Hindawi Publishing Corporation

International Journal of Aerospace Engineering

Volume 2012, Article ID 182907, 17 pages
doi:10.1155/2012/182907

Research Article
Parachute-Payload System Flight Dynamics
and Trajectory Simulation

Giorgio Guglieri
Dipartimento di Ingegneria Meccanica e Aerospaziale, Politecnico di Torino, Corso Duca degli Abruzzi 24, 10129 Torino, Italy

Correspondence should be addressed to Giorgio Guglieri, giorgio.guglieri@polito.it

Received 15 November 2011; Revised 2 February 2012; Accepted 14 March 2012

Academic Editor: C. B. Allen

Copyright © 2012 Giorgio Guglieri. This is an open access article distributed under the Creative Commons Attribution License,
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

The work traces a general procedure for the design of a flight simulation tool still representative of the major flight physics of
a parachute-payload system along decelerated trajectories. An example of limited complexity simulation models for a payload
decelerated by one or more parachutes is given, including details and implementation features usually omitted as the focus of
the research in this field is typically on the investigation of mission design issues, rather than addressing general implementation
guidelines for the development of a reconfigurable simulation tool. The dynamics of the system are modeled through a simple
multibody model that represents the expected behavior of an entry vehicle during the terminal deceleration phase. The simulators
are designed according to a comprehensive vision that enforces the simplification of the coupling mechanism between the payload
and the parachute, with an adequate level of physical insight still available. The results presented for a realistic case study define the
sensitivity of the simulation outputs to the functional complexity of the mathematical model. Far from being an absolute address
for the software designer, this paper tries to contribute to the area of interest with some technical considerations and clarifications.

1. Introduction There are a number of different kinds of parachutes that

have been designed for various applications. The different
The purpose of a parachute is to decelerate and provide applications parachutes are typically used for pilot, drogue,
stability to a payload in flight. The aerodynamic and stability deceleration, descent, extraction, supersonic drogue and
characteristics of the parachute system are governed by the stabilization, flight termination, and landing.
geometry of the parachute as such careful consideration is The dynamics of parachutes are complex and difficult
paid to this in the design process. The effects of deployment to model accurately. During both the inflation process and
and opening force are critical in the safe operation of the the terminal descent stage, the dynamics of a parachute are
parachute and the integrity of the payload. The opening char- governed by a coupling between the structural dynamics of
acteristics also feature heavily in the selection of geometry the parachute system and the surrounding fluid flow. Both
and other parameters in the design process. of these dynamic systems must be addressed as a coupled
Parachutes for aerospace applications [1–4] are in general system to gain a proper representation of the dynamic system
symmetric about the canopy axis. This axis passes through as a whole.
the center of the canopy and the confluence point of the When the parachute is in a steady state, the air flowing
suspension lines. The canopy is the cloth surface that inflates around the decelerator will separate at some location on
to provide the desired lift, drag, and stability. The suspension the canopy. The shedding of the vortices from the canopy
lines transmit the retarding force from the canopy to the can affect the stability and cause a periodic motion of both
payload either directly or through a riser attached below the parachute and payload. The wake from a porous parachute
confluence point of the suspension lines. The deceleration consists of air that flowed around the canopy and air that
force may be distributed on the payload over more than one flowed through the canopy. A payload body in the speed
mechanical joint linked to the riser by a set of short hardly range of parachute usage sheds a very turbulent wake. Part
extensible strips (bridles). of the flow that is entering the parachute is therefore of a
2 International Journal of Aerospace Engineering

