Male and female WorldSID and post mortem human subject responses in full-scale vehicle tests

ABSTRACT

Objective: This study compares the responses of male and female WorldSID dummies with post mortem human subject (PMHS) responses in full-scale vehicle tests.

Methods: Tests were conducted according to the FMVSS-214 protocols and using the U.S. Side Impact New Car Assessment Program change in velocity to match PMHS experiments, published earlier. Moving deformable barrier (MDB) tests were conducted with the male and female surrogates in the left front and left rear seats. Pole tests were performed with the male surrogate in the left front seat. Three-point belt restraints were used. Sedan-type vehicles were used from the same manufacturer with side airbags. The PMHS head was instrumented with a pyramid-shaped nine-axis accelerometer package, with angular velocity transducers on the head. Accelerometers and angular velocity transducers were secured to T1, T6, and T12 spinous processes and sacrum. Three chest bands were secured around the upper, middle, and lower thoraces. Dummy instrumentation included five infrared telescoping rods for assessment of chest compression (IR-TRACC) and a chest band at the first abdomen rib, head angular velocity transducer, and head, T1, T4, T12, and pelvis accelerometers.

Results: Morphological responses of the kinematics of the head, thoracic spine, and pelvis matched in both surrogates for each pair. The peak magnitudes of the torso accelerations were lower for the dummy than for the biological surrogate. The brain rotational injury criterion (BrIC) response was the highest in the male dummy for the MDB test and PMHS. The probability of AIS3+ injuries, based on the head injury criterion, ranged from 3% to 13% for the PMHS and from 3% to 21% for the dummy from all tests. The BrIC-based metrics ranged from 0 to 21% for the biological and 0 to 48% for the dummy surrogates. The deflection profiles from the IR-TRACC sensors were unimodal. The maximum deflections from the chest band placed on the first abdominal rib were 31.7 mm and 25.4 mm for the male and female dummies in the MDB test, and 37.4 mm for the male dummy in the pole test. The maximum deflections computed from the chest band contours at a gauge equivalent to the IR-TRACC location were 25.9 mm and 14.8 mm for the male and female dummies in the MDB test, and 37.4 mm for the male dummy in the pole test. Other data (static vehicle deformation profiles, accelerations histories of different body regions, and chest band contours for the dummy and PMHS) are given in the appendix.

Conclusions: This is the first study to compare the responses of PMHS and male and female dummies in MDB and pole tests, done using the same recent model year vehicles with side airbag and head curtain restraints. The differences between the dummy and PMHS torso accelerations suggest the need for design improvements in the WorldSID dummy. The translation-based metrics suggest low probability of head injury. As the dummy internal sensor underrecorded the peak deflection, multipoint displacement measures are therefore needed for a more accurate quantification of deflection to improve the safety assessment of occupants.

KEYWORDS:

• Biomechanics
• matched pair tests
• side impact
• WorldSID
• NCAP tests

Introduction

In the U.S. Federal Motor Vehicle Safety Standards (FMVSS) and consumer information-driven Side Impact New Car Assessment Program (SNCAP) tests the European side impact dummy with rib extension (ES-2re) is used as the midsize male, and the SID-IIs is used as the small female anthropomorphic test devices (Kuppa et al. 2003; FMVSS-214, 2008). As a potential advancement of the ES-2re device and for future crashworthiness studies, since the mid to late 1990s, another midsize male dummy, termed the WorldSID-50M, has been developed. Its biofidelity has been evaluated using different methodologies (Damm et al. 2006; Hautmann et al. 2003; Kim et al. 2016; Rhule et al. 2009; Scherer et al. 2009; Sunnevang et al. 2011). In addition, injury risk curves have been developed using survival analyses procedures based on the recommendations of a task group formed by the International Standards Organization (Petitjean et al. 2012). A small female-sized WorldSID, termed the WorldSID-5F or female dummy, has also been developed, and its biofidelity has been assessed (Been et al. 2007; Carroll et al. 2013; Eggers et al. 2009). Injury risk curves are under development for this dummy.

