Chinese Journal of Aeronautics, (2014),27(2): 272–279

Chinese Society of Aeronautics and Astronautics

& Beihang University
Chinese Journal of Aeronautics
[email protected]
www.sciencedirect.com

Study of parachute inflation process using fluid–structure

interaction method
Yu Li *, Cheng Han, Zhan Ya’nan, Li Shaoteng

College of Aerospace Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, China

Received 25 March 2013; revised 28 August 2013; accepted 11 October 2013

Available online 28 February 2014

KEYWORDS Abstract A direct numerical modeling method for parachute is proposed firstly, and a model for
Empirical formula; the star-shaped folded parachute with detailed structures is established. The simplified arbitrary
Fluid–structure interaction; Lagrangian–Eulerian fluid–structure interaction (SALE/FSI) method is used to simulate the infla-
Inflation process; tion process of a folded parachute, and the flow field calculation is mainly based on operator split-
Opening shock; ting technique. By using this method, the dynamic variations of related parameters such as flow field
Parachute; and structure are obtained, and the load jump appearing at the end of initial inflation stage is cap-
Wind tunnel test tured. Numerical results including opening load, drag characteristics, swinging angle, etc. are well
consistent with wind tunnel tests. In addition, this coupled method can get more space–time detailed
information such as geometry shape, structure, motion, and flow field. Compared with previous
inflation time method, this method is a completely theoretical analysis approach without relying
on empirical coefficients, which can provide a reference for material selection, performance optimi-
zation during parachute design.
ª 2014 Production and hosting by Elsevier Ltd. on behalf of CSAA & BUAA.
Open access under CC BY-NC-ND license.

1. Introduction Therefore, the experimental study on the parachute inflation

process mostly focuses on the opening shock load which can
Phenomena such as canopy collapse or contraction, apex be easily measured.2–5 From 2006, earlier studies on the dy-
whipping, asymmetric inflation, etc. usually occur because of namic relationship between the projected diameter and open-
the fluid–structure interaction during parachute inflation pro- ing shock of a flat circular parachute were published, in
cess.1 The inflation process is transient and shorter than 1 s which high speed camera and dynamic balance were used
in most situations, making it difficult to measure the time- based on synchronous measure technology.4,5 However, the
sequence canopy shape, the flow field and other dynamic data. corresponding literature on the dynamic relationship between
the fluid field and structure in inflating is still unavailable.
The fluid–structure interaction (FSI) problem of parachute
* Corresponding author. Tel.: +86 25 84893739.
is highly nonlinear and unsteady, which is very difficult and
E-mail address: [email protected] (L. Yu). complex. Therefore the earlier numerical studies focused on
Peer review under responsibility of Editorial Committee of CJA. the dynamics of the parachute and payload system, which need
to rely on empirical formula or data. The typical methods are
inflation distance method, inflation time method, radical
Production and hosting by Elsevier momentum equation method, etc.6,7 All of those methods need

1000-9361 ª 2014 Production and hosting by Elsevier Ltd. on behalf of CSAA & BUAA. Open access under CC BY-NC-ND license.
http://dx.doi.org/10.1016/j.cja.2014.02.021
Study of parachute inflation process using fluid–structure interaction method 273

