Mesh quality improvement by re-triangulation surface meshes on 3-dimensional CAD model of a human wrist joint

With the recent advances computing technologies nowadays, reverse engineering is rapidly developing and has been strongly established many years in biomedical industry. Three dimensional geometric modeling is now being extensively used in many applications such as surgical planning, rapid prototyping, medical implants design, numerical simulation and etc. Thus, mesh quality is very crucial especially in finite element analysis in order to produce high accuracy results. Re-triangulation surface method is employed to improve the mesh quality of the model with certain parameters. The purpose of this study are a) to construct a complex three dimensional geometric CAD model of human wrist joint which contains cortical and cancellous bones using image-based processing method, b) to investigate the effects of triangle reduction, mesh smoothing and size of triangle mesh on the accuracy and mesh quality of the wrist bones model and c) to optimize the mesh quality using response surface methodology (RSM). A three dimensional CAD model of a human wrist joint was constructed after conversion model to non-uniform rational B-spline (NURBS). Re-triangulation process was carried out by manipulating different values of the parameters like geometrical error, smoothing factor and control edge length of triangle mesh along with mesh quality analysis in order to get high quality of mesh on the triangular surface wrist joint model. Lastly, optimization was performed via RSM to optimize the high quality mesh on the triangular surface model. In the findings, triangle reduction with geometrical error of 0.05 mm had reduced the number of triangles of cortical and cancellous bones to 74.81 % and 75.34 % respectively while obtained 70.05 % and 68.69 % of high quality of mesh on the surface model respectively. In term of accuracy, no significant changed in surface area and volume. For mesh smoothing effect, smoothing factor of 0.8 was selected to obtain 87.53 % and 86.68 % of high quality of mesh on cortical and cancellous bones respectively. For control size of triangle mesh, 0.5 mm of control edge length of triangle was taken to produce 92.35 % and 92.09 % for cortical and cancellous bones respectively. In term of accuracy of the model, the surface area and volume of the cortical and cancellous bones had no significant changed with less than 0.15 % and 0.25 % compared to the initial bones model. In RSM optimization, the optimum mesh quality on cortical and cancellous bones were 90.74 % and 89.67 % respectively. In conclusions, the mesh quality and accuracy of the model show promising results with the process of re-triangulation surface and the high quality mesh of optimum models were obtained using RSM.

Text 114737.pdf Download (908kB) Official URL or Download Paper: http://ethesis.upm.edu.my/id/eprint/18178

Actions (login required)

View Item