Anisotropic hyperelastic modeling of arterial layers
This problem illustrates the use of the anisotropic hyperelastic capabilities
in Abaqus to model soft biological tissue.
More specifically, the problem shows how
these capabilities can be used to model the mechanical response of the adventitial layer of
human iliac arteries.
Numerical examples are provided
for simple tension tests of iliac adventitial strips cut along the axial and circumferential
directions of the artery. An example of a strip cut at an angle of 15° with respect to the
circumferential direction is also included. The numerical study demonstrates the significant
effect that dispersion of the collagen fiber orientations can have on the mechanical response
of soft tissue. The problem has been analyzed numerically by Gasser, Holzapfel, and Ogden (2006).
We consider the numerical analysis of simple tensile tests of adventitial strips cut along
the axial and circumferential directions of the artery, as well as a strip cut at an angle
of 15° with respect to the circumferential direction, as illustrated in Figure 1. Following the work of Gasser, Holzapfel, and Ogden (2006), the adventitial strips considered in the
study have referential dimensions of 10.0 mm length × 3.0 mm width × 0.5 mm thickness and
are assumed to be stress free in the reference configuration. It is assumed that two
families of collagen fibers are embedded in the specimens, symmetrically arranged with
respect to the axial and circumferential directions of the artery and with no component in
the radial (thickness) direction, as shown in Figure 1. The angle between the mean orientation of the fibers and the circumferential direction
is =49.98°. The specimens are loaded in the longitudinal direction, and their
end faces are not allowed to deform. Using appropriate symmetry boundary conditions, only
one-eighth of the geometry is modeled in the case of strips cut along the axial and
circumferential directions. A full scale simulation is carried out for strips cut at a 15°
angle with respect to the circumferential direction. The finite element model of the
specimens consists of a 20 × 10 × 2 mesh for the simulations considering one-eighth symmetry
and a 40 × 20 × 4 mesh for the full scale simulations.
The numerical analyses are conducted using the static analysis procedure in Abaqus/Standard. Linear solid hybrid elements (C3D8H) are
used to model the incompressible deformation of the arterial layers. For comparison, the
solution is also computed in Abaqus/Explicit assuming quasi-static loading conditions. Since Abaqus/Explicit has no mechanism for imposing the incompressibility constraint, some amount of
compressibility is introduced in the material response in the Abaqus/Explicit simulations and linear solid elements
(C3D8R) are used.
Material
The mechanical response of the adventitial layer is modeled using the anisotropic
hyperelastic strain energy function proposed by Gasser, Holzapfel, and Ogden (2006) to model arterial layers with distributed
collagen fiber orientations. Details of the modeling materials are given in Holzapfel-Gasser-Ogden Form, and a
discussion of the theory behind anisotropic hyperelastic material behavior is presented in
Holzapfel-Gasser-Ogden form.
It is assumed that the arterial layer is composed of two families of collagen fibers
embedded in a soft incompressible ground matrix. The families of fibers have mean
orientations characterized by the vectors and in the reference configuration, but the orientations of the fibers
within each family are dispersed. The model incorporates a scalar structure parameter, (), that characterizes the level of dispersion of the collagen
orientations. When =0, the fibers are perfectly aligned (no dispersion). When =1/3, the fibers are randomly distributed and the material becomes
isotropic. A value of =0.226 is used in the numerical simulations. For comparison, numerical
tests are also carried out assuming ideal alignment of the collagen fibers (=0).
In addition to its native implementation, the Holzapfel-Gasser-Ogden model can be
implemented using the constitutive artificial neural networks
(CANNs) framework (see Modeling Soft Matter Using Constitutive Artificial Neural Networks (CANNs)). Input
files with the Holzapfel-Gasser-Ogden form implemented as a
CANNs model are included to replicate the results
obtained with the native implementation.
Loading and controls
The mounting of the specimen in the testing machine is modeled by constraining both ends of
the strip. The strips are loaded in the tensile direction, and their end faces are not
allowed to deform.
In the Abaqus/Standard simulations the static procedure is used with a prescribed load of 2.0 N. Geometric
nonlinearities are considered in this step to account for the large deformations of the
adventitial strip.
Since Abaqus/Explicit is a dynamic analysis program and we are interested in a static solution to the problem,
care must be taken to avoid significant inertia effects as the adventitial strip is loaded.
A smooth step amplitude curve is used to prescribe the uniaxial displacement of the strip
and to promote a quasi-static solution in the Abaqus/Explicit simulations. The simulations are run in double precision.
Results and discussion
Results for the Abaqus/Standard and Abaqus/Explicit analyses are discussed in the following sections.
Abaqus/Standard results
Figure 2 shows the computed stress in the tensile direction for the axial (left) and
circumferential (right) specimens with distributed fibers (=0.226) for a tensile load of 2.0 N. The thickness of the specimens
remains approximately constant during loading, with small transition zones at the ends of
the strips. The corresponding results for the case of perfectly aligned fibers (=0) are shown in Figure 3. In this case the embedded collagen fibers need to rotate significantly toward the
loading direction before they can carry significant load. The combined effect of the large
rotation of the fibers and the incompressibility constraint causes the thickness of the
specimen to increase (and the width to decrease) in the middle region of the strip, away
from the restrained boundaries. The transition areas at the end of the strip resemble
deformation patterns similar to those observed in woven fabrics.