to be based on drag characteristics and associated air mass Most previous research works of parachute inflation using
experience formulas which are summed from experiments. the space–time FSI technique are based on a supposed initial
The parachute–payload system is divided into several objects canopy shape and the structure details are ignored. In this work,
described by particle or rigid body in those methods. And the folded canopy model including the structure details was
the trajectory, load and other parameters of research objects established by direct folding method, which is simple and easy
may be obtained, but the canopy shape and stress, the flow to realize. Based on this direct modeling, the folded ringslot
field surrounding the parachute and other dynamic character- inflating is simulated by SALE/FSI method for the first time.
istic cannot be obtained. The fluid–structure interaction Based on the SALE method, the 3D FSI computation of a
mechanical mechanisms of the parachute during the inflation folded ringslot inflating was realized. The accuracy and effi-
stage can’t be revealed. ciency of the simulation results in this work are also verified
With the continuous development of numerical algorithms by the experimental data. In this method, more space–time de-
and the computer hardware, FSI problem of parachute has tails such as flow field, stress, canopy shape and other character-
been attracting scholars’ attention at home and aboard. The istics can be obtained compared with the traditional dynamics
computational fluid dynamics/mass spring damper (CFD/ method of parachute–payload system.
MSD) model is used in the earlier FSI method based on This coupling calculation was carried out by the DELL
time–space dispersing. In this model, the parachute is dis- T5500 workstation with 12-core, 32 G memory. The CPU time
persed into a series of mass points connected by the springs consumption of which was about 145 h.
and dampers. And their movement is controlled by flow field
2. Mathematical model
force, spring force and damping force.8–10 The first application
of parachute based on this method can be found in Ref.8 and
2.1. Control equation of fluid field
in that work the inflating process of the 2D axial symmetry
round parachute surrounded by flow field was calculated. This
method focuses on the canopy shape calculation, but the stress At the inflation stage of the flexible canopy, the flow field grids
distribution of the canopy cannot be obtained. At present, one near the structure will produce large deformation. In order to
of the most commonly methods is the deforming-spatial-do- avoid serious distortion of the calculation grids, the simplified
main/stabilized space–time (DSD/SST) coupling method, arbitrary Lagrange–Euler (SALE) method is used to calculate
which can obtain the detailed description of the flow field the inflation stage of the parachute.
and structure during the parachute inflation process. This is The control equations of flow field are as follows:
8
an interface tracking method aiming to solve the large defor- > @q @vi @q
>
> þq þ ci ¼0
mation in coupling calculation, which was proposed in the >
> @t @xi @xi
<
1990s.11,12 In DSD/SST method, both flow field and structure @vi @vi ð1Þ
are dispersed by a large number of elements. The time–space >
> q þ qci ¼ rij;j þ qbi
>
> @t @x j
FSI problem was solved using dynamic grid technology. The >
:
i; j ¼ 1; 2; 3
method was firstly used in Ref.13 for 2D computation of the
flat circle parachute during the inflation process. The same where q is the density, t the time, vi the velocity, xi the Eulerian
method was also used in 3D FSI calculation of parafoil, circu- coordinate, ci the relative velocity in reference field, bi the body
lar parachute, cross parachute and ringsail parachute.14–19 The force; stress rij = pdij + l(vi,j + vj,i), p is the pressure, l the
initial unfolded shape was greatly simplified in those works. viscosity, dij the Kronecker coefficient.
Besides the DSD/SST technology, the immersed boundary Initial conditions: v0F ¼ vinlet ; p0F ¼ 101325 Pa, the super-
(IB) method was used to solve the FSI simulation problems script ‘‘0’’ means the initial time. Boundary conditions of the
of parachute inflating, and this method was used in simulating calculation field: inlet boundary, vinlet ¼ v; outlet boundary,
the cone parachute and cluster parachute.20 These methods @pF @vF @vF
¼ 0; ¼ 0; the around boundary, ¼ 0:
mentioned above are based on Navier–Stokes equations, while @n @n @n
the other method, such as large eddy simulation method, was
2.2. Structure control equation
used for FSI simulation of disk-gap-band parachute in super-
sonic flow field in Ref.21 In recent years, studies of the three-
dimensional FSI problem of the parachute based on the busi- The control equation of the structure calculation (coupling
ness software LS-DYNA also made great progress.22–24 The faces) is as follows:
FSI calculation on the quasi-inflation process of the slotting d2 ui
qS 2 ¼ rij;j þ qS bi ð2Þ
circular parachute, rectangular parachute and crucial para- dt
chute was studied in Refs.23,24, and the initial shape was sup- Initial conditions: v0S ¼ 0; r0S ¼ 0: Boundary condition of
posed to be a spreading shape. The same method is also the lines intersection: uS = 0.
used in simulating the inflation process of flat circular para- When the canopy elements contact each other, the contact
chute and bottom-stretching parachute in Refs.22,24, in which force F = kd is exerted. The penalty function coefficient k is
the structural details of parachute such as radial and zonal affected by the distance of coupling point d and stiffness coef-
band were ignored. The initial structural model of the para- ficient of fabric. Penalty method based on Ergun equation is
chute obtained by calculation is called load hung method. also used for the realization of FSI.
The canopy model will move from a fattened state to a folded
state when the tension load was applied to some nodes. If the 3. Numerical arithmetic
parachute is made from many types of materials or the struc-
ture is complex, the initial grid of canopy will be distorted and In the actual wind tunnel experiment, the intersection of lines
the calculation will become more complex. is fixed. Therefore the change of flow field computational do-
274 L. Yu et al.