disturbed nature and should be considered regarding the entry into the Martian atmosphere. The examples of capsules
aerodynamic performance of the parachute. For many types are well-known ones due to the intense space activities from
of parachutes, this change in oncoming airflow can be the former USSR and the USA. The capsules that were
quite significant during the time required for the parachute developed for manned flights in this period were Mercury,
to inflate. The implication of a rapid deceleration is that Gemini, Apollo and Soyuz, which are still considered a
second-order effects are likely to be present. reference for performing missions to this day. As was the
To summarize, calculation of parachute deployment, case with the large angle cones, there were found to be speed
inflation, and deceleration requires the numerical solution ranges over which the capsules were unstable [9].
to the equations of motion for a viscous, turbulent, separated The flow field that is associated with the capsule config-
airflow. The parachute is also a flexible body having dynamic uration is highly complex. For almost the entire speed range
behavior coupled with the behavior of the flow, which passes that the capsules operate over, the flow remains attached on
through and around it. From the above description it is the forward face. At the point of maximum diameter, the
obvious that a full-time dependent solution of this system is flow is accelerated such that the boundary layer rapidly grows
far from being easily feasible. To make a mathematical model to the point of separation. After the maximum diameter,
that is feasible, simplifications must be made, as long as the the flow then remains separated and turbulent. This flow
model can be validated satisfactorily by experiment or by is unstable and coupled with an unsteady near wake the
comparison with reference data. dynamic instability associated with capsules is produced
The overall behavior of parachutes is related to var- [11].
ious parameters: added masses, filling time, parachute As dynamic instability exists for blunt bodies over various
shape (inflated canopy elongation), porosity, suspension line speed ranges, to complete missions successfully, there is then
length, reefing, clustering, snatch loads at deployment, and a requirement for some kind of accurate stability assessment
aero-mechanical and inflation instability. In the past, most with a potential impact on both stability augmentation (if
of these effects could be generally modeled in an imprecise any reaction control system is implemented) and mission
way by simulation tools. A comprehensive computational design (parachute deployment sequence). This issue applies
technique is presented in [5–7] for carrying out three- for probes and capsules as thrusters and parachutes are the
dimensional simulations of parachute fluid-structure inter- unique available sources of additional damping.
actions, and this technique is applied to simulations of The measurement of stability derivatives for probes and
airdrop performance and control phenomena in terminal capsules was a concern of designers since the origin of space
descent. The technique uses a stabilized space-time for- flight [9], as analytical methods did not provide (at least in
mulation of the time-dependent, three-dimensional Navier- the past) adequate estimation of these relevant parameters.
Stokes equations of incompressible flows for the fluid The wind tunnel experimental techniques considered are
dynamics part. A finite-element formulation derived from free oscillation, forced oscillation, free flight tests, and even
the principle of virtual work is used for the parachute ballistic range experiments. It should be noted that wind
structural dynamics. The parachute is represented as a cable- tunnel experiments are not without limitations, so there
membrane tension structure. Coupling of the fluid dynamics remains a strong desire to make available numerical data, for
with the structural dynamics is implemented over the fluid- the purposes of both finding solutions and comparing with
structure interface, which is the parachute canopy surface. experimental results.
According to the different missions, several types of
payloads have been used in combination with aero-
dynamic decelerators: paratroops, equipment, hardware, 2. Background on Parachute-Payload
materiel, weapons, missiles, aircraft, unmanned aerial vehi- Modeling and Simulation
cles, aerospace lifting, and nonlifting spacecraft. The present
analysis is focused on aerospace applications for planetary Several studies analyze the descent and landing trajectories of
and atmospheric entry vehicles, where the payload is typi- parachute-payload systems. Generally, these analyses consist
cally a blunt body. The purpose of a blunt body is primarily of performing a simulation (typically a 3–6 DOFs rigid
to provide a large source of drag to facilitate a deceleration. body model is adopted) of the atmospheric entry phase to
Applications of blunt bodies can be seen with both manned predict deceleration peaks, descent attitudes, and terminal
reentry and planetary exploration missions. An outline of conditions. In addition, a stochastic dispersion analysis
aerodynamic decelerators for robotic planetary exploration (Monte Carlo or similar) is usually performed to assess
missions is given in [8]. In the development programs of the impact of off-nominal conditions that may arise to
such blunt bodies and also in subsequent studies, it has been determine the robustness of the mission design.
shown that they may be dynamically unstable in all or part of A two-dimensional parachute model is presented in [12]
the sub-, trans-, super-, and in some cases hypersonic speed to compute the various characteristics of the steady descent
regimes. of a parachute system. A three degree-of-freedom analysis
Blunt bodies for which there exist examples can be is presented and validated in [13] giving the longitudinal
classified into two categories, large angle cones and capsules motion of a typical vehicle during the recovery phase. The
[3]. An example of large angle cone that has performed actual parachute and the payload are supposed to be rigid and
missions is the Viking probe. It should be noted that in the interconnected by an elastic riser. Aerodynamic loads act-
two Viking missions confirmed the dynamic instability upon ing on the two subsystems are considered. Computer
International Journal of Aerospace Engineering 3

