10:446-53.
J.Orthop.Res. 1992;10:446-53.
26. Chapline JF, Ferguson SA, Lillis RP et al. Neck pain and head restraint
Accid.Anal.Prev. 2000;32:287-97.
194
BIBLIOGRAPHY
28. Ching RP, Nuckley DJ, Hertsted SM et al. Tensile mechanics of the
29. Cipriani AL, Bayan FP, Woodhouse ML et al. Low speed collinear impact
37. 2002.
31. Croft AC, Haneline MT, Freeman MD. Low speed frontal crashes and
Annu.Proc.Assoc.Adv.Automot.Med. 2002;46:79-91.
1998:454-68.
81.
195
BIBLIOGRAPHY
81.
2005;6:86-94.
36. Doeblin EO. Mechanical Engineering 687 Course Notes. The Ohio State
37. Duma SM. Analysis of field data regarding concussive head impact in
38. Duma SM, Manoogian SJ, Bussone WR et al. Analysis of real-time head
2005;15:3-8.
196
BIBLIOGRAPHY
41. Finan JD, Nightingale RW, Myers BS. The influence of reduced friction
2008;9:483-8.
42. Forbes PA, van Rooij L, Philippens M et al. Occupant Safety in Freefall
Japan.
44. Funk JR, Cormier JM, Bain CE et al. Validation and application of a
soccer ball impacts. 2009 SAE World Congress Paper Number 2009-
45. Funk JR, Cormier JM, Bain CE et al. An evaluation of various neck
Suppl 1:5-11.
197
BIBLIOGRAPHY
48. Gennarelli TA, Adams JH, Graham DI. Acceleration induced head injury
J.Trauma 1982;22:680-6.
50. Gennarelli TA, Thibault LE, Adams JH et al. Diffuse axonal injury and
2000;8:87-95.
52. Gilchrist A, Mills NJ. Protection of the side of the head. Accid.Anal.Prev.
1996;28:525-35.
53. Goel VK, Weinstein JN. Biomechanics of the Spine: Clinical and
54. Goel VK, Winterbottom JM, Schulte KR et al. Ligamentous laxity across
Proceedings:441-82.
198
BIBLIOGRAPHY
57. Greenwald RM, Gwin JT, Chu JJ et al. Head impact severity measures
2008;62:789-98.
58. Gurdjian ES, Lissner HR, Hodgson VR et al. Mechanism of head injury.
Clin.Neurosurg. 1964;12:112-28.
59. Gurdjian ES, Lissner HR, WEBSTER JE. The mechanism of production
1955;101:680-90.
61. Happer AJ, Hughes MC, Peck MD et al. Practical analysis methodology
for low speed vehicle collisions involving vehicles with modern bumper
injuries in rear end car impacts: influence of car seats and possible
199
BIBLIOGRAPHY
63. Hertz E. A note on the head injury criterion (HIC) as a predictor of the
1975;7:225-32.
65. Hodgson VR, Thomas LM. Breaking strength of the skull vs. impact
of Transportation.
71. Hu AS, Bean SP, Zimmerman RM. Response of belted dummy and
Proceedings:589-625.
200
BIBLIOGRAPHY
75. King AI, Yang KH, Zhang L et al. Is head injury caused by linear or
76. Knouse CL, Gould TE, Caswell SV et al. Efficacy of Rugby Headgear in
2003;38:330-5.
201
BIBLIOGRAPHY
79. Kuppa SM. Injury criteria and anthropomorphic test devices for
Administration.
2001;8:604-9.
Surg.Gynecol.Obstet. 1960;111:329-38.
J.Biomech. 1975;8:351-6.
Biomed.Sci.Instrum. 2006;42:383-8.
85. Margulies SS, Thibault LE. A proposed tolerance criterion for diffuse
2008;40:1135-48.
202
BIBLIOGRAPHY
1995;12:621-5.
91. Mertz HJ, Patrick LM. Strength and response of the human neck.
Stapp.Car.Crash.J. 1971;15:207-55.
92. Mertz HJ, Patrick LM. Investigation of the kinematics and kinetics of
93. Mertz HJ, Prasad P. Improved neck injury risk curves for tension and
Stapp.Car.Crash.J. 2000;44:59-75.
Automotive Engineers.
203
BIBLIOGRAPHY
96. Moss RT, Bardas AM, Hughes MC et al. Injury symptom risk curves for
97. Myers BS, McElhaney JH, Doherty BJ et al. The role of torsion in
98. Nahum AM, Gatts JD, Gadd CW et al. Impact tolerance of the skull and
of Transportation.
