Journal of Materials Processing Technology 120 (2002) 335±340

Finite-element simulation of stent expansion

S.N. David Chua*, B.J. Mac Donald, M.S.J. Hashmi
School of Mechanical and Manufacturing Engineering, Dublin City University, Dublin 9, Ireland
Accepted 5 September 2001

Abstract

A considerable amount of research has been carried out in order to improve the quality and feasibility of stents in recent years. However,
published works on simulation of stent expansion using computer software are relatively rare. In this paper, a nonlinear ®nite-element
method was employed to analyse the expansion of a stent using internal pressure. The pressure was applied as a surface load on the inner
surface, expanding the stent radially past its elastic limit to a maximum diameter. The total reduction in length as the stent expands radially
was also calculated. The analyses were performed using ®nite-element packages LS-DYNA for solution of the problem and ANSYS
(version 5.5) as a pre- and post-processor. Two different load application speeds were investigated in the simulations and the results were
compared. This paper concentrates on the effect of the speed of pressure applied on a stent during the expansion process. The issues
discussed in this study include the extent of expansion, foreshortening of the stent and the stress levels at different increments of pressure
due to the different speeds of application of the pressure. # 2002 Elsevier Science B.V. All rights reserved.

Keywords: Expansion of stent; Simulation; Finite-element method

1. Introduction the pressurised catheter balloon to a predetermined dia-

meter. The balloon is then de¯ated and the stent delivery
A stent is a device that is used to support arterial walls to system is removed [7]. This leaves the stent to support the
alleviate the blockage of arteries by plaque. Plaque is a fatty damaged arterial wall lining [6].
substance that is deposited on the artery wall. This plaque Several types of vascular stents are now undergoing
accumulates and slowly becomes thicker and in¯uences the clinical evaluation. In this paper, ®nite-element (FE) simu-
mobility of blood ¯ow. If this condition is left untreated, lation is used to simulate the expansion of a stent with
further development of atherosclerosis or build up of fatty pressure applied on the inner surface of the stent. The
substances will reduce the ¯ow of blood and may lead analysis presented in this paper addresses the effect of the
to heart attack. This condition is often treated by coronary speed of pressure on the degree of expansion and foreshor-
bypass surgery or a less invasive procedure called tening of the stent.
angioplasty [1,2]. Percutaneous transluminal coronary
angioplasty (PTCA) is a procedure that opens clogged coro-
nary arteries most commonly by in¯ating a tiny balloon in 2. Finite-element model
the coronary artery [3]. In angioplasty, the balloon-tip
catheter is pushed along a guide wire to the lesion site. The ANSYS ®nite-element package was used to design
The balloon is then in¯ated to compress the plaque against the geometric models presented in this paper. The simulation
the artery wall and, hence, restoring the opening of the artery was carried out to expand a stainless steel 304 stent of 3 mm
to allow increased blood ¯ow. However, this method may outer diameter and 2.9 mm inner diameter with 20 mm in
cause the lining of the arterial wall to tear or sag. length. The model contains 66 slots (cells) which are equally
One method to improve the procedure is to introduce a spaced throughout the entire model. The dimensions of the
stent [4,5]. The procedure is very similar to an angioplasty model are simply described in Table 1.
except when the stent is properly positioned in the artery at The individual stent-strut width is 0.214 mm in circum-
the location of the damaged lining, the stent is expanded by ferencial direction and the length of the bridging strut
connecting between the two adjacent cells is 0.4 mm.
Fig. 1 shows the discretised model of the stent and Fig. 2
*
Corresponding author. shows a close-up view of the mesh used. The ®nite-element

0924-0136/02/$ ± see front matter # 2002 Elsevier Science B.V. All rights reserved.
PII: S 0 9 2 4 - 0 1 3 6 ( 0 1 ) 0 1 1 2 7 - X
336 S.N.D. Chua et al. / Journal of Materials Processing Technology 120 (2002) 335±340

Table 1 75  106 MPa; density, 7:86  10 6

kg=mm3 ; Poisson's
Dimensions of the stent ratio, 0.27.
Category Stent before expansion