results showed good agreement with test results in terms available in terms of sensitivity of system performance met-
of oscillation amplitude and frequency, riser force, and rics to design parametric changes. In modeling a payload-
parachute wrapup about the vehicle for the simulation of parachute system, two bodies and one device have been used.
a pad-abort situation. The three-dimensional motion of a They are the parachute canopy and the payload, which are
freely descending parachute is studied in [14] with a five then connected by a single riser. In reality the parachute
degree-of-freedom analysis (roll motion is neglected). Exact canopy is connected to the payload by suspension lines, a
expressions are given for the longitudinal and lateral small riser section, and then a set of bridles. In the simulation only
disturbance stability of the gliding motion of parachutes. The the riser and the bridles are modeled in terms of a dynamic
analysis confirms that large longitudinal disturbance of most response (suspension lines are neglected). However, in terms
parachutes will result in a large pitching motion, whereas of connection points only the riser has been modeled. The
a large lateral disturbance will usually cause a large angle riser is assumed to have flexible connections at both ends and
vertical coning motion (coning mode). The longitudinal provides stiffness above a certain threshold distance and zero
mode damps out very quickly in the stable case. The three- stiffness below this distance (slack conditions).
dimensional motion of a nonrigid parachute and payload The payload is considered as a rigid body with six degrees
system is studied in [15]. Both the parachute and the payload of freedom. The forces and the moments that act on the
are assumed to have five degrees of freedom (roll about body are provided by its aerodynamics, by the weight, by the
axis of symmetry is again neglected). They are coupled inertial actions, and, finally, by the force applied by the riser
together by a fixed-length connector. The general nonlinear in the suspension point. The parachute canopy acts as a rigid
equations are linearized using small perturbation theory. The body and, by the action of aerodynamic drag, strains the riser
evaluation of the stability of an unstable payload decelerated which then transmits a force to the payload. Figure 1 shows
by a parachute is performed. The authors observed that the general philosophy used in modeling the system.
increasing riser length and parachute weight promotes sys- In terms of a dynamic response, the riser is modeled as
tem instability. A nine degree-of-freedom computer program having linear stiffness and damping, where the force at any
was developed in [16] for the simulation of the trajectory given point in time is given by
and the dynamic behavior of a rotating parachute system.
An accurate mathematical model of the joint between the FR = klr ε + cε̇. (1)
load and the parachute was found to be necessary to predict
Note that k is stiffness, lr is riser length, ε and ε̇ are strain and
the dynamic behavior of a rotating decelerated system. A
strain rate, respectively, and c is the damping coefficient. The
computer model based on a six degree-of-freedom analysis
above equation is implemented in the code to calculate the
is described in [17] and compared with drop test data. The
force present in the riser at any point. The equivalent stiffness
payload is rigidly connected, the aerodynamic forces on
of the bridles and riser is given by the following equation:
canopy and payload are determined by the instantaneous
angle of attack of the impinging airflow, the apparent nb · kr kb
masses are constant, but they depend on the direction of k= , (2)
kr + nb · kb
the acceleration. Full nonlinear equations of motion for
the axisymmetric parachute have been obtained in [18]. In where nb is the number of bridles, kr is riser stiffness and kb
particular, the correct form of the added mass tensor for is single bridle stiffness. After calculation of the equivalent
a rigid axisymmetric parachute in ideal flow has been im- stiffness, the damping coefficient c can be calculated by the
plemented in a six degree-of-freedom computer model following formulation (mP is the parachute mass and ξ is the
[19], and the results indicate that added mass effects are damping ratio):
significant. In particular, the component of added mass 
along the axis of symmetry has a strong effect on parachute k
c = 2ξmP . (3)
dynamic stability. However, design and testing experience mP
shows that dynamic stability of the parachute is a second
The values for strain and strain rate are computed by the
order design problem [3] for high-performance decelerators
difference in displacement and velocity of each end of the
that usually have both high static and dynamic stability due
riser, that is, as the displacement and the velocity of the
to the porosity of the canopy. Many works, including recent
payload and that of the parachute canopy. Also note that
models developed and validated by means of drop tests,
where this calculation occurs a check is in place for the
consider the bridles as rigid elements [20, 21].
condition where the riser is slack:
sc − s p − l0
ε= ,
3. The Mathematical Model of the lr (4)
Payload-Parachute System ε̇ = Vc − V p ,
A first mathematical representation of a parachute-payload where ε and ε̇ are, respectively, strain and strain rate, sc and
system (designated as SM1) is defined according to a very s p are, respectively, payload and parachute displacement, lr is
comprehensive approach that enforces the simplification the riser length, and Vc and V p are, respectively, the payload
of the coupling mechanism between the payload and the and parachute velocity. From these two calculations, the force
parachute, with an adequate level of physical insight still in the riser can be derived by application of (1).
4 International Journal of Aerospace Engineering

Aerodynamic
information

Aerodynamic
mathematical
model

Velocity and Forces and

angular velocity moments

Initial conditions Inertial equations of

motion Time history

Forces and
Required states
moments

Parachute model

Figure 1: Flow chart of the simulation model.

The performance of the parachute has been modeled whereas in the model used in the simulation the volume of
using an approximation for the added mass ma . When an ellipsoid is used. Hence, added mass has been calculated
a parachute passes through the air, it drags some of the using the following expression:
surrounding with it this means that this air must be
4
accelerated from rest up to the speed of the parachute. The ma = nP ka ρ πR2 h, (9)
parachute will also accelerate some of the air immediately in 3
front and behind. The energy that is required to facilitate where nP is the number of parachutes, ka is the added mass
this acceleration is taken from the kinetic energy of the coefficient, R is the inflated radius, and h is the inflated
parachute-payload system. These fluid inertia effects are height. The integer nP for the number of parachutes is
taken into account in the equation of motion. The idea is necessary because a cluster of multiple parachutes is being
that the mass has a component added to take into account modeled as one parachute of equivalent size. The added
the extra mass associated with the accelerated air. This can be mass remains constant throughout inflation, so the values for
seen below, where the change in momentum of the parachute parachute radius and height are constant, whereas in reality
system is equal to the sum of the drag force and gravitational they are of course changing quite significantly through the
force component: inflation process. The added mass coefficient is given by:
 
