Page 278 - Biomedical Engineering and Design Handbook Volume 1, Fundamentals

FINITE-ELEMENT ANALYSIS 255

dental bridges. In one study, Lin et al. (2003) evaluated the stresses between a prosthetic tooth and
its abutting native teeth. This 3D study utilized CT scanned geometries and isotropic material prop-
erties for the prosthesis and the native tooth enamel, dentin, and pulp. No attaching structures (bone,
ligament) were evaluated in this study. A study by Magne et al. (2002) evaluated a 2D model of
partial dentures, between two abutment native teeth, which included the periodontal membrane and
supporting bone. The cortical and cancellous bone, ligament, enamel, and dentin were assumed to
be isotropic. A number of implant materials, ranging from gold to fiber-reinforced composites, were
evaluated. A final example paper from Holberg et al. (2005) utilized a 3D FEA model to evaluate the
consequences of corrective facial surgery, including the jaw. Their simulations were based upon
patient-specific scanned images, and utilized isotropic properties for soft tissue modeling.

10.5.2 Case Studies Conclusions
From the brief outline above, it can be seen that FEA and CFD are being used in all aspects of med-
icine to great advantage. While touching on a few areas of significant interest, there is no area of the
body or function that does not lend itself to computer modeling. The major issues when comparing
the various models include how far to accept the simplifications. While a linear, 2D, steady or static,
isotropic model can still be quite useful, most studies incorporate at least some nonlinearity. The bulk
of the fluid models evaluate pulsatile flow, and many include the non-Newtonian behavior of blood.
Solid models often use isotropic material properties, while the issue of 2D versus 3D seems to be on
a case-by-case basis. As with all modeling, the bottom line is to be able to produce a validated model
that will successfully mimic or evaluate the clinical situation.
As can be seen from these examples, multiple generations of modeling are often used to evaluate
all the aspects of biological systems. Many simplifications are made initially to determine a first-
generation solution. The number of subsequent model generations will vary, depending on the sophisti-
cation of results desired. At this point in time it is perhaps unlikely that all nonlinear parameters of
the human body can be included in an FEA model, mainly due to the fact that all properties are not
yet known. Comparison with bench and clinical findings will demonstrate, however, that similar
behavioral results are possible.
The researcher should not be overly concerned with the minutiae of the model parameters.
Considering the variances between people, an overview covering most of the noted effects should be
the goal. The purpose of the models is to examine system behavior when changes are made, such as the
effects of geometric and material modifications. Although the actual behavior may not be exact, the
variances due to changes in the model may well mimic those of the device or human system.

10.6 CONCLUSIONS

The main points to be taken from this chapter are that many simplifications must be made initially
when evaluating human-related structures with FEA, but this may not be a major problem. Each of
the three main areas (geometry, material properties, and boundary conditions) is of equal importance
and should be considered separately. Depending on the detail of the construct modeled, many geo-
metric components may be simplified. Care should be taken when using isentropic models for solids,
but even these may be used for general evaluations of tissue regions. Soft tissues are difficult to fully
model without the use of nonlinear properties. Newtonian flows are reasonable for large-vessel flows
(such as the aorta or carotid artery) as long as the aspects of flows have been shown not to be affected
by the particular components. Boundary conditions may begin with steady-state values, but dynamic
components will add the complexities of the true system.
In closing, a computer model of portions of the human body or an added component may be as
simple or complex as the user desires. One should use properties to be found in this and other sources
and always confirm findings with clinical or bench values. The future is unlimited in this field as we
learn more about the body and seek to best aid or replace its many functions.