1
Appendix A – Simplified Vehicle Model (SVM) Tuning Injury Metric Comparison
2
3
4
5
6
The dashboard and deformable seat of the SVM were based on the 2001 Ford Taurus NCAC model
(Marzougui et al. 2012) and the steering wheel frontal airbag (*AIRBAG_SIMPLE_AIRBAG_MODEL card in LSDYNA) were based upon the open source NCAC inflating airbag model (Bedewi et al. 1996). A custom steering
column, capable of compressing under axial loads, was developed for the SVM. A calibrated foam material model
was implemented to model the effects of a knee airbag in the Camry case only.
7
8
Table A1: Vehicle parameter ranges for variation study to tune the SVM restraint systems in the Camry and Cobalt
cases to NCAP crash test data. The vehicle parameters selected to tune the SVM for each case are bolded.
2010 Toyota Camry
Restraint
System
Frontal
Airbag
Parameter
Units
Min
Selected
Max
Min
Selected
Max
Peak Inflation Rate
kg / s
1.92
2.55
2.71
1.5
2.9
3.5
Vent Area
mm2
800
1750
1800
800
1800
2000
N
2750
2950
3150
2400
2600
3400
N
2700
3100
3250
2400
3250
3500
0.1
0.17
0.24
Peak
Force
Seatbelt
Steering
Column/
Wheel
Knee Airbag
(KAB) /
Knee Bolster
2006 Chevrolet Cobalt
Pretensioner
Retractor Load Limiter
Force
Belt Buckle Friction
Coefficient
Unitless
Shear Bolt Fracture
Force
N
Stroke Resistance
N
N/A
4500
4750
8000
No Column Compression
1800
1800
1800
Rim Modulus
MPa
N/A
30,000
KAB Modulus
kPa
100
177
200
KAB Maximum Strain
Unitless
0.9
0.945
1.0
KAB Damping Factor
Unitless
0.25
0.55
0.6
KAB Thickness
mm
103
104
112
Bolster Modulus
MPa
50,000
150,000
No Knee Airbag
N/A
1400
2000
5600
9
10
11
12
13
14
Comprehensive, magnitude, and phase errors are reported in Table A2 comparing the crash test signals to the
simulation that resulted in the most similar response. The vehicle parameters from the simulation that produced the
most similar response (comprehensive error nearer to zero) were used to tune the SVM. The graphical comparisons
between the physical crash test and the simulated crash test with the optimal restraint parameter sets are plotted in
Figures A1 and A2.
15
Table A2: Sprague and Geers error factors for the tuned SVM for the Camry and Cobalt cases.
2010 Toyota Camry
Injury Metric
16
Comprehensive
(C)
2006 Chevrolet Cobalt
Magnitude Phase Comprehensive
(M)
(P)
(C)
Magnitude Phase
(M)
(P)
Head Acceleration
0.17
-0.15
0.08
0.25
-0.23
0.10
Chest Acceleration
0.14
-0.08
0.11
0.08
0.02
0.08
Pelvis Acceleration
0.15
-0.06
0.14
0.22
0.18
0.13
Left Femur Force
0.12
-0.04
0.12
0.07
0.05
0.04
Right Femur Force
0.19
0.15
0.12
0.12
0.10
0.06
Lap Belt Force
0.04
-0.02
0.04
0.20
0.18
0.08
Shoulder Belt Force
0.13
0.10
0.08
0.07
-0.01
0.07
Total Body
0.32
0.38
T6 Accel (CFC 180) Res
Head Accel (CFC 1000) Res
60
80
C = 0.17272
= -0.15009
70 M
P = 0.085472
C = 0.14401
M = -0.087279
50 P = 0.11455
Test Data
Simulation Data
Test Data
Simulation Data
Acceleration (G)
Acceleration (G)
60
50
40
30
40
30
20
20
10
10
0
0
0.05
0.1
Time (s)
0.15
0
0
0.2
0.05
Left Femur Force (CFC 600) Z
Test Data
Simulation Data
C = 0.19275
Test Data
Simulation Data
= 0.14804
0 M
P = 0.12344
-500
-500
Force (N)
Force (N)
0
0.2
Right Femur Force (CFC 600) Z
500
C = 0.12373
M = -0.039021
P = 0.11741
0.15
(b)
(a)
500
0.1
Time (s)
-1000
-1000
-1500
-2000
-2500
-1500
-3000
-2000
0
0.05
0.1
Time (s)
0.15
-3500
0
0.2
0.05
(c)
0.1
Time (s)
0.15
0.2
(d)
Belt Upper Shoulder Force (CFC 60)
Belt Anchor Lap Force (CFC 60)
5000
7000
C = 0.043035
= -0.023703
6000 M
P = 0.03592
C = 0.12666
M = 0.10176
4000 P = 0.075413
Test Data
Simulation Data
Test Data
Simulation Data
3000
4000
Force (N)
Force (N)
5000
3000
2000
2000
1000
1000
0
0
-1000
0
0.05
0.1
Time (s)
(e)
0.15
0.2
-1000
0
0.05
0.1
Time (s)
(f)
0.15
0.2
Pelvis Accel (CFC 1000) Res
70
Acceleration (G)
C = 0.1519
M = -0.059366
60 P = 0.13982
Test Data
Simulation Data
50
40
30
20
10
0
0
0.05
0.1
Time (s)
0.15
0.2
(g)
Figure A1: Injury metric comparisons between the simulated and physical 2010 Camry crash tests.
