Biomedical Engineering Reference
In-Depth Information
Table 1 Patch test results for uniform and non-uniform multi-element model and single element
model for uniaxial tension and simple shear
Multi element uniform
Multi element non-uniform
Single element
Tensile
S11 (Pa)
6.776
6.776
6.776
S22 (Pa)
0
:
168
10
7
0
:
168
10
7
0
:
168
10
7
S12 (Pa)
0
0
0
U1 (mm)
0
:
581
0
:
581
0
:
581
Shear
S11 (Pa)
0
:
24
10
5
!
1
:
75
10
5
0
:
23
10
5
!
1
:
73
10
5
0
:
41
10
5
S22 (Pa)
0
:
188
10
5
!
4
:
92
10
5
0
:
193
10
5
!
4
:
82
10
5
0
:
59
10
5
S12 (Pa)
0
:
072
10
5
!
2
:
86
10
5
0
:
097
10
5
!
2
:
81
10
5
0
:
86
10
5
Ranges are given for the simple shear test. Notations: S11, S22 Normal stress in 1 and 2
directions; S12 Shear stress; U1 Displacement in 1 direction
Uniaxial tensile tests were performed on samples with dimensions 60
20 mm
at 37
C (Instron
5544, Instron Corp., Norwood, MA), see Fig.
4
a. The load-
ing protocol comprised: (1) Pre-load to 1 % nominal strain at a strain rate of
50 mm/min to reduce material inconsistencies, and (2) Extension to 50 % nominal
strain at a strain rate of 200 mm/min, the strain rate stipulated by medical implant
standards [
16
,
17
]. Due to difficulties in monitoring lateral strain effects during the
Instron
tests, caused by curling effects at the sample edges, fabric samples were
strained on a flat bed at 37
C in steps of 10, 20 and 30 %. For both the Instron
and the flat bed tests, a 5
5 mm grid was marked on each sample to visualize
localized strain effects. Strengthening stitches were sewn at both ends of the
samples to minimize localized stress concentrations. Using digital images recorded
during the tests, the longitudinal and transverse strain was determined from a
single grid cell located in the center of the sample.
A quarter-symmetric FE model was used to simulate the physical uniaxial tensile
tests. The model, see Fig.
4
b, utilized a mesh of four-noded membrane elements.
The boundary conditions were selected such that edge AB was free to move hor-
izontally but constrained vertically, while edges AD and DC were free to move
vertically but constrained horizontally. A quasi-static displacement at 200 mm/min
[
16
,
17
] was applied to edge DC. The loaded boundary DC and the edge BC were
expected to have a greater variation in stress and deformation. Hence the element
mesh was refined toward these edges. The mesh sensitivity was assessed by
increasing and biasing the element density toward edges BC and DC and center
point A until the stress, strain and displacement fields became consistent.
2.3.1 Optimisation of Fabric-Specific Constitutive Coefficients
A genetic algorithm GA1, programmed using Perl
, was utilized to iteratively
optimize the fabric model coefficients to represent the physically tested fabrics.
Using a single set of fabric coefficients, GA1 ran mutually orthogonal uniaxial
Search WWH ::
Custom Search