NHTSA Roof Crush Study: Greater Intrusion = Greater Injury

The recently released NHTSA study, “Roof Crush Analysis Using 1997-2001 NASS Case Review,” which has been in the works for nearly four years, reviewed NASS rollover crashes in an effort to determine whether roof deformation patterns identified in an earlier agency study were still valid for current vehicle designs (see “Upgraded Rollover Roof Crush Protection: Rollover Test and NASS Case Analysis” NHTSA, June 1992). The new study, which examined 273 rollovers from 1997-2000 NASS, also examined the frequency of roof deformation patterns to determine if differences existed among various vehicle classes. Patterns were compared to results of 216 tests, tests with extended crush limits, SAE J996 drop tests, and the extent that windshields provided roof support.

The major findings indicate that greater roof intrusion correlates to greater injury and the damage patterns in these real-world cases don’t correlate to the baseline 216 test, but they do show similarity to the extended crush limits and drop tests using the 216 angle. Notably, the agency stated that neither of these test procedures address dual side loading and that they were unable to determine the extent to which the windshield contributed to the roof strength in rollover. Both of these issues are important as testing both sides of the vehicle would require higher roof strength, and testing with windshields in place is estimated to add as much as 25 to 30 percent of to the roof strength.

NHTSA’s findings will form the basis for the long-awaited proposed rulemaking to improve the roof crush resistance standard, which is a relic from the 1970s and was intended as a temporary stopgap to dynamic rollover testing that was unanimously opposed by the industry. Based on these findings, NHTSA is expected to propose a standard that will simply strengthen the current static requirement.

Study Details:

The 1995-2001 model vehicles included in the study weighed less than 10,000 lbs., rolled at least two quarter turns, were reported to have at least 6 inches of vertical intrusion. Vehicles pulling trailers or modified from the OE design were not included. The case selection included 95 passenger cars, 101 SUVs, 66 pickup trucks, and 11 minivans. [List of vehicles included-Appendix A]. Each NASS case file was also examined against the prior NHTSA patterns.

The damage patterns observed in this study were similar to the previous NHTSA work. A- , C- and D-pillars remained largely straight with bending occurring at or near both ends. Similarly B-pillars tended to bend at or near the ends, but a significant number of cases involved bending at the mid-point. Roof damage frequently included deformation near the A-pillar to roof junction and a number of cases resulted in roof “matchboxing” damage. Roof damage frequently included a deformed area, usually near the A-pillar roof junction. The report notes that more damage occurred on the side opposite the first vehicle to ground impact. Also noted was that this analysis could not determine the affect of windshield support in the crash. However, in the 24 percent of the cases in which the windshield was intact, less than half of those windshields were capable of providing post crash roof support.

In a comparison of roof damage to occupant injury severity the study found that there were more occupants with MAIS 3+ injuries in cases with less than average roof crush (30.8 cm), it appeared that something other than roof crush was associated with the injuries. This led to an examination of pre- and post-crash headroom. The data was reexamined to consider “negative headroom,” which was calculated as the mean headroom from the total intrusion. Once the cases were examined and those with improper intrusion reporting were eliminated, 196 cases were left showing that lower intrusion levels correspond to less injury.

Finally, the comparison between the NASS cases examined and the FMVSS 216 test and SAE J996 drop test using the 216 angle. It was noted that the current baseline 216 test does not produce damage levels comparable to the crashes in this study. As a result, comparisons were made with extended FMVSS testing and corresponding drop tests for the 10 and 15 inch crush levels (as noted in NHTSA’s 1998 report “Quasi Static and Dynamic Roof Crush Testing”). According to NHTSA, this comparison showed damage patterns that were generally comparable.

Copyright © Safety Research & Strategies, Inc.