d 1 ka = 1.068 1 − 1.465p − 0.25975p2 + 1.2626p3 , (10)
[M ∗ V ] = − ρV 2 CD SP + mP sin γ, (5)
dt 2
where p is the parachute porosity (see [3, 22] for further
M ∗ = ma + mP . (6) details concerning the effect of porosity on added masses).
This coefficient remains constant through inflation. The
After combining (5) and (6) and then rearranging, the fol- porosity for both drogue and main parachutes is set at a value
lowing expression for system acceleration (or deceleration) of 20% (p = 0.2).
can be used: The significance of fluid inertial effects is relevant. As
dV 1 dma an example, opening times and loads increase with altitude.
(mP + ma ) = − ρV 2 CD SP − V + mP g · sin γ. (7) The reason for this can be seen when it is considered that at
dt 2 dt
increasing altitude the true airspeed will increase for constant
However, for (7) to be useful, an expression or knowledge of indicated airspeed, as such the inertial effects of the air are
ma (t) must be available. Shown below is an expression for greater. After comparison with flight test data, the added
added mass, similar to that for apparent mass, that has been mass approximation tends to overestimate the deceleration
used widely: that will occur. From a design point of view, this means that
4 results obtained from the added mass approximation can
ma = ka ρ πR3 . (8) be viewed as conservative. So while the concept of added
3
mass is only an approximation, the effects of fluid inertia
The value of ka , a constant, can be found in various ways [1, in transient parachute aerodynamics are significant enough
3]. Klimas calculated values of ka for porous hemispherical to be included in modeling. At present this method of
ribbon parachutes. In the present case, added mass is accounting for the inertial effects of air is the most practical
calculated by using a slight variation. In the previously tool available to be used in the transient phase of parachute
presented method, the volume of a hemisphere is used, operation.
International Journal of Aerospace Engineering 5

The parachute as previously stated is modeled as a sep- DP

arate body (rigid canopy). The deceleration of this body is
therefore calculated by extension of (7) including the riser
tension force:

dV Fr − DP − mP g sin γ
= V̇P = , (11)
dt (mP + ma )

where a is acceleration, Fr is the riser force, DP is the

parachute drag, mP is the mass of the parachute, ma is the Length mP
FR
added mass, and γ is the flight path angle. The drag force
k
provided by the parachute is calculated using the following mP g
simple expression: c
FR

1
DP = ρVP2 CD SP · η, (12) γ m
2
V
where ρ is air density, VP is parachute velocity, CD is the mg
canopy drag coefficient, SP is the canopy projected area,
and η is an efficiency factor used to account for payload Figure 2: Layout of the suspension system (parachute-payload
configuration).
wake effects (η = 0.7 ÷ 0.75 for the drogue parachute and
η = 1 for the main parachute). The product CD SP (drag
area) of a parachute depends on its type, inflated shape,
size, Reynolds’ number, Mach’s number, Froude’s number,
material elasticity, and porosity. drag area is time-scheduled according to the planned reefing
In (11) the third term on the numerator is the gravitation stages and can be described by the following expressions:
force component. The denominator is then the addition of
the mass of the parachute and the added air mass. ⎡
i−1  ni ⎤
Due to the complexity involved with the filling process,  t − ti
filling time cannot be calculated by a purely analytical (CD SP )t = (CD SP )0 · ⎣ τk + τi ⎦,
k=1
tfi
method. The alternative once again lies with an empirical 
approach that is validated by flight test. Knacke defined the for t ∈ ti ÷ t f i , (14)
following empirical relation for filling time:

i 
(CD SP )t = (CD SP )0 · τk , for t ∈ t f i ÷ ti+1 ,
8D0 k=1
tf = , (13)
VS0.9

where D0 is nominal diameter and VS is velocity at line where reefing starting time is ti , duration t f i , and ratio τk .
stretch. It should be noted here that (13) is dimensionally The power for the growth, ni , has been set at 2 for the initial
incorrect. This means that the actual physics of the system stages and at 2.5 for the final stage, for both drogue and main
are not properly represented. While (13) has been shown to parachutes (the exponents are selected to fit the available
provide reasonable answers, care should always be taken in flight test data). Selection of these values is justified by a
using such an equation so that it is not being used outside better prediction of peak loads. This value of CD SP is then
the intended range of conditions. substituted into the equation for parachute drag.
An underlying assumption used in the modeling of the The final assumption made is with the clustered
parachute canopy is that it remains aligned with the velocity parachutes. In the simulation one parachute of equivalent
vector at all times. This corresponds to neglecting the lift and size has been used, with a cluster efficiency factor applied.
moment coefficients of the parachute. In Figure 2, the forces acting on both the parachute
The second major assumption made used in the par- and payload can be seen (with the exception of the body
achute canopy model is that the added mass remains con- aerodynamic forces and moments).
stant throughout inflation and parachute operation. The forces and moments acting on the payload are
Whilst in the model the value of added mass is kept calculated as the relevant contributions are added. The
constant at the value for the fully inflated parachute (conser- equations are
vative approach for the estimation of deceleration peak), the
value for drag area is ramped up as the parachute inflates. It
1  
is this feature of the model that provides a simulation of the X = ρV S V CX + CXq qd − FR cos α cos β,
inflation process. Inflation modeling is done by ramping the 2
value of the drag area, CD SP , this being the multiplication 1
of the drag coefficient and the projected area. The effective Y = ρV S(V CY + CY r rd) − FR sin β,
2
6 International Journal of Aerospace Engineering

