Search Results

A friction induced rollover event generally comprises successive stages of control loss, lift-off, transition from lift-off to launch, and roll to rest. This transition is, of course, not instantaneous due to vehicle inertia. It is analyzed by numerical integration of the equations of motion to determine roll angle and time to launch, plus roll velocity at launch, as well as intermediate values of roll angle, roll velocity and surface force versus time. Representative curves and launch values are presented. The time to launch, which is several times greater than typical vehicle collision contact time, is seen to be dependent on the magnitude of surface friction; while roll angle from over-center at launch and roll velocity at launch do not vary substantially with changes in vehicle and surface conditions. It is also seen that the surface force initially rises above static weight, producing lateral force greater than the friction coefficient times vehicle weight as rotation progresses.

It is well understood and supported by more than a decade of statistical investigation that vehicles involved in on-road rollovers are generally out of control (e.g. spinout, side slide) prior to the roll, and that lateral stability is the primary vehicle parameter governing such occurrences. Although a rollover is unstable from lift-off to tip-over, an impulse of sufficient duration is required to input the work necessary to lift the center of gravity to a position over the leading wheels. The force in this impulse is, absent a geometric trip, limited to friction between the vehicle tires and the travelled surface. The coefficient of friction of a tire sliding laterally on dry pavement is generally at or below 0.9, although in certain conditions of unusual tire shoulder wear somewhat higher friction has been observed. Therefore vehicle side stability above this level (0.9 at GVW) is a reasonable design criterion for light vehicles.

The well known vehicle planar collision model having constant stiffness is reexamined with the inclusion of elastic rebound after reaching common contact velocity. Quantitative analysis of ground friction and windage shows that they can generally be disregarded for collision impulses involving significant barrier equivalent velocity. On the other hand, elastic rebound is significant at a 30 mph barrier equivalent velocity. Comparisons of the inputs to analyses which include crush energy and the inputs to analyses which consider only impulse-momentum show that the principal direction of force and the intervehicle coefficient of friction are equivalent estimates. Further, the energy correction for a non-normal impulse in CRASH3 is seen to be an assumption of isotropic stiffness, while the stiffness coefficient commonly described as preload describes the elastic deformation.