Modelling of Frontal Collisions from Onboard Recorder and Full Width Barrier Data

Abstract

Estimating vehicle collision severity is a key factor in biomechanics as it allows correlation of injuries with impact severity in real accidents. Key vehicle parameters for injury are the mean, peak and shape of the acceleration pulse. Real world frontal collision patterns reflect the distributions of overlap, impact angle, vehicle compliance and car-to-car frontal structural interactions. In this paper, a model relating displacement to crush as a function of overlap is combined with a full-width barrier characteristic to predict the acceleration-displacement response for given car types. Application yields predictions for 30 degree angled barrier and 45% overlap rigid barrier tests which closely match the corresponding experimental results. Separately, regression analysis of data from onboard collision recorders of real world collisions for 16 car types in 269 collisions yielded relationships between mean acceleration, velocity change and displacement for individual car types. This shows that the mean crumpling stress/density is equivalent to the normalised mean acceleration and is independent of size or mass. The geometric crush model combined with a single full-width barrier characteristic for the car population is used to model the variability of structural characteristics in frontal collisions. These are combined with overlap and crush profile distributions to predict overall collision characteristics using Monte Carlo methods. The results confirm that the collision recorders data reflects the distribution of frontal collisions. A new method of estimating the frontal response of individual cars at various overlaps from a single full-width barrier test is presented, with predictions that correlate closely with collision recorder data.