1  
Z = ρV S V CZ + CZq qd − FR sin α cos β, +X
2
1    
L = ρV Sd V Cl + Cl p pd − FR sin β zcg − z p d +CL
2
 
− FR sin α cos β ycg − y p d,

1    
M = ρV Sd V CM + CMq qd + FR cos α cos β zcg − z p d α
2 V
 
β
+ FR sin α cos β xcg − x p d, ⊕

1  
N = ρV Sd(V CN + CNr rd) − FR sin β xcg − x p d
2 Center
  of gravity +CM
− FR cos α cos β ycg − y p d,
+CN +Y
(15)
+Z
where X, Y , and Z are the body forces, FR is the riser force
Figure 3: The reference frame and the aerodynamic angles for the
(tension), L, M, and N are the body moments, V is velocity,
payload.
d is payload diameter, S is reference surface area, p is roll
rate, q is pitch rate, r is yaw rate, α is angle of attack, β is
angle of sideslip, and x p , y p , and z p are the parachute-payload
attachment coordinates. All coordinates are normalized with into sections, six rigid body equations of motion, four qua-
a reference length. ternion equations, three position equations, and the velocity
In the above-presented model, there are a number of and displacement of the parachute.
featured that need to be discussed. The aerodynamic data The system written in residual form can be seen below:
for the payload—which the simulation uses as input—are  
X + mgCBE (1, 3)
given in the longitudinal plane only. In order to calculate r1 = u̇ − − qw + rv ,
m
the correct coefficient, the center of gravity must be shifted  
accordingly. This is done using the following equations: Y + mgCBE (2, 3)
r2 = v̇ − − ru + pw ,
    m
CM = CM − CZ xcg − xref − CX zcg − zref ,  
    (16) Z + mgCBE (3, 3)
r3 = ẇ − + qu − pv ,
CMq = CMq − CZq xcg − xref − CXq zcg − zref , m
r4 = IXX ṗ − IXY q̇ − IXZ ṙ − IXZ pq + (IZZ − IY Y )qr
where CM is the moment coefficient, CZ is the force  
coefficient in the Z direction, CX is the force coefficient in + IXY pr + r 2 − q2 IY Z − L,
the X direction, CMq is the moment coefficient due to pitch r5 = IY Y q̇ − IXY ṗ − IY Z ṙ − IY Z pq + (IXX − IZZ )pr
rate, CZq is the force coefficient in the Z direction due to  
− IXY qr + p2 − r 2 IXZ − M,
pitch rate, CXq is the force coefficient in the X direction due
to pitch rate, xcg and zcg are the x and z locations of the r6 = IZZ ṙ − IXZ ṗ − IY Z q̇ + IXZ qr + (IY Y − IXX )pq
 
payload center of gravity, and finally xref and zref are the x − IY Z pr + q2 − p2 IXY − N,
and z locations of the reference data center of gravity. The 1  (17)
coordinates are normalized with a reference length, that is, r7 = ė0 + e1 p + e2 q + e3 r ,
2
for axis-symmetric bodies the payload diameter d. 1 
The lateral-directional coefficients are obtained assuming r8 = ė1 − e0 p − e3 q + e2 r ,
2
the geometric symmetry of the payload. In taking this 1 
approach, the aerodynamic response to angle of attack and r9 = ė2 − e3 p + e0 q − e1 r ,
2
angle of sideslip have been decoupled. The implications 1 
of this decoupling are that there are no aerodynamic roll r10 = ė3 + e2 p − e1 q − e0 r ,
2
moments (L = 0), no sideslip effects in the longitudinal
plane, and finally no angle of attack effects in the lateral r11 = ẋ − CBE (1, 1)u − CBE (1, 2)v − CBE (1, 3)w,
plane. It should also be noted that, since the body is axis- r12 = ẏ − CBE (2, 1)u − CBE (2, 2)v − CBE (2, 3)w,
symmetric, total angle of attack would need to be considered r13 = ḣ + CBE (3, 1)u + CBE (3, 2)v + CBE (3, 3)w,
for aerodynamic coefficients interpolation, as implemented
r14 = ṡ p − V p
in [23].  
The body axis system has been adopted, and this system Fr − D p − m p g sin γ
can be seen in Figure 3. The system of differential equations r15 = V̇ p −   ,
that describes the parachute-payload system can be broken m p + ma
International Journal of Aerospace Engineering 7

where u, v, and w are the body axis velocities in the X, Y , The flight path angle is calculated using a sequence of C
and Z directions, respectively, p, q, and r are the roll rate, transformation matrices. There exists a problem in extract-
pitch rate, and yaw rate in body axis, e0 through to e3 are the ing the flight path angle as it approaches 90◦ , in that there
quaternion values for orientation, γ is the flight path angle, exists a singularity in the inverse sine function. To get around
x, y, and h give the position in the Earth fixed axis, s p is the this problem, a third axis system has been introduced that is
absolute displacement of the parachute canopy, and finally rotated 90◦ in pitch with respect to the wind axis. The flight
V p is the parachute canopy total velocity. path angle is then given by the following:
The magnitude of the velocity vector, that is, the airspeed,  
is found from the three body-axis velocities: CWB = C1 (0) · C2 (α) · C3 −β ,
−1
 CBW = CWB ,
V = u2 + v2 + w2 . (18) ◦
CGW = C1 (0) · C2 (90 ) · C3 (0),
(24)
The derivative of the velocity vector is then calculated by CWE = CWB · CBE ,
CGE = CGW · GWE ,
(uu̇ + vv̇ + wẇ)
V̇ = . (19) γ = sin−1 (−CGE (1, 3)) − 90◦ .
V
The angle of attack is calculated by The coordinates of the payload along the trajectory are
 
