Test system, vehicle and occupant response repeatability evaluation in rollover crash tests: the deceleration rollover sled test

Abstract

The goal of this study was to evaluate the repeatability afforded by a rollover test system in terms of the test conditions, vehicle and occupant response, and vehicle deformations. Eight full-scale rollover tests were performed using three 2002 Ford Explorer vehicles, instrumented with anthropomorphic test devices and arrays of accelerometers and angular velocity sensors, to examine both intra- and inter-vehicle repeatability in five non-destructive low-speed (LS) and three full-thrown high-speed (HS) rollover crash tests using the deceleration rollover sled method. The cart was accelerated to the target velocity (LS: 19.3 km/h, n = 5; HS: 48.3 km/h, n = 3) and then decelerated (soil trip simulation) to initiate vehicle roll. All five LS and the first two HS tests showed a high degree of repeatability (peak lateral acceleration: 5.3 ± 0.2 g LS and 5.5–6.5 g HS; peak roll rate: 134 ± 12 deg/s LS and 237.3–237.4 deg/s HS; peak curb force: 92.4 ± 2.0 kN LS and 92.6–92.9 kN HS; peak dummy head resultant acceleration: 6 ± 0.2 g LS and 45–56 g HS; peak dummy upper neck axial load: 162 ± 35 N LS and 10.7–13.6 kN HS). However, despite nearly identical deceleration pulses, the third test exhibited significantly different kinematics (237 vs. 250 deg/s, four quarter turns vs. two). These results demonstrate that the test system can generate repeatable test conditions, which result in repeatable vehicle and dummy responses, but that these responses are highly sensitive to variations in the test conditions.