Assessment of the skull fracture prediction capability of finite element head models

The development of drones has raised questions about their safety in case of high-speed impacts with the head. This has been recently studied with dummies, postmortem human surrogates and numerical models but questions are still open regarding the transfer of skull fracture tolerance and procedures from road safety to drone impacts. This study aimed to assess the performance of an existing head FE model (GHBMC M50-O v6.0) in terms of response and fracture prediction using a wide range of impact conditions from the literature (low and high-speed, rigid and deformable impactors, drones). The fracture prediction capability was assessed using 156 load cases, including 18 high speed tests and 19 tests for which subject specific models were built. The GHBMC model was found to overpredict peak forces, especially for rigid impactors and fracture cases. However, the model captured the head accelerations tendencies for drone impacts. The formulation of bone elements, the failure representation and the scalp material properties were found of interest for future investigation. The model still predicted a sizable proportion of skull fractures. With failure enabled, it reached a sensitivity of 86.6% and a specificity of 82.0% (n=156). With failure disabled, risk curves with a rating of good according to ISO/TS 18506:2014 were developed using the second principal strain in the outer table cortical solid elements.