Abstract
This report describes a method of evaluating the real safety of a car or a design feature by correlating data from mathematical simulations and occupant data from accidents. Extensive accident material with a certain car model is used to describe facial injuries for belted drivers in frontal impacts. In a carefully validated model, simulations are carried out at different crash test speeds and for different dummy sizes. The data from the accident material and from thesimulations are correlated with special attention to crash severity, occupant size, trajectory, and contact speed. (A)