calculated from
w 
α = tan−1  
u . (20)
s = x2 + y 2 + (h − h0 )2 , (25)
The angle of sideslip is calculated by where h0 is the initial altitude.
  The atmospheric profile (density and speed of sound)
v
β = sin−1  
V  . (21) is approximated with a cubic curve fit of the reference
atmosphere.
Note that the angles of attack and sideslip are put into A second simulation model (designated as SM2) was
the correct phase after this calculation, since the inverse defined, implemented, and validated in [23]. The major
trigonometric function will only ever return values between differences with respect to the less complex SM1 model are
±90◦ . (i) the aerodynamics of the bodies are defined in terms
The conversion from the Euler angles to quaternion of total angle of attack due to the geometrical
values is computed by the following set of equations: symmetry of payload and parachute;
     
ψ θ φ (ii) the dynamics of the parachute are modeled with a
e0 = cos · cos · cos full-state rigid body six degree-of-freedom represen-
2 2 2
      tation;
ψ θ φ
+ sin · sin · sin ,
2 2 2 (iii) the added masses of the parachute are defined
     
ψ θ φ by a tensor including rotational inertial properties
e1 = cos · cos · sin
[18, 19];
2 2 2
     
ψ θ φ (iv) the aerodynamics of the parachute include lift and
− sin · sin · cos ,
2 2 2 damping coefficients;
      (22)
ψ θ φ (v) the suspension system is represented by a more real-
e2 = cos · sin · cos
2 2 2 istic layout with distributed elements and suspension
     
ψ θ φ links whose strain is estimated with an iterative
+ sin · cos · sin ,
2 2 2 numerical method;
     
ψ θ φ
e3 = sin · cos · cos (vi) the effects of atmospheric turbulence and asymme-
2 2 2 tries (either geometrical or inertial unbalance) can be
     
ψ θ φ included.
+ cos · sin · sin .
2 2 2
The transformation matrix that relates from Earth fixed to 4. Software Implementation
body axis is calculated using quaternion values:
The simulation software is written in Fortran language.
CBE This program simulates the dynamics of the parachute-
⎡ ⎤ payload system (time domain integration). The user is
e02 + e12 − e22 − e32 2(e1 e2 + e0 e3 ) 2(e1 e3 − e0 e2 ) offered predefined initial conditions from reference flight
⎢ ⎥
= ⎣ 2(e1 e2 − e0 e3 ) e02 − e12 + e22 − e32 2(e2 e3 + e0 e1 ) ⎦. conditions or the option of trimming the payload at any
2 2 2 2
2(e0 e2 + e1 e3 ) 2(e2 e3 − e0 e1 ) e0 − e1 − e2 + e3 desired altitude and using those generated initial conditions.
(23) The inertial properties of the payload (mass, moments
8 International Journal of Aerospace Engineering

Model
Balance
V

Δθ

Amplifier
filter

A/D
1500

converter

Step
PC
motor
Servo
controller
RS232

Tunnel floor
(a) (b)

Figure 4: The experimental setup [10].

of inertia, and position of the center of gravity) and the vector with components x(1), . . . , x(n) and f is a vector of
parachute characteristics (size, mass, opening, and staging nonlinear functions. Each equation is of the form:
sequences) are defined for a selected altitude ranges (mission
table lookup). fk (x(1), . . . , x(n)) = 0, for k = 1, . . . , n. (27)
The solver used in the simulation program is DASSL (dif- The algorithm used for the solution is based on an iterative
ferential algebraic system solver) originally developed by method, which is a variation of Newton’s method using
Petzold [24–26] for the solution of systems of differen- Gaussian elimination in a manner similar to the Gauss-
tial algebraic equations (DAEs). This routine is based on Seidel process. Convergence is roughly quadratic. All partial
backward differentiation formulas (BDFs). The solution derivatives required by the algorithm are approximated by
algorithm then attempts to make the residual r(t) equal to first difference quotients. The convergence behavior of this
zero at each time step: code is affected by the ordering of the equations, and
it is advantageous to place linear and mildly nonlinear
equations first in the sequence. The convergence is started
r(t) = f (ẋ, x, t) = ẋ − f  (x, t). (26) from statically stable conditions (initial payload tumbling
is avoided). Numerical integration was performed assuming
a time step of 0.001 s with an absolute and relative error
Equation (26) also shows that it is very simple to modify tolerance of 10−7 . Longitudinal and directional planes are
existing explicit ODE (ordinary differential equation) sys- considered separately. The system of equations for the
tems to be used with an implicit formulation. With DAE longitudinal plane is defined as follows:
solvers [27], the equations of motion can be implemented f1 = M, f2 = X − mg sin θ, f3 = Z + mg cos θ,
directly in the form of residuals. Therefore, no symbolic (28)
expansions are needed to identify acceleration terms, and f4 = θ − γ − α,
no ad-hoc algorithms need to be used to determine the where
vector of derivatives at every time step. The DAE solver may 1  
automatically adjust the order of the integration formula M = ρV 2 SdCM + FR cos α zcg − z p d
2  
and the integration step size to achieve the desired accuracy,
whereas typical solvers have fixed order, fixed step size, and + FR sin α xcg − x p d,
no automatic accuracy control. 1 (29)
The initial conditions for the payload (with or without X = ρV 2 SCX − FR cos α,
2
decelerator deployed) may be computed for a given initial 1
altitude h0 , returning the quasi-trim values for V velocity, α Z = ρV 2 SCZ − FR · sin α.
2
angle of attack, β angle of sideslip, θ pitch angle, and γ flight
path angle (stabilized fall). Note that it has been decided 5. Results
to leave roll as the untrimmed equation. The purpose is to
solve a system of n simultaneous nonlinear equations in n The reference mission described in [23] was used to evaluate
unknowns. It solves the problem f (x) = 0 where x is a the two simulation models. The data available from the
International Journal of Aerospace Engineering 9