In the sled testing condition, nearside pure lateral impact responses of the male dummy, such as chest and abdominal deflections, have been compared with the responses from post mortem human surrogates (PMHS), in matched pair tests (Maltese et al. 2002; Pintar et al. 1997; Yoganandan et al. 2002). Recent studies have compared the response of the dummy to the angled load vector under the sled testing condition with PMHS (Yoganandan et al. 2008). The ability of this male dummy to better mimic the deflection responses of PMHS has been demonstrated using differing types and arrangements of internal sensors and chest bands (Belcher et al. 2011; Yoganandan et al. 2011). While the dummy has been extensively tested in many laboratories, few matched pair PMHS full-scale tests have been conducted under standardized testing conditions (Pintar et al. 2007; Pintar et al. 2006; Tylko et al. 2006). Such evaluations are important because the ultimate goal is to use the WorldSID male and female dummies in full-scale vehicles, and conduct standardized crashworthiness tests and assess occupant safety. While these two dummies have been used in research studies around the world in sled, pendulum, and full-scale vehicle tests, their use is not promulgated in U.S. federal standards including SNCAP tests (Belcher et al. 2011; Eggers et al. 2009; Scherer et al. 2009; Tylko and Dalmotas 2005; Tylko et al. 2006; Yoganandan et al. 2011; Yoganandan et al. 2013b; Yoganandan and Pintar 2008). They are, however, intended to be used in future crashworthiness standards. Consequently, the objective of this study is to compare the responses of the male and female WorldSID dummies with those of male and female PMHS in full-scale vehicle tests.

Methods

General protocol

Matched pair tests were conducted using the male and female PMHS and male and female WorldSID surrogates according to the FMVSS-214 protocols and using SNCAP change in velocity. Fundamental data from PMHS experiments have been previously presented, and their test methods have been included in the appendix (Yoganandan et al. 2015a; Yoganandan et al. 2015b). Data from these tests were used to compare with the male and female dummy responses in similar vehicles. Full-scale vehicle moving deformable barrier (MDB) and 75-degree pole tests were conducted using restrained surrogates according to guidelines specified by the National Highway Traffic Safety Administration (NHTSA) for SNCAP tests. Four vehicles were used: two each for the MDB and pole tests, one set with the PMHS and the other set with dummies. Midsize adult male specimens were selected for the MDB and pole left front seat (driver) positions and a midsize adult female specimen was selected for the MDB left rear seat (passenger) position. The male dummy (Build level F) was positioned in the MDB and pole driver positions, and the female dummy (Build level C) was positioned in the MDB passenger or left rear seat position. The midsize sedan type vehicles were from the year 2010 and the same model (Kia Forte). All vehicles had seat-mounted torso airbags in their front seats and head curtains spanning from the front A-pillar region to the rear header-roof rail (portion of the rail close the rear seat occupant). The timing of the airbag deployments was approximately 30 ms. Torso side airbags were not present in the rear seat. The seat upholsteries were fabric. The driver seat did not have power adjustments.

Dummies

The positioning of the dummies was also based on the NHTSA SNCAP procedures for the MDB test. As the WorldSID-50M and ES-2re dummies are different, and the WorldSID-5F and SID-IIs dummies are also different, modifications to the seating procedure were made to accommodate the WorldSIDs. The male dummy was positioned in the middle of the left front seat for both MDB and pole tests, and the female dummy was positioned in the middle of the left rear seat for the MDB test. The head was level, head restraint in contact with the back of the head, and restrained using the three-point belt. The lap and shoulder portions were positioned for each dummy following the specifications for the ES-2re and SID-IIs.

