This thesis examines the effects of changes in architecture on apparent mechanical properties of a rod-like trabecular bone structure representative of that found in the human vertebra. A rod-like hexagonal columnar solid model was developed, based on a repeating unit cell structure, which incorporated real architectural measures. Parametric solid models were then created to examine changes in bone volume fraction (BVF), degree of anisotropy (DA), horizontal strut thinning (HST), and vertical strut loss (VSL). Using the finite element method (FEM), vertical struts were found to be largely responsible for apparent stiffness While horizontal struts provided reinforcement of the vertical struts to prevent them from buckling under load. Apparent elastic modulus was shown to be highly dependent on the minimum cross-sectional area of the vertical struts and was, therefore, affected by changes in BVF and VSL. Ultimate (buckling) strength was influenced by all of the architectural modifications considered due to changes in vertical strut slenderness (BVF, DA, HST) or increased stress caused by reduced vertical support (VSL).
Selected solid models were manufactured from DuraForm GF material using selective laser sintering (SLS), a form of rapid prototyping, to validate the FEM models. However, the mechanical tests could not be completed due to repeated instrumentation problems. Nevertheless, preliminary results showed good agreement with the apparent modulii predictions of the FEM models. Ultimate strengths of the rapid prototypes were significantly lower than the FEM predictions because failure likely occurred by inelastic buckling, which could not be predicted by the FEM models since a linear-elastic material model had been adopted. Failure stresses and failure modes observed in the rapid prototyped samples were found to correspond with those reported for real trabecular bone. The simulated bone structures were also able to effectively reproduce several characteristics of real trabecular bone mechanical behaviour. However, variability in the sintering process presents difficulties when using SLS as a tool for validation.