1 4
200
0 3
Drogue snatch

−1
150 2
Drogue disreef
−2
Airspeed (m/s)

Deceleration (g’s)
1
−3

SM1

SM2
100 Main snatch
0
−4

−1
Main disreef −5
50
−2
−6

−7 −3
0
6000 8000 10000 12000 14000
h (m) 6000 8000 10000 12000 14000
h (m)
SM1
SM2 SM1
Flight test SM2
(a) (b)
−55 −55

−60 −60

−65 −65
Flight-path angle (deg)
Flight-path angle (deg)

−70 −70

−75 −75

−80 −80

−85 −85

−90 −90

−95 −95
6000 8000 10000 12000 14000 6000 8000 10000 12000 14000
h (m) h (m)

SM1 SM2
SM2 Flight test
(c) (d)

Figure 5: The simulation of the trajectory profile for the parachute-payload system.

reference flight tests [28, 29] are used for the validation and a cluster of three main polyconical slotted parachutes
of the simulation of capsule terminal reentry dynamics. (nominal diameter 22.9 m) inflated in three reefing stages.
The capsule is substantially a 70%-scaled version (diam- The sequence of inflation (see Table 1) and the parachute
eter is 2.8 m) of the original Apollo Command Module drag profile are time-scheduled according to the planned
decelerated by a single conical ribbon drogue parachute reefing stages and inflation times. Note that staging is
(nominal diameter 5.8 m) inflated in two reefing stages obtained through parachute reefing (i.e., through restricted
10 International Journal of Aerospace Engineering

50000 50000

40000 40000

30000 30000
Load (N)

Load (N)
20000 20000

10000 10000

SM1 SM2
0 0
7500 10000 12500 15000 7500 10000 12500 15000
h (m) h (m)

Riser Riser
Parachute Parachute
(a) (b)
200000 200000

150000 150000
Load (N)

Load (N)

100000 100000

50000 50000

SM1 SM2
0 0
5000 5500 6000 6500 7000 5000 5500 6000 6500 7000
h (m) h (m)

Riser Riser
Parachute Parachute
(c) (d)

Figure 6: The loads acting on the riser and the decelerator (drogue and main parachute).

canopy deployment) with the purpose of limiting peak loads Torino [10]. The wind tunnel is a closed-loop tunnel with a
and deployment shocks. A detailed description of parachute cylindrical working section of 3 m diameter by 5 m length.
deployment sequences for aerospace applications is given in The model reproduces the scaled geometry of the NASA
[2, 3]. Apollo Command Module (model diameter is 340 mm). The
The aerodynamic data for the payload—which the sim- longitudinal loads on the scaled vehicle (axial force, normal
ulation uses as input—were obtained with wind tunnel force, and pitching moment) are measured by an internal
static and forced oscillation tests performed at Politecnico di 3-component balance fit within the capsule model. Body
International Journal of Aerospace Engineering 11

14000 200

12000 100
h (m)

y (m)
10000 0

8000 −100

6000 −200

−2000 −1000 0 1000 2000 −2000 −1000 0 1000 2000

s (m) s (m)

SM1 SM1
SM2 SM2
(a) (b)

Figure 7: The simulation of the trajectory profile for the parachute-payload system.