Instrumentation

Five infrared telescoping rods for assessment of chest compression (IR-TRACCs) were used: three for the thorax and two for the abdomen. A 59-gauge chest band was also used on the first rib of the abdomen. The vertical distance between the h-point and the top of the first rib of the abdomen was 196 mm and 160 mm for the male and female dummies, respectively. The jacket and padding of the dummy covered the chest band. Other instrumentation included accelerometers inside the head and at the T1, T4, T12, and pelvis regions, and an angular rate sensor inside the head. All three PMHS were prepared with accelerometers on the head, spine, and pelvis. The head was instrumented with a pyramid-shaped nine-axis accelerometer package, termed a PNAP (Yoganandan et al. 2006). A triaxial angular velocity sensor was secured to the PNAP device. The spine was instrumented with a 6DX Pro (DTS, Inc., Seal Beach, CA) to record linear accelerations and angular velocities. These sensors were attached to the spinous processes of the T1, T6, and T12 vertebrae. The same type of sensor was attached to the pelvis. Chest bands (model 3055 and 4592, Denton, Inc., Rochester Hills, MI) were wrapped as follows. The first band was placed at the fourth thoracic vertebral level (approximately under the axilla, covering regions of the upper thorax), the second one was placed at the eighth thoracic vertebral level (inferior to T6 instrumentation), and the third one was at the tenth thoracic vertebral level (superior to T12 instrumentation). All chest bands had 59 gauges except for the middle chest band for the driver PMHS in the MDB test, which had 40 gauges.

MDB and pole tests

The honeycomb structure specified in the NHTSA test procedure was attached to the face of the moving cart (NHTSA 2012). The alignment of the MDB face with respect to the vehicle was such that the left edge of the honeycomb face was aligned at a distance of 940 mm forward of the centerline of the wheelbase. This alignment procedure resulted in the MDB contacting both nearside door structures. The mid coronal plane of the head was aligned with respect to the centerline of the pole, and the vehicle was positioned at an angle of 75 degrees with respect to the pole.

Data acquisition and analysis

Data were acquired from 181 and 317 channels for the pole and MDB dummy tests, and 299 and 546 channels for the PMHS tests, respectively. They were gathered according to the Society of Automotive Engineers (SAE) specifications at 20 kHz (SAE 2003). Vehicle accelerations were filtered at SAE Class 60 and PMHS accelerations were filtered at Class 180. Class 1000 filter was used to compute temporal chest deformation contours, and the methodologies are described in the following for the two surrogates.

PMHS chest band data

Contours were computed at 0.05-ms increments (NHTSA 2010). The spine and sternum locations on the pretest contour were identified based on their actual locations. The gauge closest to the spine was used as the reference to determine the origin location for contour outputs. The midpoint of the line joining the spine and sternum defined the origin. The distance from the spine to the origin was recorded. The origin was situated in all contours at the same distance from the spine to the origin along the spine–sternum line. The deflection of each gauge along the contour at each time step was defined as the change in the distance between the line joining this point and the origin in the initial pretest contour and the deformed contour. Deflection values were computed at all gauges along the impacted side. The gauge on the contour resulting in the largest change in magnitude from its initial gauge-to-origin distance determined the maximum deflection. The angle from the spine–sternum line to the line connecting each gauge to origin was measured at each point in time for each gauge, and the angle at which maximum deflection occurred was determined. The angle between the spine–sternum axis and peak deflection vector represented the angulation (Yoganandan et al. 2015a).

Dummy chest band data

Because the sternum is free to rotate relative to its spine, more than the human or PMHS sternum relative to its spine, different methods for measuring chest deflection were used for the two surrogates. A common origin was used to compare deflections measured from the IR-TRACCs and from the contours. For this new deflection measurement method, the common origin was the same as that of the IR-TRACC. Based on the location of the IR-TRACC origin for each rib, its distance along the x- and y-axes to the center of the rear face of the spine box of the dummy was obtained from engineering drawings. To locate the IR-TRACC origin using the chest band, one gauge was placed at the center of the spine during the mounting of the instrument. Assuming parallelism between the line joining the gauges on either side of the spine and the y-axis, a perpendicular line (the length of the distance between the spine and IR-TRACC origin in the x direction) was drawn from the gauge at the center of the spine toward the center of the dummy. Then another perpendicular line (the length of the distance between the spine and the IR-TRACC origin in the y direction) was drawn from the end of the first perpendicular line to the location of the origin. This series of two perpendicular lines in combination with the known x- and y-distances of the origin to the center of the spine (from engineering drawings) was used to locate the IR-TRACC origin with respect to the chest band. Since this origin represented a fixed location with respect to the spine, deflection of each gauge was calculated from this origin at every point in time. It was assumed that the three gauges along the rear face of the spine remain fixed to the spine. Since the x- and y-distances to the spine were known from engineering drawings, the three gauges along the spine were used to initially locate the IR-TRACC origin with respect to the chest band at time zero. The location of this initial IR-TRACC origin was recalculated based on two additional gauges on the side of the dummy opposite from the impact. For every temporal point after time zero, these two gauges were used to locate the IR-TRACC origin, and the calculated chest band deflections were compared to the IR-TRACC-measured deflections. Dummy maximum deflection results are presented as direct measurements from the chest band and half chest breadth-normalized deflections, equal to the percent change in length between the chest band gauge and the IR-TRACC origin on the initial and maximum deflection contours.