Number of cells 66 …33  2†

Cell size 1.1128 mm2 3. Loading and solution
Cell areas 73.4448 mm2
Metal surface area 188.4955 mm2 Due to the symmetry of the stent by utilising the correct
Stent surface area 115.0508 mm2 boundary conditions, a half model was used to simulate the
Outer diameter of the stent 3 mm
expansion process. Symmetric boundary conditions were
Inner diameter of the stent 2.9 mm
Length of stent 20 mm imposed on the nodes of the stent in the plane of symmetry.
Both ends of the stent were free from any constraints. The
pressure load was applied as a surface load on the inner stent
surface, expanding the stent radially past its elastic limit to a
maximum diameter before failure stress was reached.
Two simulations with different load speeds (Fig. 3) were
investigated during the analysis. For load case I, the stent
was subjected to a uniform internal pressure increasing at a
constant rate of 400 MPa/ms. For load case II, the stent was
subjected to a uniform internal pressure increasing at a
constant rate of 11.3 MPa/ms.
In order to reduce the computational time, it was neces-
sary to set a short simulation time for the analysis. Although
this is much shorter than the actual time required to expand
the stent, the kinetic energy was checked throughout the
entire analysis to ensure that it was negligible compared to
the internal energy as recommended by the developers of
LS-DYNA [8]. Fully integrated elements were used in the
Fig. 1. Finite-element model of half of the stent. model despite the fact that reduced integration elements
were available and are very computer time ef®cient. This is
model consists of a total of 9504 elements. The stent was because reduced integration elements are prone to hourglass
discretised by 200 elements along its modelled length and 88 deformation. Owing to the structure of the model, which
elements in circumference with one element across the was thin and slotted, choosing reduced integration elements
thickness. was not a good option, as a high level of hourglass deforma-
The elements used in the model are 8-node explicit solid tion would be expected. A number of simulations were
brick elements [8]. A bi-linear elasto-plastic material model run using reduced integration elements, and the results
was assumed for the stent material. The material properties were compared with the results obtained using fully inte-
were chosen to approximately represent stainless steel 304 grated elements. The resulting solutions showed that the
and are as follows [7]: Young's modulus, 193 GPa; tangent stress levels obtained using reduced integration elements
modulus, 692 MPa; yield strength, 207 MPa; shear modulus, were less than those obtained using fully integrated ele-
ments. The level of hourglass energy was checked and
found to be much higher than acceptable limits, de®ned

Fig. 2. Selected discretised geometry. Fig. 3. Two different load cases.

 S.N.D. Chua et al. / Journal of Materials Processing Technology 120 (2002) 335±340 337

by the developers of LS-DYNA as less than 10% of the and they were located at a region close to the middle of the
internal energy of the process [8]. stent.

4.1. Load case I

4. Results and discussion
The distribution of von Mises stress in the expanded stent
The nodes on the plane of symmetry were constrained just before failure produced by load case I is indicated in
accordingly to allow the stent radius to change according to Fig. 4(a). Fig. 4(b) shows a close-up view of the deformed
the applied load. The critical pressure was determined based mesh in the centre portion of the stent after enlargement.
on the ultimate stress or tensile strength of the material As can be seen in Fig. 4(b), the regions of high stress are
which was approximately 517 MPa [7]. For the stent, concentrated at the four corners of the cells rather than in
maximum equivalent stresses were used for the direct the middle of the struts or the bridging strut itself. This
comparison with the ultimate material strength. All max- was because the struts pull apart from each other to form
imum radial displacements were measured by taking the a rhomboid shape of cells during the expansion. This
lowest displacement values of the nodes in radial direction also resulted in foreshortening of the stent. As the stent is

Fig. 4. (a) Distribution of von Mises at maximum expansion; (b) closer view of the central stent-strut.
338 S.N.D. Chua et al. / Journal of Materials Processing Technology 120 (2002) 335±340

Fig. 5. Maximum radial expansion achieved under load case I.

expanding, the bridging struts move closer and closer to each

other until the maximum applied pressure is reached.
Fig. 5 shows the maximum radial displacement of the
stent at equal increments of pressure for load case I. The
maximum pressure just before failure is 13 MPa. At that
pressure level, the highest stress value in the stent is
517 MPa which is equal to the ultimate tensile strength of
the material (Fig. 4(a)). The maximum radial displacement
obtained is found to be 5.6 mm from the central axis of the
stent. The maximum foreshortening of the stent in axial
direction at that point is 2.06 mm.

