The objective of this study is to apply computer modelling, supplemented by studies of "hard copy" documentation of rollover accidents and crash tests, to access the influence of the crash severity indicators. The crash events in tripping, pure roll, and first vehicle impact are separated and analyzed independently. The STI Tripped Rollover Model suggests that roll rate is strongly related to intitial velocity for tripped rollover. The roll rates predicted by the STI model were in excess of 5 rev/sec for the range of crash speeds which account for 50% of the rollovers. However, analysis of the National Accident Sampling System (NASS) rollover data suggests that roll rates above 2.5 rev/sec are rare. The high tripping forces permitted by the STI model may account for this disparity. The pure roll simulations show that high roll rates, independent of other impacts, can produce conditions to induce ejection. The simulations of first vehicle impact provide an insight into those variables which influence roof crush and velocity loss. It appears that roof crush and roll velocity loss are largely independent of lateral velocity. Additional data collection elements will be required to evaluate the significance of the tripping mechanism, the roll rate and the vertical velocity on rollover crash severity.
Abstract