A commercially-available two-dimensional software program, validated to model high speed collisions, was extended to analyze rear-end collisions involving speed changes below 10 miles per hour. Simulation results were compared to the results of several series of published full-scale staged collisions. A total of 84 rear-end crash tests, involving 20 vehicles of different makes and models, were analyzed. Test conditions included free-rolling as well as braked vehicles, and in-line as well as oblique collision configurations.
The analysis demonstrates that the simulation model provides accurate and reliable predictions of vehicle delta-V's for rear-end collisions, under aligned and oblique conditions, and with free-rolling and braked conditions for the foam core and piston-equipped bumper types examined. Using baseline crush stiffness data derived from higher speed barrier tests, the model was found to over-predict peak acceleration values, and correspondingly, under-predict collision durations in low speed collisions. In modeling collisions with closing speeds below 15 mph, a proportionate reduction of the baseline stiffness coefficients was found to significantly improve the model's prediction of the crash pulse shape, and the corresponding peak acceleration and duration. The ability of the model to predict vehicle deformation under these crash conditions was not examined.
It is intended that this technique be used in modeling real-world collisions in conjunction with a separate closed-form momentum analysis. Subsequent to such a momentum analysis, the current technique can be used to model the vehicle speed changes predicted by momentum, and then to examine time-based aspects of the collision such as the timing of vehicle collisions and/or the acceleration time histories and peak accelerations of the involved vehicles, as well as the effects of vehicle alignment and braking.