main is limited. In order to simulate the conical ringslot inflat- 3.2. Smoothing grids
ing in this state, the referenced grid of flow field in SALE/FSI
method can be fixed in this state. In this work, the inlet velocity is constant and the referenced
The parachute structure simulation method is mainly based grids of the flow field are fixed during parachute inflation.
on Lagrangian arithmetic. The calculation grids are fixed on After the Lagrangian time step calculation of the flow field is
the structure and move with the material. Thus the informa- carried out, the fluid material grids will be distorted. Thus they
tion of mobile boundary of the parachute can be described must be smoothed by equipotential method which was devel-
accurately. oped by Winslow25 in order to guarantee the grid quality.
The inlet velocity is constant and set as 18 m/s in the infinite Equipotential zoning is a method to establish a structured
mass state of parachute inflation process. The SALE/FSI mesh for finite difference or finite element calculations by solv-
method is processed to solve the serious distorted flow field ing the Laplace equations (later extended to Poisson equation).
grids during calculation. The flow field calculation is mainly The topological relation of the smoothed mesh is the same as
based on splitting technique. For each step forward in time, the original one. The new nodal position of mesh is determined
there are three distinct sub-steps: (A) solving the flow field by the following equation based on equipotential arithmetic:
based on Lagrangian arithmetic, and getting state parameters
in Lagrangian coordinate; (B) smoothing the distorted fluid a1 @ n1 n1 x þ a2 @ n2 n2 x þ a3 @ n3 n3 x þ 2b1 @ n1 n2 x þ 2b2 @ n1 n3 x
material grids based on ‘‘Equipotential’’ zoning method, in þ 2b3 @ n2 n3 x ¼ 0
which the topology relationship of grids does not change;
(C) transporting the flow field information to the updated grids where
based on advection methods. These sub-steps are shown in ai ¼ @ ni x21 þ @ ni x22 þ @ ni x23
Fig. 1. At the end of Step (C), the solution is prepared for b1 ¼ ð@ n1 x  @ n3 xÞð@ n2 x  @ n3 xÞ  ð@ n1 x  @ n2 xÞ@ n3 x2
the next time step.
b2 ¼ ð@ n2 x  @ n1 xÞð@ n3 x  @ n1 xÞ  ð@ n2 x  @ n3 xÞ@ n1 x2
3.1. The Lagrangian arithmetic b3 ¼ ð@ n3 x  @ n2 xÞð@ n1 x  @ n2 xÞ  ð@ n3 x  @ n1 xÞ@ n2 x2
In order to obtain more even mesh, the Simple Averaging
In the Sub-step (A), the grids move with the material deforma- arithmetic is also used. The coordinates of a node are the sim-
tion. Therefore the convective terms and the mass equation in ple average of the coordinates of its surrounding nodes as
the governing Eq. (1) can be ignored. The control equation of follows:
the flow field and structure can be simplified as follows: 1X N
nþ1
xSA ¼ xn ð6Þ
ui þ Fint
M€ ext
i ¼ Fi ði ¼ 1; 2; 3Þ ð3Þ N i¼1 i

where, the three terms stand for inertia force, internal force A composite algorithm is generated using a weighted aver-
and external force respectively from left to right, M is the mass age of those two algorithms above, where the subscripts ‘‘E’’
and u€i the acceleration. and ‘‘SA’’ refers to the equipotential and simple averaging
The nodal displacement is calculated based on the discrete algorithm respectively, and w is the weighting factor.
Eq. (3) by using finite element method (FEM). After calculat-
xnþ1 ¼ wE xnþ1
E
nþ1
þ wSA xSA ð7Þ
ing the nodal acceleration, centered-differences are used to find
the velocity at time n + 1/2, and the new position of the node
at time level n + 1, n is the time step. 3.3. Advection algorithms
Then the stress, strain and other parameters of elements can
be got at time n + 1. In general, the time step Dt must be cho- As a fluid element moves from its position at the end of the La-
sen small enough to guarantee calculation stability. In this pa- grange calculation to its new position after mesh smoothing,
per, the time step is chosen as follows: there are mass, momentum and energy flux through the six
Dt < Dtc ¼ aðLmin =cÞ ð4Þ faces of the element. Since advection is merely a geometric up-
date of the mesh, the source terms of the governing equation
where Lmin, c, a stands for the minimum cell size, sonic speed disappear. The advection equation can be described as follows:
of the material and zooming factor respectively (a = 0.7 in this
work). @/ @ci /
þ ¼0 ð8Þ
@t @xi
where / is the generalized variable.
As for the three-dimensional problem, the convection term
is discrete which is expressed by the following common form:
X
6
/nþ1
A VA
nþ1
¼ /nA VAn þ fj / ð9Þ
j¼1

