Detailed FE rib modelling for fracture prediction

Le résumé fourni par la source

The purpose of this thesis was to investigate if rib fracture can be predicted in dynamic analysis\nusing the first principal strain estimate on subject specific finite element (FE) models of human\nribs. The investigation was first to be conducted on models represented by an all hexahedral\n(all-hex) element based mesh. If the rib fracture was captured with a detailed all-hex modeling\napproach, the aim was to determine what would be the maximum level of simplification that could\nbe used in the FE model, without losing the capability to estimate the fracture location.\nSubject specific FE models were developed to reproduce dynamic end-to-end rib displacement tests,\nconducted prior to this thesis, on twelve number sixth rib specimens. Pre-test, high resolution CT\nimages of the rib geometries were taken to be individually processed with a cortical bone mapping\nalgorithm to provide subject specific cortical bone thickness. The thickness distributions enabled a\nmanual Hexa-Block meshing procedure of the cortical and trabecular bone of the ribs. The all-hex\nmeshed cortical bone was in a later step converted to, eight node thick shell elements, followed\nby quadrilateral shell elements with nodal thicknesses. Material tension tests were conducted on\ncoupons of the cortical bone to provide subject specific isotropic material properties. Isotropic\nlinear elastic heterogeneous material properties for the rib specimens’ trabecular bone were obtained\nbased on density estimates from the same CT data, though processed in an earlier Master’s\nthesis on the same set of rib specimens. In a later step, the heterogeneous material data was also\nhomogenized to a linear elastic isotropic material representation.\nFour different modeling approaches were analyzed. The most detailed model used an all-hex\nmesh with heterogeneous trabecular material properties. The first simplification step used an allhex\nmesh with a subject specific homogenized trabecular material property. In the second and\nthird simplification steps, the cortical bone was represented with thick and thin shell elements,\nrespectively. Both employing heterogeneous trabecular material properties.\nBecause of numerical instabilities unable to be resolved within the timeframe of the project, the\nthick shell modeling approach failed the energy balance assessment. Hence, the pertaining results\nwere deemed unfeasible to evaluate.\nOverall, the model validations showed that seven out of twelve rib models had a non optimized\nreaction force-displacement response which agreed well with the original experiment. In addition,\nother metrics measured in the experiment were also accurately captured for the seven ribs, independent\nof modeling approach (excluding thick shell). However, only the all-hex models accurately\ncaptured the correct fracture locations.\nThe five ribs that did not capture the force-displacement response also did not predict the correct\nfracture location. The disagreements for the five ribs are hypothesized to be owed to cortical\nporosity effects. These effects influence the quality of the coupon tension tests as well as the\ncortical bone mapping algorithm, which lack the ability to correctly represent changes in tissuelevel\nproperties due to cortical porosity.\nThe results imply that the human rib fracture is strain controlled and can be successfully captured\nin a subject specific finite element model which uses elements supporting a 3D stress state to model\nthe cortical bone. Consequently, the most simplified model to capture the fracture location was\nan all-hex model with homogeneous material properties. The results provide guidelines for further\ndevelopment of the thorax used in impact biomechanic human body models.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.