Professional Documents
Culture Documents
4 130
E(z) = 20 + z (22)
A x i i 162
G (z ) = i =1 (12)
4 for z [0, 160]. Substituting equation (22) into equation
Ai (9), it is possible to determine the displacement
i =1 distribution along the lateral side of the composite stem
and compare it with the ones obtained with the finite
L(z) = ztg30 + R cos r (13) element analysis and with the experimental study.
The numerical integration of equation (9) was
where Ai and xi are the correspondent area and performed with Mathematica software.
coordinate x for each of the geometric figures of the
whole cross section (figure 3). For L(z), 96<z<162. The 2.3. Numerical (FEA) displacement distribution
G(z) function can be obtained replacing the following
equations: A finite element model was used to test if the second
moment of area function, equation (16), would provide
4
r 2
identical results. Hence, the analytical displacement
A = Ri
2
rR cos + 2rB(z) +
2
(14) distribution was compared with the same resulting from
i =1 the finite element analysis (FEA), where constant
4 material modulus prosthesis was considered. For the
A x i i =
2 3
3
2
R sin + rR 2 cos 2 + rB2 (z) 2rB(z )R cos
3 finite element analysis, the prosthesis was modelled with
i =1 SolidWorks computer aided design software and
r 2 2
R cos + B(z) r 2 + r 3 (15) simulated with Hypermesh finite element analysis
2 2 3 software. The finite element model was composed of
21765 tetrahedral four-node elements and 5340 nodes.
The final function for the second moment of area Figure 4 shows the finite element mesh of the prosthesis.
was obtained considering the sum of the second moment The same was loaded laterally-medially by a single load
of area of each of the elementary geometries: of 700 N.
The displacement data in the prosthesis front plane
4 were very identical (maximum difference of 32 m).
I y (z ) = I
i =1
y i (z) = I y1 (z) + I y 2 (z) + I y 3 (z) + I y 4 (z) (16) This fact gave us the necessary confidence to use the
Iy(z) equation to derive the lateral side displacements
and compare them with the same obtained
where experimentally.
sin 2 4 sin 2
I y1 = R 4 + + R 2 2R sin + G (z ) (17)
4 8 9 3 2.4. Experimental displacement distribution
y( z) = fringe _ order (23)
2
Figure 4 Finite element mesh of the prosthesis.
For the laser used, =632 nm, and replacing this
In both, the numerical and experimental models, the
value in equation (23), the displacements were obtained
prosthesis was loaded as a cantilever beam. No torsional
along the lateral side of the prosthesis.
force due to the anterior loading component of the hip
joint force was taken into account within the
3. RESULTS AND DISCUSSION
experimental setup. The objective within this part of the
study was to validate the controlled stiffness concept, by
The applicability of the analytical second moment of
determining an optimal variation composite thickness,
area function of the prosthesis model was verified by
assuming that the bending stiffness of the optimised
comparison of the displacements calculated with the
prosthesis is the same as the bending stiffness of a
finite element method and the numerical integration of
prosthesis composed of a metal core and a composite
equation (9). The results obtained for both methods were
layer. For an equivalent torsional stiffness, the variation
similar along the prosthesis length. The maximum
of the optimal composite layer thickness was inferior to
displacement observed was 1.704 mm for the numerical
the one obtained for the equivalent bending stiffness,
integration of the analytical model and 1.720 mm for the
therefore torsion played a second-order effect
finite element method. This fact gave us the necessary
concerning the objectives of the study hereby presented.
confidence to use the analytical second moment of area
With this technique (ESPI), the assessment of the
function to obtain the displacements of the controlled
displacement field of the prototype was possible with no
stiffness prototype analytically [7].
need of contact or surface preparation and with very
Table 1 shows the displacement values measured
high resolution, on the order of magnitude of the laser
experimentally and those obtained analytically. Figure 6
wavelength. The experimental displacements were
shows a graphical comparison of the ESPI and
obtained with a manufactured prototype.
numerical integration displacement fields. A relatively
A stiff support was used to hold the prosthesis 30
good correlation between these was obtained using the
mm from its distal end. The magnitude of the
method described, and so we can state with some
displacement field measured was of the order of a
certainty that the manufactured prosthesis prototype
micron. A very low load of P = 0.2 N was used, which
possesses equivalent optimised stiffness. This means that
allowed us to obtain a reasonable number of fringes to
the prosthesis was manufactured with an equivalent
calculate the displacement values. The experimental
Youngs modulus as stated in the Materials and Methods
setup used is shown in figure 5. The interferometer arms
section. However, at the distal part of the prosthesis
are oriented in such a way that the sensitivity vector was
where it was rigidly fixed, the highest displacement
coincident with the direction of the displacements
difference was observed, which was due to the nature of
measured, that is, perpendicular to the lateral surface of
the experimental set-up.
the prosthesis.
The displacement difference near the prosthesis
fixation device was 400% and gradually diminished up
to a difference of 2.5% at the last observed fringe, which
was localised at a distance of 15 mm from the applied
load.
The higher differences observed at the fixation
region of the prosthesis may have resulted from a less
efficient fixation of the prosthesis. In the analytical-
numerical model, the fixation of the prosthesis was
considered to be ideal. We must say that the measured
displacements are of the order of a micron and can
easily be influenced by the experimental setup. The
maximum displacement obtained with the ESPI
technique was 4.3 m; for the analytical model, the
Figure 5 ESPI set-up.
maximum displacement was 4.1 m, a relative
difference of 5%.
analytical model. Relative to the prosthesis concept, it is
possible to generate optimised Youngs modulus data by
overcoming possible composite manufacturing
difficulties.
REFERENCES
4. CONCLUDING REMARKS