Human Body Models (HBMs) provide increased biofidelity and additional measurement capabilities compared to dummies. However, at present injury prediction is mostly limited to binary assessments based on physical thresholds and therefore doesn’t allow for a population‐based probabilistic estimation. In this study, different Injury Risk Curves (IRCs) predicting pelvic fractures were developed for the Total Human Model for safety (THUMS) v4.02 AM50. These IRCs are based on two types of metrics, namely global metrics and local metrics. Global metrics are based on force measurements whereas local metrics are based on strains. A large set of post‐mortem human subject (PMHS) experimental tests was reproduced with the THUMS v4, then the different metrics were measured and correlated to the injuries observed in the experiments. Different indicators were used to assess and compare the fracture prediction capabilities of the different IRCs.
Keywords:
Pelvic fracture, Injury Risk Curve, THUMS, Side impact, Human body models