8
>
> cj jcj j 
>
> f / ¼ ð/ þ /þ j Þþ ð/j  /þ j Þ
< j 2 j 2
þ
/j ¼ Sjþ1 ðxj þ 2 Dtcj  xjþ1 Þ þ /njþ1
n n 1 n ð10Þ
>
> 2 2
>
> 
2
: /j ¼ Snj1 ðxnj þ 12 Dtcj  xnj1 Þ þ /nj1
2
Fig. 1 Operation of each time step. 2 2
Study of parachute inflation process using fluid–structure interaction method 275

Table 1 Parameters of textile material.

Structure Density (g/m3) Thickness/diameter (mm) Poisson ratio Width (mm) Young’s modulus (Pa)
Canopy 130 0.03 0.14 5.87 · 109
Suspension 7 2.5 6.52 · 108
Reinforcement belt 5 0.1 15 3.42 · 109

8    
>
> /njþ3  /njþ1 Dx2j þ /njþ1  /nj1 Dx2jþ1
>
> n  2
>
< Sjþ12 ¼
2 2 2

Dx Dx j Dx þ Dx
jþ1 j jþ1
ð11Þ
>
>
> Dxj ¼ xnjþ1  xnj1
>
>
2 2
: Dxjþ1 ¼ xn 3  xn
jþ2 jþ12

where, the expression of Snj1 is the same as Eq. (11). The sub-
2
script ‘‘j’’ refers to calculation element and its surrounding ele-
ment respectively. V is element volume. The value of fj/
depends on the velocity at node j, which defines the upstream
direction. The staggered mesh algorithm and monotone up- Fig. 3 Initial model of parachute.
wind schemes for conservation laws (MUSCL) scheme with
the second order precision are applied to calculating the flux
of mass, momentum and energy through the element faces.

4. Structure and numerical model

4.1. Parameters of the parachute

A kind of conical ringslot is involved in this work. The para-

chute is made up of 20 gores and 5 rings. The textile fiber mate-
Fig. 4 Initial grid of parachute and flow field.
rials of the canopy, suspension, and reinforcement line are
brocade silk, aramid and brocade silk respectively. And the
material parameters are shown in Table 1. The symmetrical
section of the ringslot from a top view is shown in Fig. 2, in folding method also makes it convenient to program and to
which the AA0 is the symmetrical line, the apex angle is establish folded numerical model. The canopy is meshed by
16.9, the vent diameter is 0.54 m, the nominal area is 30 m2, 36200 triangular elements. The suspension and reinforcement
and suspension length is 7.4 m. belt are meshed by bar elements with 4760 nodes. The hexahe-
dral elements are used to mesh the flow field, and flow field
4.2. Finite element model grids are over 690000. The computational domain along the
axial, radical and zonal is 10D0 · 5D0 · 5D0 respectively (D0
is the nominal diameter, and the length of wake flow is about
The direct folding modeling technology26,27 was applied to
6D0). The initial element model of the flow field and parachute
establishing the initial finite element model of parachute with
is shown in Fig. 4.
completely straightened suspensions. The canopy looks like
‘‘*’’ from overlooking (see Fig. 3). The direct folding modeling
method is simple and does not need calculation compared with 5. Numerical results and discussion
load hung method. Therefore more structure details including
reinforce belt, geometric porosity, etc. can be described. More- 5.1. Flow field and shape of the structure
over, the folded extent and the initial inlet area can be adjusted
conveniently. In this work, the folded angel of the canopy is Fig. 5 shows the time-sequence results of coupling calculation
only 1 and the initial inlet area is just 0.085 m2. The direct such as the canopy shape change including stress contour,
pressure and velocity of the flow field in a symmetrical section.
The following conclusions can be obtained according to the
numerical results.