PMHS data

PMHS data were normalized using the equal stress equal velocity approach (Yoganandan et al. 2014b; Yoganandan and Pintar 2005). Deflections were also normalized based on one-half of the chest breadth. The normalized deflection represented the percent change in length between the chest band gauge and origin on the initial and maximum deflection contours. The initial one-half chest breadth was determined using the chest band contour origin and the same gauge used in the determination of the peak deflection.

Results

Deflection profiles from the dummy internal sensors were unimodal at t thoracic and abdominal regions (Figures 1–3). Maximum internal sensor deflections are given (Table 1). Maximum deflections ranged from 17.1 mm to 19.9 mm for the thorax, and from 24.6 mm to 30.5 mm for the abdomen ribs for the driver dummy in the MDB test. These data were respectively 7.5 mm to 32.6 mm and 17.9 mm to 18.3 mm for the passenger dummy. In the pole test, peak deflections ranged from 21.5 mm to 25.2 mm for the thorax, and 33.1 mm to 33.6 mm for the abdomen ribs.

Figure 1. Rib deflections from the IR-TRACC sensors and PMHS chest bands for the MDB-driver test.

Figure 2. Rib deflections from the IR-TRACC sensors and PMHS chest bands for the MDB-passenger test.

Figure 3. Rib deflections from the IR-TRACC sensors and PMHS chest bands for the pole-driver test.

Table 1. Peak deflections (mm) from the IR-TRACC sensors in the dummy.

Location	MDB driver	MDB passenger	Pole driver
Thorax rib 1	17.1	15.5	24.7
Thorax rib 2	19.4	7.5	21.5
Thorax rib 3	19.9	32.6	25.2
Abdomen rib 1	24.6	17.9	33.6
Abdomen rib 2	30.5	18.3	33.1

The maximum deflections for the PMHS and dummies computed from the chestband (termed as contour), are shown (Table 2). The maximum deflections from the chest band placed on the first abdominal rib were 31.7 mm and 25.4 mm for the driver and passenger dummies in the MDB test, and 37.4 mm for the driver dummy in the pole test. The maximum deflections computed from the contours at a gauge equivalent to the IR-TRACC location were 25.9 mm and 14.8 mm for the driver and passenger dummies in the MDB test, and 37.4 mm for the driver dummy in the pole test. The maximum normalized contour deflections for the PMHS were 76.0 mm, 74.1 mm, and 66.4 mm for the MDB driver, MDB left rear seat passenger, and pole driver. Table 2 also shows the normalized chest deflections based on one-half chest breadth. A comparison of time histories is shown (Figure 4) for the passenger surrogate in the MDB test. Comparisons of time histories (Figures A-1 and A-2) for the driver dummies in the MDB and pole tests are given in the online appendix.

Table 2. Deflections and compressions data from dummy and PMHS.

				Test, occupant
	Description			MDB, driver	MDB, passenger	Pole, driver
IR-TRACC		Dummy	Deflection (mm)	24.6	17.9	33.6
			Percent half chest compression	22.4	18.1	34.7
Chest band	Maxima	Dummy	Deflection (mm)	31.7	25.4	37.4
			Percent half chest compression	29.9	27.1	38.6
		PMHS	Deflection (mm)	76.0	74.1	66.4
			Percent half chest compression	43.8	54.3	40.4
	At IR-TRACC location	Dummy	Deflection (mm)	25.9	14.8	37.4
			Percent half chest compression	23.6	14.9	38.6

Figure 4. Comparison deflection-time histories from the IR-TRACC at the first abdominal rib, maximum chest band, and location corresponding to the IR-TRACC on the chest band for the MDB-passenger test.