4.2. Load case II

Fig. 6(a) shows the distribution of von Mises stress just

before failure in the expanded stent produced by load case II.
Fig. 6(b) shows a close-up view of the deformed mesh in the
central part of the stent after enlargement.
Fig. 6. (a) Distribution of von Mises at maximum expansion; (b) closer
Distending of the stent started from both ends and then view of the central stent-strut.
developed towards the centre part of the stent. As the
pressure increased, the cells formed a rhomboid shape
(Fig. 6(b)) and the stent experienced foreshortening. Loca-
lised expansion of the struts was found to occur if the applied
internal pressure was too small.
Fig. 7 shows the radial displacement achieved as the
pressure increased gradually. It was found that maximum
expansion was obtained by using a maximum pressure of
1.35 MPa. At this level of pressure, the maximum radial
displacement obtained was 6.72 mm and the ultimate fore-
shortening of length was 6.5 mm. The maximum von Mises
stress was 517 MPa. The path of stent expansion process in
load case II was very similar to load case I. Apart from
Fig. 7. Maximum radial expansions achieved under load case II.
greater level of displacement and foreshortening, the only
difference between load cases was that load case II seemed
to produce a greater bending at both ends. In reality, it is not 5. Comparison of results
preferable to have an excessive bending at the ends of the
stent as it may injure the artery wall during stent deploy- Potential diameter of the distending stent was calcu-
ment. During the expansion process, it was observed that lated from the deformed shape at each equal increment of
after pressure of 1.1 MPa, the stent begun to retract long- pressure and compared between the two load cases. Fig. 8
itudinally along its length at a greater rate as well as further shows the plots of radial displacement with increasing
expanding in diameter. pressure. It is clearly shown that the stent will experience
 S.N.D. Chua et al. / Journal of Materials Processing Technology 120 (2002) 335±340 339

Fig. 8. Displacement against percentage increment of pressure at equal intervals.

Fig. 9. Development of stress in stent under load cases I and II.

greater expansion under the load case II despite the fact

that the pressure load in load case I was much higher than
load case II (13 MPa for load case I and 1.3 MPa for load
case II). This is primarily due to the speed of application of
pressure.
Fig. 9 shows the development of von Mises stress
throughout the expansion process. It is clearly shown that
the stresses in the stent developed at a faster rate within the
®rst millimetre of radial expansion regardless of what Fig. 10. Comparison between displacement and foreshortening at each
pressure speeds were applied. However, the stresses in load pressure point: (a) load case I; (b) load case II.
case I seemed to develop quicker at every subsequent 1 mm
of increment in radial displacement after the ®rst millimetre
of expansion. As a result, the stent under load case I reached when the pressure was approaching the critical value. In load
the failure point at a lower level of radial expansion than load case II (Fig. 10(b)), the stent experienced a small amount of
case II. foreshortening until it reached a pressure of 0.8 MPa at
Fig. 10(a) and (b) shows the comparison of displacements which point it tended to pick up rapidly. When the pressure
and foreshortenings at each pressure point for both load reached 1.1 MPa, the stent reduced its length at a quicker
cases. During the expansion process, the radial displacement rate than the diameter expanded. At the maximum pressure,
and foreshortening increased gradually as the pressure the total foreshortening and radial displacement were nearly
increased to 13 MPa as can be seen in Fig. 10(a). It was equal for load case II. For radial displacement of 5.6 mm, the
observed that foreshortening of stent always occurs slower foreshortening was only about 2 mm in load case I and about
than radial displacement, and it tended to go even slower 4.5 mm in load case II.
340 S.N.D. Chua et al. / Journal of Materials Processing Technology 120 (2002) 335±340

ity of stress±strain plots of the structure against material

property data and smooth continuity of contour lines across
element boundaries.
It was observed that the stent assumed a concave shape at
maximum expansion and the end struts showing more
concave-shaped deformation. During stent deployment, it
is not desirable to have such abrupt bending because it may
cause injury and damage to the arterial wall [9].