2008.
204
BIBLIOGRAPHY
2003;10:85-90.
103. Newman JA. A generalized model for brain injury threshold (GAMBIT).
1986.
106. Niehoff P, Gabler HC. The accuracy of winsmash delta-v estimates: the
Annu.Proc.Assoc.Adv.Automot.Med. 2006;50:70-86.
and loading rates affect buckling modes and injury mechanisms in the
205
BIBLIOGRAPHY
J.Biomech. 2007;40:535-42.
109. Nightingale RW, Chancey VC, Luck JF et al. The human cervical spine
J.Biomech. 1996;29:307-18.
112. Nuckley DJ, Van Nausdle JA, Eck MP et al. Neural space and
113. Nusholtz GS, Di DL, Shi Y et al. Studies of neck injury criteria based on
2002;16:220-42.
206
BIBLIOGRAPHY
115. Ommaya AK, Hirsch AE. Tolerances for cerebral concussion from head
116. Ommaya AK, Hirsch AE, Martinez JL. The role of whiplash in cerebral
660804. 1966.
1996;28:493-9.
1980.
120. Panjabi MM, Crisco JJ, Vasavada A et al. Mechanical properties of the
121. Patrick LM, Chou CC. Response of the human neck in flexion,
207
BIBLIOGRAPHY
2003;53:799-812.
Neurosurgery 2006;58:78-96.
47.
1995;117:474-8.
Stapp Car Crash Conference Journal SAE Paper Number 841656. 1984.
208
BIBLIOGRAPHY
128. Ronnen HR, de Korte PJ, Brink PR et al. Acute whiplash injury: is there
1996;201:93-6.
129. Rupp JD, Reed MP, Van Ee CA et al. The tolerance of the human hip to
130. Schmitt KU, Muser M, Niederer P. A new neck injury criterion candidate
for rear-end collisions taking into account shear forces and bending
tissue neck injuries in cases with long sick leave times. Traffic.Inj.Prev.
2003;4:162-8.
Clin.Neurosurg. 1964;12:226-36.
133. Schneider RC, REIFEL E, CRISLER HO et al. Serious and fatal football
2000;8:55-67.
209
BIBLIOGRAPHY
135. Seeman MR, Muzzy WH, Lustick LS. Comparison of human and hybrid
III head and neck dynamic response. Stapp Car Crash Journal
1986;30th Proceedings:291-311.
137. Siegmund GP, King DJ, Montgomery DT. Using barrier impact data to
138. Siegmund GP, Heinrichs BE, Chimich DD et al. The effect of collision
Accid.Anal.Prev. 2005;37:275-85.
2005;6:267-77.
140. Smith PK, Hamill J. The effect of punching glove type and skill level on
1986;12:153-61.
141. Smith TA, Bishop PJ, Wells RP. Three dimensional analysis of linear
210
BIBLIOGRAPHY
142. Snell Memorial Foundation I. Standard for protective headgear for use
Automotive Engineers.
146. Spitzer WO, Skovron ML, Salmi LR et al. Scientific monograph of the
J.Neurosurg.Spine 2005;3:471-6.
Med. 2008.
211
BIBLIOGRAPHY
149. Strother CE, Woolley RL, James MB et al. Crush energy in accident
68). 2006.
2003;47:107-33.
152. Tencer AF, Mirza S, Cummings P. Do "whiplash" victims with neck pain
Annu.Proc.Assoc.Adv.Automot.Med. 2001;45:203-14.
153. Thibault LE, Gennarelli TA. Brain injury: An analysis of neural and
J.Athl.Train. 2008;43:578-84.
212
BIBLIOGRAPHY
155. Torg JS, Quedenfeld TC, Thieler ER, III et al. Collision with spring-
Statistics.
157. United Kingdom Rail Safety and Standards Board. Improving the
1972;4:24-6.
the human muscular and ligamentous cervical spine. Stapp Car Crash
J. 2000;44:85-102.
213
BIBLIOGRAPHY
Annu.Proc.Assoc.Adv.Automot.Med. 2003;47:363-81.
2006;124:69-76.
2005;57:1154-72.
214
BIBLIOGRAPHY
168. Walilko TJ, Viano DC, Bir CA. Biomechanics of the head for Olympic
Proceedings:137-48.
Proceedings:313-31.
204.
215
BIBLIOGRAPHY
176. Yoganandan N, Pintar FA, Zhang J et al. Upper neck forces and
Ann.Biomed.Eng 2008;36:406-14.
178. Zhang L, Yang KH, King AI. A proposed injury threshold for mild
216