Page 271 - Biomedical Engineering and Design Handbook Volume 1, Fundamentals

248 BIOMECHANICS OF THE HUMAN BODY

clinical results. Most FEA codes allow orthotropic material properties in 2D or 3D. This being stated,
it must also be noted that since people vary widely in shape and size, material properties may have
standard deviations of 30 percent or more. Once a determination of what material properties are
going to be used is made, it may be beneficial to vary these parameters to ensure that changes of
properties within the standards will not significantly alter the findings.
Distinction must be made between cortical and cancellous bone structures. The harder cortical
outer layer carries the bulk of the loading in bones. Although one should not neglect the cancellous
bone, it must be recalled that during femoral implant procedures for hip implants, the cancellous
canal is reamed out to the cortical shell prior to fitting the metal implant.
Joint articular cartilage is a particular area of concern when producing an FEA model. Work
continues on fully describing the mechanical behavior of these low-friction cushioning structures
that allow us to move our joints freely. As a first approximation, it is generally assumed that these
surfaces are hard and frictionless. If the purpose of the model is to investigate general bone behavior
away from the joint, this will be a reasonable approximation. To isolate the cartilage properties, one
should perform a data search of the current literature. Of course, implant modeling will not require
this information because the cartilage surface is generally removed.
When evaluating intact joints, after the articular cartilage has been modeled, one is forced to
examine the synovial fluid. This material is a major component to the nearly frictionless nature of
joints. Long molecular chains and lubricants interact to reduce friction to levels far below that of the
best human-made materials. At a normal stress of 500 kPa, a sample of a normal joint with synovial
fluid has a friction coefficient of 0.0026, whereas a Teflon-coated surface presents coefficients in the
range of 0.05 to 0.1 (Fung, 1996). In addition, the articular cartilage reduces friction by exuding
more synovial fluid as load increases, effectively moving the articular surfaces further apart. When
the load is reduced, the fluid is reabsorbed by the articular materials.
Implants require their own careful study as pertains to the implant-bone interface. If an unce-
mented implant is to be investigated, the question is one of how much bone integration is expected.
If one examines the typical implant, e.g., the femoral component of a hip implant or the tibial com-
ponent of a knee implant, one will see that a portion of the surface is designed for bone ingrowth
through the use of either a beaded or wire-mesh style of surface. These surfaces do not cover the
entire implant, so it should not be assumed that the entire implant will be fixed to the bone after healing.
Cemented implants are expected to have a better seal with the bone, but two additional points must
be made. First, the material properties of bone cement must be included. These are isotropic and are
similar to the acrylic grout that is the primary component of bone cement. This brings up the second
point, which is that bone cement is not an actual cement that bonds with a surface but rather a tight-
fitting grout. As a first approximation, this bonding may be considered a rigid linkage, but it must be
recalled that failure may occur by bone separation from the implant (separation from the bone is less
likely as the bone cement infiltrates the pores in the bone).
Ligaments and tendons may be modeled as nonlinear springs or bundles of tissue. In the simplest
case, one may successfully model a ligament or tendon as one or more springs. The nonlinear behavior
is unfortunately evident as low force levels for ligaments and tendons. The low force segment of their
behavior, where much of the motion occurs, is nonlinear (forming a toe or J in the stress-strain curve)
due to straightening of collagen fibers of different lengths and orientations (Nowak, 1993; Nowak
and Logan, 1991). The upper elastic range is quasi-linear. As multifiber materials, ligaments and
tendons do not fail all at once but rather through a series of single-fiber microfailures. If one is
interested in evaluating behavior in the permanent deformation region, this ropelike behavior must
be accounted for.
Dental implants area also nonisotropic in nature, as are many of the implant materials, especially
composite-reinforced materials (Nowak et al., 1999). Fiber-reinforced composites generally require
a 3D model for analysis, and the properties must be applied using a local coordinate system. This is
so because the composites are generally wrapped around a post, and the lengthwise properties no
longer run simply in one direction. Choosing a local coordinate system around curves will allow the
modeler to keep the high-modulus direction along the cords of the reinforcements.
Although skin is an anisotropic material having different properties in each direction, Bischoff
et al. (2000) have presented an interesting model as a second step beyond isotropic material selection.