(1) The canopy skirt opens first, and then the canopy
inflates from apex to skirt, which reflect the influence
of airflow entered into the canopy. The space–time
change of canopy stress is also decided by the space–
time deformation of canopy. The canopy is not com-
pletely symmetric inflated because of the asymmetric
Fig. 2 Top view of conical ringslot model (half). flow (see Fig. 5).
276 L. Yu et al.

Fig. 5 Structure and flow field of time-sequence during inflation process (left: canopy shape with stress contour; right: pressure contour).

(2) During the inflation process, the maximum effective around which is similar. At the main inflation stage,
stress appears at the initial inflation phase, the 0.42 s. each seam works as the outlet to reduce the opening
The position is at the middle part of the second ring shock and inflation speed of the canopy with the increas-
which departs form the apex at about one-third of the ing internal pressure (see Fig. 5(b) and (c)).
radius length. The canopy will not be damaged even (4) The wake vortex begins to develop near the canopy sur-
with the maximum effective stress (1.907 · 108 Pa). face at first. Then it moves downstream, and at the
(3) At the initial inflation stage, each seam between rings meantime, becomes bigger and longer. Finally, the vor-
can be an assistant inlet and accelerate the inflation, tex breaks into pieces and moves far away (see Fig. 5(b)
and the velocity and pressure distribution of flow field and (c)).
Study of parachute inflation process using fluid–structure interaction method 277

respectively. Then the initial inflation time and the fully in-
flated time are 0.46 s and 2.00 s, which are shorter than the
ones in wind tunnel experiment (0.58 s and 2.17 s respectively).
The canopy is soft, light and large, and the inflation status of
parachute is greatly influenced by air flow, package manner,
manufacturing process of parachute and other factors. In the
actual situation, which is not the ideal state, the canopy infla-
tion is usually asymmetrical but not synchronized. Therefore
the inflation time will be prolonged significantly in tests. In
short, the numerical results such as inflation time and pro-
jected diameter can satisfy the demand of engineering analysis.

5.3. Swinging
Fig. 6 Numerical results of projected diameter.
A wind tunnel test was carried out and its state was set the
same as the calculation state (see Fig. 7). The force balance
was used to measure parachute opening shock load, and the
binocular camera method was used to obtain the canopy
swinging change. According to definition of Descartes’s coor-
dinate system, the vertical, horizontal and the central axis
direction of the parachute is defined as X, Y, Z coordinate
directions respectively. Numerical calculation results of swing-
ing are shown in Fig. 8 and the results comparison of numer-
ical simulation and experiment is shown in Table 2. The
experimental swinging angle is slightly bigger because it is dif-
Fig. 7 Wind tunnel test. ficult to keep the flow field in wind tunnel absolutely symmetric
and the canopy structure is also not symmetric due to the man-
ufacturing process and design. However, the numerical swing-
ing angle is consistent with the result of wind tunnel test, and
both results show that this parachute has good stability.

5.4. Opening shock and drag characteristics

It can be found that the numerical results of drag area and

opening shock are consistent well with the wind tunnel test
data in Table 2. The drag area and opening load of the infla-
tion process are shown in Figs. 9 and 10. Moreover, Fig. 10
shows the experience formula which is based on the inflation
time method.7 The results show that:

(1) During the same parachute inflation state, the change of

Fig. 8 Swinging of the parachute during inflation process. opening load is similar to the drag area, but this corre-
lation is nonlinear. Because the opening shock of para-
chute is not only related to drag area but also related
to the change rate of drag area.
5.2. Inflation time and the projected diameter (2) Both the maximum load and maximum drag area occur
at the time when the canopy is completely open, which
In Fig. 6, the evenly projected diameter of the vent and skirt is follows the change regulation of inflation process in infi-
0.52 m and 3.90 m at terminal stage respectively. The ratio of nite mass condition. Both results based on FSI method
the inlet projected diameter and nominal diameter is 0.63. It and experimental method appear the first opening load
is consistent with the statistical data which ranges from 0.60 peak at the time when the vent is completely open (about
to 0.65.7 The start time of the vent opening and skirt opening at 0.50 s). But this phenomenon cannot be reflected by
is defined as the initial inflation time and fully inflated time inflation time method.

