Biomedical Engineering Reference
In-Depth Information
liver was segmented from CT data and a physical model of the liver was produced
using rapid prototyping. Then a duplicate of the model was cast from soft silicone
gel (see e.g., [
15
]). The elastic modulus of the gel was chosen to approximately
match the elastic modulus of a porcine liver. In order to be able to track deformation
in a CT scan, small marker balls with a diameter of 1.6 mm were arranged directly
beneath the surface and throughout the volume of the model.
In order to build a FE model from the phantom, a CT image of the liver phantom
was semi-automatically segmented using a region-growing algorithm. Subse-
quently, a quadratic tetrahedral mesh was generated using the commercial software
package 3-Matic 5.1.
4 Results and Discussion
4.1 Convergence Analysis
A convergence is performed in order to compare the performance of the quadratic
corotated FEM with the conventional linear corotated tetrahedral formulation. We
consider two simplemodel problems for the analysis (see Fig.
1
). The first deformation
pattern includes a large rotational component (bending of a 2
3
20 cm beam).
Fig. 1 Beam model (
upper left
) and cube model (
lower left
) considered for the convergence
analysis.
Right
: Logarithmic RMS error in terms of DOF for linear tetrahedra (
dashed line
) and
quadratic tetrahedra (
solid line
) for the beam bending (
round markers
) and the cube stretching
(
rectangular markers
) simulation