Head Accel (CFC 1000) Res
80
T6 Accel (CFC 180) Res
60
C = 0.25279
Test Data
Simulation Data
= -0.23341
70 M
P = 0.09707
C = 0.081393
M = 0.018907
Test Data
Simulation Data
50 P = 0.079166
Acceleration (G)
Acceleration (G)
60
50
40
30
40
30
20
20
10
10
0
0
0.05
0.1
Time (s)
0.15
0
0
0.2
0.05
(a)
1000
0.1
Time (s)
0.15
0.2
(b)
Left Femur Force (CFC 600) Z
1500
Right Femur Force (CFC 600) Z
2000
C = 0.068758
M = 0.052905
P = 0.043917
C = 0.11858
M = 0.10351
1000 P = 0.057867
Test Data
Simulation Data
Test Data
Simulation Data
0
500
Force (N)
Force (N)
17
0
-1000
-2000
-500
-1000
0
-3000
0.05
0.1
Time (s)
(c)
0.15
0.2
-4000
0
0.05
0.1
Time (s)
(d)
0.15
0.2
Belt Upper Shoulder Force (CFC 60)
Belt Anchor Lap Force (CFC 60)
8000
6000
5000
C = 0.19801
M = 0.18009
P = 0.082305
C = 0.070579
M = -0.01021
4000 P = 0.069836
Test Data
Simulation Data
Test Data
Simulation Data
Force (N)
Force (N)
3000
4000
2000
2000
1000
0
-2000
0
0
0.05
0.1
Time (s)
0.15
-1000
0
0.2
0.05
0.1
Time (s)
0.15
0.2
(f)
(e)
Pelvis Accel (CFC 1000) Res
70
C = 0.22318
M = 0.1809
Test Data
Simulation Data
Acceleration (G)
60 P = 0.13072
50
40
30
20
10
0
0
0.05
0.1
Time (s)
0.15
0.2
(g)
18
Figure A2: Injury metric comparisons between the simulated and physical 2005 Cobalt crash tests.
19
Appendix B – Injury Metrics and Risk Evaluations
20
21
22
23
24
25
26
The case occupants for each CIREN case reconstruction were repositioned according to the five occupant
and vehicle position variable ranges summarized in Table B1. The CT scan of the Cobalt case occupant’s left lung
was segmented to identify injured tissue. The CT reconstruction, highlighting air, healthy lung tissue, and contused
lung tissue, is shown in Figure B1. The peak maximum principal strain in each of approximately 34,000 left lung
elements of THUMS v4.01 in each crash simulation were plotted as a function of cumulative lung volume exceeding
a given peak maximum principal strain. The peak strain of these elements in the baseline Cobalt simulation is
plotted in Figure B2. The estimated injured lung tissue volume for this case was 40.5% of the total lung volume.
27
28
29
30
31
32
Table B1: Occupant positioning parameter ranges for variation study to simulate various pre-crash occupant
positions and postures in the Camry and Cobalt cases. Steering column angle was measured between the axis of the
steering column and horizontal axis, while the steering column position was positive when moved closer to the
occupant from the mid-position. Seat back angle was measured between the back surface of the seat and the vertical
axis. The mid-track seat position was referred to as the zero position, while positive seat track positions indicated
that the seat was moved away from the dashboard. The zero position for D-ring height was the lowest anchor point.
2010 Toyota Camry
2006 Chevrolet Cobalt
Positioning Variable
Units
Min
Baseline
Max
Min
Baseline
Max
Steering Column Angle
Deg
23.2
25
26.8
17
20
23
Steering Column
Telescoping Position
mm
-21
0
+21
-25
0
+25
Seat Back Angle
Deg
-1.5
+10
+23.5
0
+6
+25
Seat Track Position
mm
-142
-71
+71
-26
+72
+145
D-Ring Height
mm
0
0
75
0
50
75
33
34
35
36
Figure B1: Approximately 49% of the Cobalt case occupant’s left lung was identified as contused lung tissue (red)
from the CT scan. Air within the pleura is shown in blue and healthy lung tissue in gray.
