Multi-scale modeling of human cortical bone: aging and failure studies

At LMSSC, Cnam, Paris, January 10th 2007
Elisa Budyn (Assistant Professor, Computational Mechanics Laboratory, University of Illinois at Chicago, USA)

A multi-scale analysis for unit cells of human cortical bone is presented. Two studies are conducted: the first study concerns the effect of aging over the structural and mechanical properties of human cortical bone; the second study is devoted to the failure mechanism and the development of cracks in cortical bone under various loading conditions. Experiments are conducted on human specimen of different ages in order to measure relevant geometrical and mechanical parameters and obtain microscopic data that will be injected into finite element models. First a continuum FEM model will compute macroscopic information that will be validated through comparison with the experimental measurements. For the failure mechanism study, an XFEM model will be developed in order to allow the growth of multiple cracks until complete failure of the cell. An elastic-damage criterion will be used in order to place the initial cracks in maximum strain locations. To follow the global response of the cell, the stress intensity factors are computed at each crack tip and a load parameter is adjusted so that the stress intensity factors remain at the critical value. In the case of competitive crack tips, a stability analysis is performed by computing the second derivative of the potential energy for each crack. Fatigue loading will be also investigated. The discretization utilizes the extended Finite Element Method and requires no remeshing as the cracks grow. The crack geometries are arbitrary with respect to the mesh, and are described by a vector level set. Special boundary conditions and the algorithm for detecting crack bridging and crack entering Haversian canals which allows the cracks to grow until maximum failure and/or percolation is presented.


Laboratoire de Mécanique des Structures et des Systèmes Couplés - LMSSC