Development and Validation of a 2D/3D Finite Element Model of a Composite Hemipelvis

The use of mechanical analogue composite bone models for a range of biomechanical analyses and testing procedures has grown rapidly since their introduction by Sawbones (Pacific Research Laboratories, Inc., Vashon, WA). The advantages of these composite bones over cadaveric human bones include less variability among specimens, ready availability, lower costs and ease of handling. The fourth generation of Sawbones is now commercially available, which include human femurs, tibiae, humeri and hemipelves. A number of these composite bone models have been mechanically evaluated, i.e. the femur and tibia models, but others such as the hemipelvis have been neglected. However, the composite hemipelvis has been used in several biomechanical research studies; therefore, mechanical validation of the hemipelvis is required. For this study, a robust finite element (FE) model was constructed to investigate the mechanical behavior of a composite left hemipelvis bone model. A computer tomography (CT) scan of the analogue was obtained to produce a computer aided volumetric model. This model was imported and discretized in ABAQUS (Simulia, Providence, RI). In order to reduce computational costs, two-dimensional (2D) shell elements were used to mesh the thin cortical bone layer, while the cancellous bone region was meshed with solid, three-dimensional (3D) tetrahedral elements. A series of FE tests were performed on various shell-solid element domains, to ensure the use of 2D shell elements to model the cortical layer. Once the shell-solid approach was confirmed, a FE model of the hemipelvis was constructed and validated against strain gauge data from quasi-static loading experiments. Three rosette strain gauges (Vishay Micro-Measurements, Raleigh, NC) were mounted on regions of interest along the pubic body, inferior ramus and ischium of the composite hemipelvis. The hemipelvis was fully restrained in a custom-built fixture while quasi-statically loaded using an MTS Mini Bionix II to control the application of 600 N (MTS Systems Corp, Eden Prairie, MN). Maximum and minimum principal strains were calculated from the strain gauge readings and compared to FE predictions of strain at the mounting location of the strain gauges.