37
38
39
Figure B2: The baseline case was estimated to have 40.5% of the lung tissue contused, according to the maximum
principal strain threshold (0.343) identified by Gayzik et al, 2011 (Gayzik et al. 2011).
40
41
42
43
Regional-level injury metrics (HIC15 and CTI) were plotted as functions of the five occupant positioning
variables in Figure B3. The baseline simulations for each case are identified with a red star. These injury metrics
were converted to injury risk probabilities using the injury risk functions summarized in Table B2. Head injury risk
distributions, as calculated using Table B2, are plotted across all occupant positions in Figure B4 for each case.
44
45
46
47
Figure B3: HIC15 and CTI as a function of five positioning variables in the Camry and Cobalt cases. HIC15 and
CTI values in the baseline simulation that most closely matched the occupant position documented in CIREN are
indicated by the red star.
48
Table B2: Injury risk functions used for this study.
Head Injury Risk
AIS 1:
AIS 2:
(NHTSA 1995)
1
1+
e(1.54+200/HIC15 −0.0065∗HIC15)
1+
e(2.49+200/HIC15 −0.00483∗HIC15)
1
AIS 3:
1
1 + e(3.39+200/HIC15 −0.00372∗HIC15)
AIS 4:
1
1 + e(4.9+200/HIC15 −0.00351∗HIC15)
AIS 5:
1
1+
e(7.82+200/HIC15 −0.00429∗HIC15)
Chest Injury Risk
(Eppinger et al. 1999)
AIS 2:
1
1 + e(4.847−6.036∗CTI)
1
AIS 3:
1+
e(8.224−7.125∗CTI)
AIS 4:
1
1 + e(9.872−7.125∗CTI)
AIS 5:
1
1 + e(14.242−6.589∗CTI)
Knee, Thigh, Hip Injury Risk
AIS 2:
AIS 3:
(Kuppa et al. 2001)
1
1+
e(5.7949−0.5196∗Max Femur Force)
1
1 + e(4.9795−0.3260∗Max Femur Force)
49
50
51
52
53
54
Figure B4: Head injury risk (AIS 1+, 2+, 3+) as a function of HIC15 (NHTSA 1995) for all simulated positions in
the Camry and Cobalt cases. Head injury risk in the baseline simulation that most closely matched the occupant
position documented in CIREN is highlighted with darker shading.
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
Appendix C – Modeling Limitations
Due to the limited availability of occupant specific HBMs and detailed vehicle models, several assumptions
and simplifications used in this study may have influenced the results. Many of the most significant assumptions
involved the tuning of the SVM. Frontal NCAP crash tests are only performed at one speed and are not performed
on every vehicle model each year. Because each vehicle was not tested each year, sisters and clones were used to
match the crash test vehicle calibration to the CIREN case vehicle. For a given model year, the occupant restraint
systems could vary from a sister or clone vehicle despite having matching vehicle stiffness and crash pulse
characteristics. Additionally, if the simulated H3 occupant response closely matched the frontal NCAP crash test, it
was assumed that the THUMS occupant response within the same vehicle model would match a human’s response
in crash events occurring with similar velocities. It was assumed that the properties varied within the SVM tuning
LHD were not dependent upon the speed of the crash. Despite the inability to tune the SVM at varying crash speeds,
performing vehicle-specific FE reconstructions is an improvement on past studies (Danelson et al. 2015; Golman et
al. 2014).
Generating boundary conditions for the frontal crash reconstruction simulations from EDR reports resulted
in a few limitations as well. While the vehicle tuning simulations incorporated pitching of the vehicle, EDR reports
did not present pitching of the vehicle throughout the crash event and therefore the crash pulses applied to
reconstruct the CIREN crashes only incorporated the linear acceleration pulse. Additionally, low magnitude lateral
acceleration pulses during frontal crashes may not be recorded by the EDR in frontal crash events and therefore not
implemented into the crash simulation. Additionally, the current SVM used for this study is not capable of
simulating occupant compartment intrusions seen in more severe or small overlap crashes.
Much of the SVM geometry originated from a 2001 Ford Taurus, which may not be representative of the
geometry in all modern vehicles. The dashboard and knee bolster geometry may account for discrepancies in the
lower body response between the simulated and physical crash tests. Future studies may investigate the use of more
generic occupant compartment geometries within the SVM.
Using the THUMS 50th percentile male model for all case reconstructions was another limitation in this
study. The chosen scaling method was selected to maintain the same regional and organ-level mesh geometries.
There was no variation in the occupant girth or weight to create occupant-specific models. To account for significant
variation of occupant girth from the scaled THUMS model, morphing techniques could be used to modify the shape
of the case occupant or scaling could be performed on the 5 th percentile female or 95th percentile male THUMS.
Similarly, anatomical material properties remained the same for each case reconstruction. Material properties of
individual organs and bones could be modified to account for age and sex differences.