This paper first checks the validity of "observed" EDSMAC damage, prior to its use as a reference in accident reconstruction, by the reconstruction of speeds and times using elementary physics, for both normal and oblique deformation. Functional peculiarities are discovered in simulations of severe fully symmetric frontal impact. The paper then checks, against known experimental impacts, the validity of the reference EDSMAC simulations established by R.L. Woolley in 1994. Functional peculiarities are discovered in those reference simulations of offset impact. Checking against the remainder of this reference family, it is shown that for frontal, corner, and side impact the CRASH methods usually give substantial and sometimes gross errors, mainly due to classical neglect of travel and rebound during impact. However, its extension, CRASHEX, gives results usually quite close to those of EDSMAC. Comparison of various approaches with the reference cases indicates that a fully developed, properly calibrated reconstruction method can substantially match the reference method of simulation at its best, while sometimes in relevance and always in speed of solution it clearly outperforms the reference. (A)
Abstract