Arc far Side Impact Collaborative Research Program – Task 5b: Test Procedures Crash Tests and Sled Tests for the Far-side Environment

6.6Simulation Results

Table 1 presents the final sled and instrument panel angle for each test modelled in this study. Although the moving barrier in the SNCAP test had a direction of travel 63 degrees to the struck vehicle centreline, the appropriate sled angle was almost 90 degrees (87^o). The Y damage tests had the same sled angle (87°), but required a rotated interior that resulted in the greatest total angle of 95.4°.
Table 1. Summary of Sled and Instrument Panel (IP) Angle

TEST CONFIGURATION	TEST IMPACT ANGLE	SLED ANGLE	IP ANGLE	TOTAL ANGLE
63^o SNCAP – 25.8 k/hr	90°	87°	0.0°	87.0°
90^o Y Damage ~ 26 k/hr	90°	87°	8.4°	95.4°
30° Corner Impact - 57.6 k/hr	30°	43°	8.5°	51.5°

The instrument panel angle is equal to the vehicle rotation when the event is considered ‘over’ in each case. For this reason, it is dependent upon the location that the force of impact acts and the moment that it places on the vehicle with respect to the vehicle’s center of gravity [7]. This effect is observed by comparing the SNCAP and Y damage tests, which both contact the side of the vehicle. The MDB in the SNCAP test contacts the subject vehicle near its center causing the PDOF to pass near the vehicle center of gravity. The bullet vehicle in the Y damage test contacts the target vehicle in the front of the vehicle creating the 8.4° of rotation observed in the test. It follows that the amount of energy transmitted by the bullet vehicle in a Y damage crash configuration could influence the sled configuration.

The total angle for each test configuration is composed of the sled angle, or angle of the sled buck, and the instrument panel angle, which is equal to the amount of rotation experienced by the vehicle during the event. An increase in the impact angle of the bullet vehicle to the target vehicle typically increases the sled angle, but a direct correlation can not be made between the two angles.

For the three cases simulated, the occupant kinematics of the sled tests accurately mimicked the occupants of their corresponding crash tests. The sled test configuration determined from the lap belt tests was also found to be the proper configuration for occupants wearing 3-point belt restraints.

The accuracy of the sled test occupant excursion is greatly dependent upon the amount of vehicle rotation in the crash test. This is seen in the Head Excursion Cross Plots as the sled tests occupants move in a relatively linear motion while the vehicle rotates underneath the crash tests occupants. The Head PDOF plots reiterate these findings as the Sled curves remains relatively constant and the Pulse curves change dependent on the amount of vehicle rotation. In all cases the Head PDOF of the Sled and Pulse curves are nearly identical when the event is considered ‘over’ for the crash mode.

6.7Conclusions

The results from this study show that a sled test can effectively mimic the kinematic response of a far-side occupant in the side and corner impacts for which there was crash test data. The impact angle, initial vehicle velocities, delta-V, and impact location are all important factors in developing a proper sled test configuration. However, no direct correlations were able to be determined between a single variable and the sled pulse or correct sled or instrument panel angle to be used in the test configuration. Of equal importance to each of these variables is how the structure of each vehicle interacts with one another. This interaction is highly dependent upon the previously mentioned variables but can drastically change with small variations in the impact conditions.

Although the results from this study conclude that crashes are too complicated to correlate a single input variable to the sled configuration, certain sled configurations may accurately represent a wide range of crash environments. For example, a sled test angle of 87 degrees appeared to be appropriate for both the SNCAP and the Y-damage test conditions that were simulated. The use of the approach documented by this paper will provide a basis for determining the best sled configuration to mimic any far-side crash condition.

6.8References

Alonzo, B., 2005, “Validation of the MADYMO Human Facet Model for Far-Side Crash Simulation”, Report to the National Crash Analysis Center, GW University.

Bostrom, O., Fildes, B., Morris, A., Sparke, L., Smith, S., and Judd, R., 2003, “A Cost Effective Far Side Crash Simulation,” UCrash 8 (3) pp. 307-313.

Bundorf, R., 1996, “Analysis and Calculation of Delta-V from Crash Test Data”, SAE Paper No. 960899.

Cheng, P. and Guenther, D., 1989, “Effects of Change in Angular Velocity of a Vehicle on the Change in Velocity Experienced by an Occupant during a Crash Environment and the Localized Delta V Concept,” SAE Paper No. 890636.