7. Conclusion

From the analyses shown, it is clear that only moderate

pressure speed applied on the inner surface of the stent can
produce the maximum potential diameter. However, smaller
Fig. 11. Comparison between the length of foreshortening and radial pressure and lower pressure speed will probably generate
displacement. greater potential diameter, but it can also introduce gross
buckling of the stent which is not desirable on stent deploy-
Fig. 11 shows the total length of foreshortening for every ment. On the contrary, higher pressure speeds will result in
1 mm increment of radial displacement. It is evident that the higher stresses in the stent for similar resultant. Higher speed
application of high-speed pressure has a direct effect on the of pressure generates higher stress but lower radial displace-
foreshortening of stent during the expansion. Therefore, ment.
faster speed of pressure application implies greater fore- In order to make the simulation more realistic, a balloon is
shortening of stent. Load case II has the tendency of having a used as a medium to distend the stent in future work. This
slower expansion in the radial direction compared to fore- will require modelling the balloon and contact conditions
shortening in axial direction. In comparison to load case I, it between the stent and balloon. The effect of using different
can be seen that it is not necessarily the case that the stent stent geometries will also be studied.
will have smaller displacement in radial direction during the
expansion. In fact, the stent will experience greater radial
displacement as illustrated in Fig. 11. The graph also shows References
that load case I was producing less foreshortening than load
case II for every equal interval increment of radial displace- [1] S.B. King III, N.J. Lembo, W.S. Kosinski, et al., A randomized trial
comparing coronary angioplasty with coronary bypass surgery, New
ment. The percentage difference of foreshortening between
Engl. J. Med. 331 (1994) 1044±1050.
load cases I and II was nearly 100% when 5 mm of radial [2] D.L. Fischman, M.B. Leon, D.S. Baim, R.A. Schatz, M.P. Savage, A
displacement was achieved. randomized comparison of coronary±stent placement and balloon
angioplasty in the treatment of coronary artery disease, New Engl. J.
Med. 331 (1994) 496±501.
[3] D.A. Morrison, J. Sacks, F. Grover, K.E. Hammermeister, Effective-
6. Study limitations
ness of percutaneous transluminal coronary angioplasty for patients
with medically refractory rest angina pectoris and high risk of adverse
This study presents a ®nite-element analysis of predicting outcomes with coronary artery bypass grafting, Am. J. Cardiol. 75
the stent diameter and foreshortening under different speeds (1995) 237±240.
of pressure application during the entire expansion process. [4] R.A. Hontz, M.D. Tripp, L.P. Kline, Stents keep occluded vessels
open, RN 54 (1991) 50±54.
In the actual situation, a balloon is used as the medium to
[5] C. Palmaz, Intravascular stenting: from basic research to clinical
distend the stent. In such a case, the analysis will be more application, Cardiovasc. Inter. Rad. 15 (1992) 279±284.
complicated due to the fact that the balloon will have to be [6] N. Morris, Nurse and Technologist Guide, Johnson & Johnson
modelled and contact conditions between the balloon and Interventional Systems Co., New Jersey, 1995.
stent will have to be considered. [7] R.C. Hibbeler, Mechanics of Materials, 2nd Edition, Macmillan, New
York, 1994.
It should be noted that no experimental works have been
[8] J.O. Hallquist, LS-DYNA3D Theoretical Manual, Livermore Software
conducted so far. Validation of the results was simply based Technology Corporation, Livermore, CA, 1993.
on the standard general validation checks of the ®nite- [9] A.J. Carter, D. Scott, L.R. Bailey, et al., Stent design: in the ``ends'' it
element method such as smooth load development, similar- matters, J. Am. Coll. Cardiol. 29 (1997) 2259A.