Biomedical Engineering Reference
In-Depth Information
(Li, Lopez, and Rubash 2001; Donahue et al. 2002). Other studies have investigated the changes of
cartilage stress associated with different flexion angles (Moglo and Shirazi-Adl 2003; Shirazi-Adl and
Mesfar 2005), calculated from full extension to deep flexion. The stress and strain field of ligaments
was also predicted when solid models of these parts were introduced (Pena et al. 2006).
In this chapter, we present an FE model of a knee in a kneeling position to investigate the stress dis-
tribution of cartilage, menisci, and bones. Based on this case, we discuss the FE modeling procedure
and show how these models can help us to better understand the relevant biomechanical issues.
7.2 develoPment oF a kneelIng model
MR images were obtained from a 26-year-old healthy male volunteer. Scanning was carried out
on his right knee flexed through 90 degrees. During scanning, the volunteer was asked to remain
relaxed in order to eliminate the influence of pre-stressing. The resolution of these images was
0.75 mm, with slice distance of 0.94 mm.
The images were used to build a geometrical model using the commercial software MIMICS,
including the femur, fibula, tibia, patella, cartilage, medial and lateral menisci, patellar tendon,
and ligaments. The points of attachment of the biceps and semimembranosus were also identified.
When performing the mask editing process, instead of directly forming the precise contour for each
part, an overlap of the masks was carefully created using a Boolean operation. For example, when
we constructed the femur and its cartilage, we precisely shaped their outer surfaces and created
an overlap by manually moving the inner surface of the cartilage into the femur. The reason for
this procedure was that if we created masks such as that shown in Figure 7.1b, then there would be
both overlap and gaps at the interface, that could not be eliminated by a simple Boolean subtract.
Because forming thin objects is usually more difficult, the cartilage rather than the femur was used
as the subtracted part. The .stl format files of all parts, which described the surface as triangles,
were imported into Rapidform XOR for improvement of triangle qualities and further modification.
Following these procedures, all parts were imported into ABAQUS. A Boolean operation was
carried out at this stage to avoid any possible errors caused by shifting among different software.
Then all parts were meshed by hexahedral elements, which can provide more accurate results over
tetrahedral elements with the same element size. This meshing strategy requires manual partition
of the parts. The method used can be seen in Figure 7.2.
(a)
(b)
(c)
FIgure 7.1
Geometric models were created based on magnetic resonance images. (a) MR images that are
used to create the geometric model. (b) Direct contouring of the cartilage, which may cause gaps between
cartilage and bone of the 3D models. (c) Manually moving the inner surface of the cartilage into the femur to
cause an overlap while keeping its outer surface precisely shaped. The overlapped part will be subtracted by
Boolean operation, by which all the possible gaps between cartilage and bone can be eliminated.
Search WWH ::
Custom Search