The relationship between the accuracy of results and the number of material properties used in finite element analysis of the human tibia

摘要: In finite element (FE) analysis of long bones it is now common practice to calculate the material properties based on CT data. Although a unique property calculated for each element, assigning an individual results in excessively large models. To avoid this, usual group elements their and assign single (Zannoni 1998). No study has analysed effect number used bone FE model accuracy results. The aim this was evaluate variation mechanical environment as function model. An mesh cadaveric human tibia containing 47,696 ten-node tetrahedron 75,583 nodes created using scans. Material were previous work (Rho 1995, 1996). Eleven models by varying groups (1, 2, 4, 8, 16, 32, 64, 128, 256, 512, 1024) divided into. A assigned group. All subject axial point load 300N applied medial condyle tibial plateau while distal end fixed. maximum minimum principal strains deflections, at 17 well distributed surface 65 randomly within plotted against groups. total strain energy also errors strain, deflection, assuming that 1024 accurate. parameter converge with least energy. At 512 error less than 0.001% (0.7% two model). next displacements. Using materials displacement (4.7% 2 model), internal 0.53% (36.7% convergence occurred strains. when 1.06% (57.7% model) 3.02% (104.5% respectively. This demonstrates relationship between While will allow analyst make informed decision modelling helps examine validity studies which, usually due limited resources, fewer properties.