Cuadrado, J., “Comparison Of Crash Pulses And Sled Responses For Far-Side Impacts”, Master’s Thesis The George Washington University, March 6, 2008.

Digges, K. and Dalmotas, D., 2001, “Injuries to Restrained Occupants in Far-Side Crashes” Proceedings of the 17^th ESV Conference.

Fay, R., Raney, A., and Robinette, R., 1996, “The Effect of Vehicle Rotation on the Occupants’ Delta V,” SAE Paper No. 960649.

Gabler, H., Digges, K., Fildes, B., and, Sparke, L., 2005, “Side Impact Injury Risk for Belted Far Side Passenger Vehicle Occupants,” SAE World Congress, SAE Paper No. 2005-01-0287.

“MADYMO V6.2 Theory Manual,” TNO MADYMO BV, 2004

Marine, M. and Werner, S., 1998, “Delta-V Analysis from Crash Test Data for Vehicles with Post-Impact Yaw Motion,” SAE Paper No. 980219.

Palmer, S., 2006, “NHTSA’s Final Ruling for Automotive EDRs Will Revolutionize Auto Insurance,” Injury Sciences, LLC.

Roberts, V. and Compton, C., 1993, “Relationship Between Delta V and Injury”, SAE Paper No. 930311.

Smyth, B and Smith J., 2007, “Developing a Sled Test from Crash Test Data. SAE World Congress,” SAE Paper No. 2007-01-0711.

Stolinski, R., Grzebieta, R., Fildes, B., Judd, R., Wawrzynczak, J., Gray, I., McGrath, P., and Case, M., 1999, “Response of Far Side Occupants in Car-to-Car Impacts with Standard and Modified Restraint Systems Using H-III and US SID”, International Congress & Exposition, SAE Paper No. 1999-01-1321.

7Findings of Studies to Determine Crash and Sled Test Conditions

7.1Summary of Study Objectives

An objective of this task was to study issues associated with far-side sled testing and crash testing using computer simulation.

Chapter 1 conducted finite element simulations of vehicle and moving barrier crashes into a Ford Taurus. A principal purpose was to determine the degree to which the barriers produced damage patterns that were similar in shape and extent to the vehicle impacts. A secondary purpose was to determine the damage patterns for a frequently occurring crashes at 60 degrees and crashes that impact the front wheels, producing Y-damage.

Chapter 2 conducted MADYMO simulations to determine a dummy or human model that best matched the kinematics of a cadaver that was tested in a far-side crash. A principal purpose was to establish a validated model for use in evaluation occupant response in the far-side crash environment. A second objective was to compare existing dummy models to the cadaver kinematics.

Chapter 3 used the MADYMO human facet model to explore a restraint issue associated with sled testing in far-side crashes. The cadaver test program summarized in Task 2 required a console to provide restraint to the pelvic region. The height of the console should be sufficient to provide pelvic restraint without causing excessive loading of the abdomen. The modeling effort in Chapter 3 explored the occupant kinematics when subjected to consoles of varying heights.

Chapter 4 used the MADYMO human facet model to explore a crash pulse issue associated with sled testing in far-side crashes. The cadaver test program summarized in Task 2 required a crash pulse that was representative of a vehicle-to-vehicle crash. However, the test sled had limited capability to duplicate the highest accelerations that were observed in a crash test. The MADYMO model was used to determine an appropriate sled crash pulse.

Chapter 5 used MADYMO models to explore the interaction between five different dummy models and five different countermeasures. The dummy models were:

1. 
   Hybrid III
   
2. 
   BioSID
   
3. 
   EuroSID I
   
4. 
   EuroSID II
   
5. 
   SID2s
   

With these models, five vehicle configurations were simulated:

1. 
   Baseline (just vehicle interior)
   
2. 
   Reverse Belts
   
3. 
   Base with chest airbag
   
4. 
   Base with shoulder airbag
   
5. 
   Base with chest and shoulder airbag
   

Chapter 6 used data from crash tests and MADYMO occupant modeling to explore the relationship between the crash PDOF as documented in the accident data and the test sled configuration required to simulate the same far-side crash environment.

Although not part of the far-side project, a complementary series of six crash tests with both near-side and far-side dummies was conducted by The Australian Department of Infrastructure [Newland, 2008]. One purpose of the tests was to examine dummy to dummy interaction.