Table 2 Results comparison of numerical simulation and experiment.

Value Vertical maximum Horizontal maximum Drag area at fully Opening shock (N)
swinging angle () swinging angle () inflated state (m2)
Test 3.13 3.09 18.2 3724
Calculation 2.69 2.01 18.4 3890
278 L. Yu et al.

process and is proven to be able to solve the large deformation

problem of parachute. The numerical results are in good agree-
ment with the test data, which shows that this method has high
accuracy. Moreover, it can provide more space–time fluid–
structure–motion information such as the canopy shape, the
opening shock, swinging, etc. compared with the empirical
method, and can better reveals the mechanical mechanism of
parachute inflation process. All the results show that the
space–time SALE method is recommendable in numerical sim-
ulation of parachute FSI problem and can satisfy the following
purposes:

(1) Predicting time-sequence change of the canopy shape

Fig. 9 Drag area of ringslot. with stress and providing reference for the material
selection and structure optimization.
(2) Predicting aerodynamic characteristics of parachute and
providing reference for the payload–parachute system
design and trajectory study.
(3) Predicting the swinging and stability of the parachute
and taking further researches on the stability and twine
problems of the payload–parachute system.

In a word, the FSI method of the parachute inflation pro-

cess is a complete theoretical method which does not depend
on empirical coefficient. Based on this method, the interactive
mechanical mechanism of fluid force and structural force dur-
ing parachute working can be revealed and more space–time
information of fluid–structure–motion can be obtained.
Fig. 10 Opening load of ringslot.
Acknowledgments

Affected by the preceding air flow peak, the canopy opens The authors are grateful to Yan Xiaoxue for her careful polish-
gradually from bottom to apex. The flexible canopy has light ing. This study was co-supported by the National Natural Sci-
mass and cannot support the bending. Therefore, at the initial ence Foundation of China (No. 11172137) and the
inflation stage, the upper folded part of the canopy will pro- Aeronautical Science Foundation of China (No.
duce some second folding due to its swinging and twisting 20122910001).
aroused by the fluid force and structural force. This may cause
the parachute opening to be difficult. When the apex is com-
References
pletely open the second folding disappears rapidly, and the
opening load will jump because of the rapid growth of deploy
1. Orlik-Rueckemann KJ, Maydew RC, Peterson CW. Design and
area. Therefore the first opening shock is influenced by the sec- testing of high-performance parachutes. NASA-ADA-246343;
ond folding of the upper canopy. The first opening shock 1992.
appearing at initial inflation stage reflected by numerical sim- 2. Wolf N, Whitmore T, Jensen S. Design and testing of collapsible
ulation or experimental method reveals the interaction mecha- drogue parachute for the X-37 Vehil. AIAA-2005-1660; 2005.
nism of flow field and structure, which is consistent with the 3. Jason PS, Hart A. Canned telemetry system for sensor data on
actual situation. However, the time-sequence drag area is as- parachute system. AIAA-2005-1513; 2005.
sumed continuous and smooth by the inflation time method,7- 4. Yu L, Ming X, Hu B. Experimental investigation in parachute
which does not agree with the practical situation. Moreover, opening process. J Nanjing Univ Aeronaut Astronaut
2006;38(2):176–80 [Chinese].
the empirical coefficient selected may not be appropriate for
5. Yu L, Zhang XH, Li SS. Experimental on canopy payload
the specified geometric dimension of parachute or specified
performance of parachute. J Beijing Univ Aeronaut Astronaut
working state. 2007;33(10):1178–81 [Chinese].
6. Peng Y, Zhang QB, Cheng WK. Summary for the research
6. Conclusion methods of parachute inflation process. Chin Space Sci Technol
2003;7–12 [Chinese].
7. Wang LR. Parachute theory and applications. Beijing: Aerospace
In this paper, the direct folding method is applied to establish-
Press; 1997. p. 187–229.
ing the initial folded parachute model and shows that: (A) this 8. Stein KR, Benney RJ, Steeves EC. A computational model that
method is convenient in modeling the structure details; (B) the couples aerodynamic structural dynamic behavior of parachutes
application is simple without calculation compared with the during the opening process. NASA-ADA-264115; 1993.
load hung method; (C) it is easy to adjust the folding rate 9. Yu L, Ming X. Study on transient aerodynamic characteristics of
and inlet area. Then, the space–time FSI technique based on parachute opening process. Acta Mech Sin 2007;23(6):627–33
SALE method is used to simulate the parachute inflation [Chinese].
Study of parachute inflation process using fluid–structure interaction method 279