Figure 4 shows the computed load versus displacement curves for the circumferential and axial
specimens. The dashed curves correspond to the simulations with ideally aligned fibers,
and the continuous curves correspond to the simulations that include dispersion. As seen
in the figure, the material response is very soft at low stretches; only a small force is
needed to achieve significant extension. Once the collagen fibers are approximately
aligned with the loading direction, the material stiffens rapidly. This is particularly
evident in the case of the circumferential specimen with =0; the alignment requires very large average stretches, and the specimen
stiffens at a displacement of about 4 mm. In contrast, when dispersion is included in the
simulation, the collagen fibers need to rotate less before they carry load compared with
the ideally aligned case. Therefore, the dispersion of the collagen fibers leads to a
stiffer macroscopic response of the specimens. Specifically, the dispersion parameter controls the elongation at which the specimen stiffens.
These numerical results for axial and circumferential specimens are in agreement with the
results reported in Gasser, Holzapfel, and Ogden (2006).
Figure 5 shows the stress in the loading direction in a specimen cut at an angle of 15° with
respect to the circumferential direction. The plot on the left corresponds to the case of
distributed fibers, and the plot on the right corresponds to the case of ideally aligned
fibers. Again, we observe that significantly more rotation is required for ideally aligned
fibers before they carry load.
Abaqus/Explicit results
For comparison, the axial and circumferential specimens with distributed fibers are also
analyzed using Abaqus/Explicit. Figure 6 shows the stress in the tensile direction for the axial (left) and circumferential
(right) specimens at the end of the Abaqus/Explicit simulations. As illustrated in Figure 7, the load-displacement response computed in Abaqus/Explicit compares well with that obtained in Abaqus/Standard. Small discrepancies in the results can be attributed to the use of different element
types as well as minor dynamic effects and some amount of material compressibility in the
Abaqus/Explicit simulations.
Holzapfel-Gasser-Ogden model implemented as a
CANNs model; simple tension of adventitial strip
cut at an angle of 15° with respect to the circumferential direction; distributed
fibers (=0.226); C3D8H
elements.
Holzapfel-Gasser-Ogden model implemented as a
CANNs model; simple tension of adventitial strip
cut at an angle of 15° with respect to the circumferential direction; ideally
aligned fibers (=0); C3D8H elements.
Holzapfel-Gasser-Ogden model implemented as a
CANNs model; simple tension of adventitial strip
cut along axial direction; distributed fibers (=0.226); 1/8 symmetry model; compressible;
C3D8R elements.
Additional Abaqus/Standard input files not used for discussion
Simple tension of adventitial strip cut at an angle of 15° with respect to the
circumferential direction; distributed fibers (=0.226); plane stress 1/4 symmetry model; incompressible;
M3D4 elements.
Simple tension of adventitial strip cut at an angle of 15° with respect to the
circumferential direction; ideally aligned fibers (=0); plane stress 1/4 symmetry model; incompressible;
M3D4 elements.
References
Gasser, T.C., G. A. Holzapfel, and R. W. Ogden, “Hyperelastic
Modelling of Arterial Layers with Distributed Collagen Fibre
Orientations,” Journal of the Royal Society
Interface, vol. 3, pp. 15–35, 2006.
Holzapfel, G.A., T. C. Gasser, and R. W. Ogden, “A
New Constitutive Framework for Arterial Wall Mechanics and a Comparative Study
of Material Models,” Journal of
Elasticity, vol. 61, pp. 1–48, 2000.
Holzapfel, G.A., G. Sommer, and P. Regitnig, “Anisotropic
Mechanical Properties of Tissue Components in Human Atherosclerotic
Plaques,” Journal of Biomechanical
Engineering, vol. 126, pp. 657–665, 2004.
Tables
Table 1. Assumed material properties for iliac adventitial layer.
Holzapfel-Gasser-Ogden energy function
coefficients:
= 3.82 kPa
= 996.6 kPa
= 524.6
= 0.226
= 0 ( = 1× 10−6 for compressible case)
Fiber directions (for strips cut along
circumferential direction):
with = 49.98°.
Figures
Figure 1. Adventitial layer with two embedded families of fibers with mean orientations and (left). Definition of circumferential, axial, and 15° specimens for the
tensile tests (right). Figure 2. Stress in the direction of applied load for iliac adventitial strips cut in the axial
(left) and circumferential (right) directions. Results correspond to an applied load of
2.0 N; dispersion of collagen fibers is included (=0.226). Figure 3. Stress in the direction of applied load for iliac adventitial strips cut in the axial
(left) and circumferential (right) directions. Results correspond to an applied load of
2.0 N; collagen fibers are perfectly aligned (=0). Figure 4. Load-displacement response of circumferential and axial specimens. Figure 5. Stress in the direction of applied load for iliac adventitial strips cut at an offset
angle of 15° with respect to the circumferential direction; with dispersion of collagen
fibers included (left) and without dispersion (right). Figure 6. Abaqus/Explicit results for the stress in the direction of applied load for iliac adventitial strips
with distributed fibers (=0.226) cut in the axial (left) and circumferential (right) directions.
Results correspond to the end of the Abaqus/Explicit simulations. Figure 7. Comparison of Abaqus/Standard and Abaqus/Explicit results for the load-displacement response of circumferential and axial specimens with
distributed fibers (=0.226).