The online appendix also includes comparisons of dummy and PMHS contours (Figures A-3 to A-8) at the initial position and maximum deflection for the dummy and PMHS. The initial and final distances from the chest band gauge to the origin are shown for maximum deflection and deflection at the gauge of the IR-TRACC location. The morphology of the contour shapes at maximum deflection matched for each test pair, with the exception of the MDB passenger dummy and PMHS. The driver dummy chest shapes showed translation and rotation. However, the passenger PMHS in the MDB test demonstrated a compressed chest shape. The static vehicle deformation profiles at the mid-door level (Figures A-9 and A-10) and the acceleration of the head, T1, T6, T12, and sacrum (Figures A-11 to A-58) are presented in the appendix. The peak resultant PMHS and dummy head accelerations were 42.2 g and 43.2 g for the front (Figure A-13) and 76.4 g and 81.6 g for the rear seat positions (Figure A-29) in the MDB test. For the pole test, they were 73.4 g and 66.1 g (Figure A-45).

The head injury criterion (HIC) and brain rotational injury criterion (BrIC) for the dummy were 212 and 0.94 for the driver, and 640 and 0.44 for the passenger in the MDB test. The HIC and BrIC values for the driver in the pole test were 491 and 0.50. For the PMHS, the HIC and BrIC were 235 and 0.76 for the driver and 508 and 0.62 for the passenger in the MDB test, and 500 and 0.36 in the pole test. The peak resultant accelerations of the thoracic spine (Figures A-17 Driver/A-33 Passenger for T1, A-20 Driver/A-36 Passenger for T4/T6, and A-23 Driver/A-39 Passenger for T12; see online appendix) were greater in the PMHS than in the dummy (MDB front seat: 64.9 g, 72.3 g, and 86.1 g versus 48.6 g, 41.4 g, and 53.7 g; MDB rear seat: 146.5 g, 120.7 g, and 119.2 g versus 85.8 g, 75.5 g, and 99.2 g; pole (Figures A-49, A-52, and A-55): 63.0 g, 51.9 g, and 62 g versus 41.6 g, 39.4 g, and 45.9 g). The peak resultant pelvic accelerations were greater in the PMHS only for the driver in the MDB test (Figure A-26; see online appendix).

Discussion

While the matched pair protocol was used to compare the dummy and PMHS responses, the number of tests limits the generalizability of the results. In addition to the inherent biological nature of the PMHS, the complexity of the experiment, that is, positioning of the dummy or biological surrogate in different vehicles due to the destructive nature of the test, adds to the variability in response. Repeated matched pair tests also introduce their own statistical variance. Because these types of tests consume significant resources in terms of equipment, personnel, cost, and time, they are not routinely done in this field. The present study is unique from this perspective. It is intended to provide insights into the responses of the surrogates, instead of serving as the full validation spectrum of the dummy response under repeated testing conditions for different vehicles.

The^{Figure 4} equal stress equal velocity approach is one method to normalize data. Other approaches are available (Yoganandan et al. 2014b). The impulse-momentum method has been used in previous sled and pendulum impact tests, wherein in forces, deflections, and velocities are recorded. A recent study comparing different methods for determining deflection response corridors from sled tests reported that the equal stress equal velocity approach produces responses very similar to the other methods, and the authors suggested the use of either method for analyzing such data (Yoganandan et al. 2014a).

