Earlier studies by the authors have proposed separating rollover crashes according to belt use, ejection status, and single vs. multiple harmful events. These different classifications were associated with differences that could substantially alter the risk of serious injury. For each classification, metrics to characterize rollover severity were presented. For most single vehicle crashes, the number of roof contacts with the ground was found to predict injury risk. For multi-harmful event crashes the extent of damage caused by the most severe non-rollover harmful event, combined with the number of roof impacts was found to predict injury risk.
This paper examines NASS/CDS 1995-2003 to determine the injury distribution by body region for the most frequently occurring rollover classifications that result in MAIS 3+ injuries from sources inside the vehicle. The examined classifications of rollovers include: belted not-ejected and unbelted not-ejected. For each category the injury patterns by body region were presented. Differences in injuries in near-side and far-side rollovers were evaluated.
In general, head injuries were the most frequent MAIS 3+ injury for belted occupants. However, trunk injuries were more frequent for belted occupants in near-side rollovers. It was found that a higher fraction of severe injuries occurred in far-side rollovers compared to near-side rollovers. This tendency held for rollovers with one roof impact or less as well as higher severity rollovers.
The frequency of injury and ejection for near and farside rollovers was examined. The MAIS 3+ HARM distribution by body region was examined as a function of number of roof impacts and direction of roll for not ejected front seat occupants. About 46% of the occupants were exposed to far-side rollovers, but more than half of the injuries occurred in far-side rollovers.
To examine occupant kinematics in injury producing rollovers, a MADYMO 6.1 model of a front occupant compartment of a mid-size SUV with a belted Hybrid III dummy was used. The model was validated against an available staged test with a similar configuration.
Computer modeling suggest that a higher tripping acceleration results in higher roll rates which, in turn, can lead to increased number of roof impacts. Associated with the increase in roll rate was an increase in the maximum head velocity.
The data analysis and computer modeling suggest the need to assess the severity of the vehicle loading that causes the vehicle to rollover. The severity of the tripping forces may be related to the risk of injury.