10. Yu L, Shi XL, Ming X. Numerical simulation of parachute during 19. Kenji T. Fluid structure interaction modeling of spacecraft
opening process. Acta Aeronaut Astronaut Sin 2007;28(1):52–7 parachutes for simulation-based design. J Appl Mech
[Chinese]. 2012;79(1):1–9.
11. Tezduyar TE, Behr M, Mittal S. A new strategy for finite element 20. Kim YS, Peskin CS. 3-D parachute simulation by the immersed
computations involving moving boundaries and interfaces––the boundary method. Comput Fluids 2009;38(6):1080–90.
deforming-spatial-domain/space–time procedure: I. The concept 21. Karagiozis K, Kamakoti R, Cirak F, Pantano C. A computational
and the preliminary numerical tests. Comput Methods Appl Mech study of supersonic disk-gap-band parachutes using Large-Eddy
Eng 1992;94(3):339–51. simulation coupled to a structural membrane. J Fluids Struct
12. Tezduyar TE, Behr M, Mittal S. A new strategy for finite element 2011;27(2):175–92.
computations involving moving boundaries and interfaces––the 22. Tutt BA, Taylor AP. The use of LS-DYNA to simulate the
deforming-spatial-domain/space–time procedure: II. Computation inflation of a parachute canopy. AIAA-2005-1608; 2005.
of free-surface flows, two-liquid flows, and flows with drifting 23. Tutt BA, Taylor AP, Jean CB, Gargano B. The use of LS-DYNA
cylinders. Comput Methods Appl Mech Eng 1992;94(3):353–71. to assess the performance of airborne systems North America
13. Stein KR, Benney RJ, Kalro V, Johnson AA, Tezduyar TE. candidate ATPS main parachutes. AIAA-2005-1609; 2005.
Parallel computation of parachute fluid–structure interactions. 24. Coquet Y, Bordenave P. Improvements in fluid structure interac-
AIAA-1997-1505; 1997. tion simulation of parachute using LS-Dyna. AIAA-2011-2590;
14. Kalro V, Tezduyar T. A parallel 3D computational method for 2011.
fluid–structure interactions in parachute systems. Comput Methods 25. Winslow AM. Numerical solution of the quasilinear Poisson
Appl Mech Eng 2000;190(3–4):321–32. equation in a nonuniform triangle mesh. J Comput Phys
15. Stein K, Benney R, Kalro V, Tezduyar TE, Leonard J, Accorsi M. 1997;135(2):128–38.
Parachute fluid–structure interactions: 3-D Computation. Comput 26. Ma CS, Yue H, Huang SL. A study on airbag folding patterns for
Methods Appl Mech Eng 2000;190(3):373–86. improving occupant protection effectiveness. Automot Eng
16. Stein K, Benney R, Tezduyar T. Fluid–structure interaction of a 2005;27(3):350–3.
cross parachute: numerical simulation. Comput Methods Appl 27. Cheng H, Yu L, Yin ZW. A new method of complicated folded
Mech Eng 2001;191(6–7):673–87. fabric modeling. J Harbin Inst Technol 2012;19(2):43–7.
17. Tezduyar TE, Sathe S, Pausewang J, Schwaab M, Christopher J,
Crabtree J. Fluid–structure interaction modeling of ringsail Yu Li received Ph.D. degree from Nanjing University of Aeronautics
parachutes. Comput Mech 2008;43(1):133–42. and Astronautics (NUAA) in 2006. Now, she is a professor and doc-
18. Takizawa K, Wright S, Moorman C, Tezduyar TE. Fluid– toral tutor in College of Aerospace Engineering, NUAA. Her main
structure interaction modeling of parachute clusters. Int J Numer research interests are fluid structure interaction, ADS design, and
Methods Fluids 2011;65(1–3):286–307. environmental control systems.