The acceleration and angular velocity signals from the PMHS were noisier than those from the dummies due to the biological nature of the PMHS. The peak resultant head accelerations (Figures A-13, A-29 and A-45; see online appendix) for the biological surrogate and the dummy differed by 1 g for the MDB left front occupant, while they were 5.2 g lower for the PMHS as the left rear occupant. In contrast, in the pole test, the peak accelerations were greater by 7 g for the PMHS. These are attributed to the seating position, interaction with the internal upper structures, and the type of test. The probability of AIS 3+ head injury based on the HIC ranged from 3% to 13% for the PMHS and from 3% to 21% for the dummy from all tests, and based on the HIC, they ranged from 0 to 21% for the biological and 0 to 48% for the dummy surrogates (Hertz 1993; Takhounts et al. 2013). The decreased dummy spine accelerations relative to those of the PMHS (Figures A-17, A-20, A-23, A-33, A-36, A-39, A-49, A-52, and A-55; see online appendix) were attributed to the nature of the surrogate and instrumentation, and differences between the local segmental vertebral mass in the PMHS and the rigid spine box in the dummy. It should be noted that the lateral accelerations matched better than the vertical accelerations. Fractures occurred in both MDB-tested specimens, with higher number and bilateral involvement of rib fractures for the left rear seat location (Yoganandan et al. 2015b). Local mass recruitments also influence responses of the human body under dynamic loading (Yoganandan et al. 2014c). The frangible nature of the biological specimen in contrast to the infrangible dummy construction explains the differences in the acceleration responses at different body regions. The peak resultant sacrum accelerations (Figures A-26, A-42, and A-58; see online appendix) between the PMHS and dummy differed by 0.7 g for the MDB left front occupant, while they were 13 g and 23 g lower for the PMHS in the left rear seat and pole test. The match and differences are attributed to test type, seating position, and door interaction.

One of the purposes of this matched pair test series was to examine the behavior of the dummy using different metrics and from different body regions and assess the adequacy of internal sensors. The differences in the maximum deflections computed at the location of the abdominal IR-TRACC sensor using the contours compared to the internal sensor-measured magnitudes were less than 4 mm (Table 2 and Figure 3). These results provide confidence in the measured deflections of both methods. Based on the deflection contours, the MDB test for the driver induced lateral loading at the lower thorax (PMHS) and abdomen rib 1 (WorldSID) chest locations for both surrogates (Figure A9; see online appendix). The female dummy responded with a posterior to lateral loading pattern, while the biological surrogate responded with the left to right lateral modality. The deformed shape of the chest of the female dummy reflected translation and rotation rather than the compression response, observed in the biological surrogate. The maximum deflections occurred at the abdominal rib level for both male dummies and at the lower thorax level for the biological surrogates in both tests (Figures 1–3), that is, left front seating position in the MDB and pole tests, demonstrating a biofidelic trend in responses of the male dummy from two different tests at two different velocities and vectors. For the female dummy, however, the maximum deflection occurred at the third thorax rib, while the biological surrogate responded with the peak deflection at the lower thorax level. The difference in the location of the occurrence of maximum deflection was attributed to the seating height variations: The biological surrogate was 100 mm taller than the female dummy.

For the dummy in the pole test, the locations of maximum deflection and IR-TRACC coincided, and hence, the measured deflections for these locations on the chest band were the same, 37.4 mm. For the driver dummy in the MDB test, the gauge of maximum deflection was one gauge away from the internal sensor, and deflections measured from the contour were 31.7 mm for the maximum and 25.9 mm for the IR-TRACC. However, for the female dummy, the IR-TRACC gauge was two gauges away from the peak deflection location, and this factor accounted for the larger difference: 25.4 versus 14.8 mm for maximum and IR-TRACC (Table 2). These results indicate that the measurement is less reliable when the deflection sensor is not in close proximity to the region of loading. Consequently, a sensor with multiple points of deflection measurement should be used to obtain accurate peak deflections in vehicle crash tests. The contours at the middle thoracic level in the biological surrogate seated in the left rear position showed anterior oblique loading. These considerations lead to the choice of multipoint sensing. It should be noted that previous research has also emphasized the use of multipoint deflections to detect angular loading (Yoganandan et al. 2013a; Yoganandan et al. 2015a). Thus, the findings from the present and the cited studies suggest that a multipoint displacement sensor is needed to improve the safety assessment of occupants.

Funding

This research was supported in part by the Department of Veterans Affairs Medical Research, the Department of Neurosurgery at the Medical College of Wisconsin, and NHTSA. This material is the result of work supported with resources and use of facilities at the Zablocki VA Medical Center, Milwaukee, Wisconsin, and the Medical College of Wisconsin. The first three authors are part-time employees of the Zablocki VA Medical Center, Milwaukee, Wisconsin. Any views expressed in this article are those of the authors and not necessarily representative of the funding organizations.