Objective biofidelity rating of a numerical human occupant model in frontal to lateral impact.

Both hardware crash dummies and mathematical human models have been developed largely using the same biomechanical data. For both, biofidelity is a main requirement. Since numerical modeling is not bound to hardware crash dummy design constraints, it allows more detailed modeling of the human and offering biofidelity for multiple directions. In this study the multi-directional biofidelity of the MADYMO human occupant model is assessed, to potentially protect occupants under various impact conditions. To evaluate the model's biofidelity, generally accepted requirements were used for frontal and lateral impact: tests proposed by EEVC and NHTSA and tests specified by ISO TR9790, respectively. A subset of the specified experiments was simulated with the human model. For lateral impact, the results were objectively rated according to the ISO protocol. Since no rating protocol was available for frontal impact, the ISO rating scheme for lateral was used for frontal, as far as possible. As a result, two scores show the overall model biofidelity for frontal and lateral impact, while individual ratings provide insight in the quality on body segment level. The results were compared with the results published for the THOR and WorldSID dummies, showing that the mathematical model exhibits a high level of multi-directional biofidelity. In addition, the performance of the human model in the NBDL 11G oblique test indicates a valid behavior of the model in intermediate directions as well. A new aspect of this study is the objective assessment of the multi-directional biofidelity of the mathematical human model according to accepted requirements. Although hardware dummies may always be used in regulations, it is expected that virtual testing with human models will serve in extrapolating outside the hardware test environment. This study was a first step towards simulating a wider range of impact conditions, such as angled impact and rollover.