7.2Results

The Finite Element Models used in Chapter 1 indicate that the IIHS barrier produced similar damage on a Taurus patterns to those produced by a full size pickup truck. The NHTSA barrier and the Y damage test produced less damage to the Taurus front door than the IIHS barrier.

The average CDC extent of damage produced in actual IIHS crash tests is considerably less that the average extent of damage to vehicles with far-side occupants injured at the AIS 3+ severity. This result suggests that the barrier test speed of a far-side test should be higher than the speed used by IIHS in their side impact tests for consumer ratings.

The MADYMO modeling effort summarized in Chapter 2 found that the human facet model matched the cadaver kinematics very well and it was considered suitable for evaluation of occupant motion in the far-side crash that was simulated. The MADYMO dummy models (Hybrid III, BioSID, EuroSID 1, EuroSID2 and SID2s) did not accurately reflect the motion of a human cadaver under the same impact configurations as the cadaver test.

The conclusion of Chapter 3 was that the force exerted by the occupant on the center console increased as the height of the center console increased. However, when the center console remained low, the belt restraint system restrained the pelvis rather than the center console. As the center console height increased above 8 inches, it loaded the occupant’s abdomen and ribs

The modeling documented in Chapter 4 shows that a square wave sled test pulse can produce similar occupant kinematics to the desired car-to-car pulse.

The modeling research reported in Chapter 5 found that side-impact dummies, designed primarily to test for occupant injuries on the near-side are limited in their ability to emulate occupant kinematics for a far-side impact. Also a Hybrid III test device, designed primarily for frontal impacts, is limited in its ability to test for far-side impacts. Five anthropomorphic test devices (ATDs) were simulated in MADYMO for far-side impacts and all failed to produce desired kinematics. However, the human faceted MADYMO model did show promise by properly reproducing occupant kinematics.

Despite the short-comings of the dummies for reproducing far-side kinematics, the reaction of these dummies in MADYMO to certain countermeasures offers some insight into future studies. A reverse 3-point seatbelt effectively restrained the occupant, however, significantly increased neck force levels, almost crossing injury threshold levels. Chest and shoulder airbags on the inside of the occupant contained the occupant and prevented excursion. However, left the head and neck unrestrained and showed awkward movement of the head. In addition, the use of a petite dummy exposed some vulnerability of odd sized occupants to airbags.

The results from this study reported in Chapter 6 show that a sled test can effectively mimic the kinematic response of a far-side occupant in the side and corner impacts for which there was crash test data. It is important to understand that a side impact with a 60 degree PDOF does not translate to a sled test with a sled angle of 60 degrees! For example the SNCAP test has a direction of force of 63 degrees but an appropriate sled angle of 87 degrees. The difference is due to the rapid rotation of the struck vehicle to align with the striking vehicle.

The impact angle, initial vehicle velocities, delta-V, and impact location are all important factors in developing a proper sled test configuration. However, no direct correlations were able to be determined between a single variable and the sled pulse or correct sled or instrument panel angle to be used in the test configuration. The use of the approach documented by this study can provide a basis for determining the best sled configuration to mimic any far-side crash condition.

Tests conducted in Australia have shown that the presence of a far-side dummy does not interfere with the side protection measurements made by the near-side dummy. However, there was interaction between the near-side and far-side dummies during the rebound of the near-side dummy. The interaction occurred well after the far-side dummy slipped out of the shoulder belt [Newland 2008]. These test results indicate that it may be feasible to incorporate a far-side dummy in the vehicles used in consumer information tests. Such a proposal was presented at the ESV meeting in June, 2009 [Digges 2009]. An analysis of safety belt geometry and pretensioning indicates that relatively simple countermeasures can improve far-side protection [Eschemendia 2009].

7.3References

Digges, K., Echemendia, C., Fildes, B. and Pintar, F.,“A Safety Rating For Far-Side Crashes”, Paper Number 09-0217, Proceedings of the 21^st ESV Conference, June 2009. .

Echemendia, C., “PMHS Far-Side Sled Test Validation with MADYMO”, Report to the National Crash Analysis Center, GW University, March 2009. .

Newland, C., Belcher, T., Bostrom, O., Gabler, H., Cha., J., Wong, H., Tylko, S., and Dal Nevo, R., “Occupant-to-Occupant Interaction and Impact Injury Risk in Side Impact Crashes”, Stapp Car Crash Journal, Vol 52, pp..327-347, 2008. .