[SY-E14] Characterization and multi-scale modeling of the mechanical response of the human humerus under dynamic loading
The relevance and biofidelity level of the human numerical models are key issues in car accidents related trauma research. To limit the risk of injuries of upper extremities and plan the preventive intervention, the humerus biomechanical properties must be correctly assessed. However, the constitutive laws used are largely derived from experimental characterizations carried out at the macroscopic scale without taking into account the bone architecture. A multi-scale approach coupled with nanoindentation experiments revealed to be more appropriate when the robustness of computation and accuracy of results are of interest.
In this study, we propose a multi-scale approach for the accurate characterization and modeling of the mechanical behavior of the human humerus under dynamic loading. The present model is based on the coupling between the Mori-Tanaka homogenization scheme, and an isotropic elastic damage model in the thermodynamic framework. In order to consider the strain rate effect on the humerus behavior, the standard model of Johnson-Cook is adopted. The obtained model is implemented using a User Material routine within the code LS-DYNA. The validity of the resulting FE model has been validated by comparing numerical predictions with experimental observations from characterization tests at different length-scales. To this end, a first experimental campaign was undertaken by means of nanoindentation tests on prismatic samples extracted directly from the same humerus diaphysis in order to determine its microscopic elastic properties. Then, local measurements of the damage effects were performed using several tension/compression and bending tests on small humerus specimens. Once the humerus material parameters were determined, a set of global validation bending impact tests composed of nine humeri were carried out using a drop tower for the determination of the mechanical response and the damage growth until complete humerus fracture.
The outcome of the proposed multi-scale model appears to correctly predict the general trends observed experimentally via the good estimation of the humerus ultimate impact load. The fracture patterns predicted by the proposed damage model are consistent with the physical humerus rupture even if this model is limited only to the estimation of the failure initiation.
In this study, we propose a multi-scale approach for the accurate characterization and modeling of the mechanical behavior of the human humerus under dynamic loading. The present model is based on the coupling between the Mori-Tanaka homogenization scheme, and an isotropic elastic damage model in the thermodynamic framework. In order to consider the strain rate effect on the humerus behavior, the standard model of Johnson-Cook is adopted. The obtained model is implemented using a User Material routine within the code LS-DYNA. The validity of the resulting FE model has been validated by comparing numerical predictions with experimental observations from characterization tests at different length-scales. To this end, a first experimental campaign was undertaken by means of nanoindentation tests on prismatic samples extracted directly from the same humerus diaphysis in order to determine its microscopic elastic properties. Then, local measurements of the damage effects were performed using several tension/compression and bending tests on small humerus specimens. Once the humerus material parameters were determined, a set of global validation bending impact tests composed of nine humeri were carried out using a drop tower for the determination of the mechanical response and the damage growth until complete humerus fracture.
The outcome of the proposed multi-scale model appears to correctly predict the general trends observed experimentally via the good estimation of the humerus ultimate impact load. The fracture patterns predicted by the proposed damage model are consistent with the physical humerus rupture even if this model is limited only to the estimation of the failure initiation.