Table 1: The parachute deployment sequence. longitudinal stability derivatives (measured from frequency
response) is σ ≈ ±1%.
Sequence Time (s) Altitude (m) The reentry profile is correctly reproduced by the two
Mortar firing (Pilot Jettison) 0.00 13994 simulators—as presented in Figure 5—and the trajectory,
Back cover separation — 13666 shaped by the parachute opening sequence, is matched by
Drogue snatch/1st stage inflation 3.34 13430 both SM1 and SM2 models. The trend of the load factor
Drogue 2nd stage exhibits a set of marked peaks given by the staging of both
9.34 12546 main and drogue parachute. The SM1 results exhibit an
inflation/disreef
Drogue bridles cut — 6461 oscillatory behavior in the second part of the parachute
Main snatch/1st stage inflation 82.30 6364
deployment phase, induced by the oscillatory behavior of the
payload, not found for SM2 model results. Note that low-
Main 2nd stage inflation 88.30 6092
altitude flight test data are affected by the asymmetry of the
Main 3rd stage inflation/disreef 94.55 5950 suspension system induced by the cut of one of the bridles of
Main bridle 1 cut — 5360 the main parachutes (see Table 1).
The ability of the SM1 model to estimate the riser and
parachute loading is outlined in Figure 6. The trend of
axes are adopted for the reduction of experimental results forces is coherent with the deceleration profile presented in
centered in the actual reference center of gravity. The angle Figure 5. The offset between the force acting on the canopy
of attack is set to 180◦ when the spherical base (thermal and the riser is well marked for the main clustered parachute,
shield) is exposed to the wind. The model is supported modeled as an equivalent single decelerator. This offset is
(see Figure 4) by a vertical strut that can rotate over the due to the mutual influence of the added masses (inertia-
complete range of angle of attack (α = ±180◦ ). Positioning induced delays) and the elongation of the riser. The effect
is performed by a step-motor that is controlled by a driving of the elasticity of the main cable is also visible in terms of
unit interfaced to the data acquisition PC. The stability damped oscillatory strain, triggered by the bouncing of the
derivatives are evaluated according to the small amplitude payload (and by its attitude dynamics) after each inflation
direct forced oscillation technique [30]. The oscillation is phase (staged reefing).
generated by the step-motor (Δθ = 1◦ − 5◦ ) and f = 0.5– The projected trajectory is compared in Figure 7. Accu-
5 Hz. The position of the driving shaft is acquired by means rate trajectory coordinates were not available from flight test
of a digital encoder. The measurement repeatability for the data. The profiles (mirrored for the purpose of checking
averaged static coefficients is σ ≈ ±0.1% estimated over their symmetry) are very similar in the vertical plane.
the full range while the measurement repeatability for the Nevertheless, a slightly different path is found comparing the
12 International Journal of Aerospace Engineering

230 230
190 190

Flight test
SM2
220 220

SM2
SM1
180 180
210 210
170 170

Angle of attack (deg)

Angle of attack (deg)

200 200
160 160
190 190
150 150
180 180
140 140
170 170
130 130
160 160
120 120
6000 8000 10000 12000 14000 6000 8000 10000 12000 14000
h (m) h (m)

SM1 Flight test

SM2 SM2
(a) (b)
10 30
10 30

5 25
Flight test

5 25
SM1

0 20
0 20
Angle of sideslip (deg)

−5
Angle of sideslip (deg)

15
−5 15

SM2
SM2

−10 10
−10 10

−15 5
−15 5

−20 −20 0
0

−25 −5 −25 −5

−30 −10 −30 −10

6000 8000 10000 12000 14000 6000 8000 10000 12000 14000
h (m) h (m)

SM1 Flight test

SM2 SM2
(c) (d)

Figure 8: The simulation of the angles of attack and sideslip for the payload.

traces for the lateral-directional plane. This is the impact of The aerodynamic angles of the payload are plotted in
the aerodynamic model adopted for the payload, which for Figure 8. The major discrepancy between the two models
the SM1 case decouples and superimposes the effects for the is the level of dynamic stability of the modal response
longitudinal and the lateral-directional planes. The use of (short-period dynamics). This can be explained with the
total angle of attack for the interpolation and reconstruction extreme sensitivity of the aerodynamic coefficients to center
of payload aerodynamic coefficients (as in SM2 model) of gravity location and angle of attack (see Figures 9 and
provides a more accurate fit. 10). Remind that the center of gravity location is updated
International Journal of Aerospace Engineering 13

Table 2: The index of similarity for the pitch rate spectral response.

Flight test ⇒ SM1 Flight test ⇒ SM2

Drogue parachute Drogue parachute
I1 I2 I3 I1 I2 I3
1.392 1.005 0.989 0.645 1.145 1.110
Main parachute Main parachute
I1 I2 I3 I1 I2 I3
0.578 1.022 1.029 0.463 1.088 1.098

0.15 providing a more realistic behavior of the parachute-payload

xCG = 0.657 − zCG = 0 system. Differently, the parachute trajectory in the SM1
model is aligned instantly with the velocity vector (flight
path) of the payload. Furthermore, a multilink suspension
system (as modeled in SM2 implementation) induces a
stronger steering effect on the payload.
CM

0 xCG = 0.657 − zCG = 0.035 In order to verify the modal response, at least for
the longitudinal short-period natural frequency, the pitch
Stable trim rate spectral response for the two simulation models is
compared with the elaboration of available flight logs in
Figures 11 and 12 (note that from flight test reports gyro
−0.15
90 105 120 135 150 165 180
signal measurements are very noisy after drogue release). The
comparison of the spectral data is based on the crosscheck
Angle of attack (deg)
of the index of similarity (see Table 2), defined according to
Figure 9: The aerodynamic coefficients of the payload (pitching (30), where Ai is the amplitude for the given frequency ωi :
moment static stability) [10].

N
s0 = Ai ,
i=1
xCG = 0.657 − zCG = 0 s0 σ1a
σ1 = I1 =
N σ1b

N s1 σ2a
Unstable s1 = Ai · ωi −→ σ2 = s −→ I2 =
σ2b
CMq

0 (30)
∗

0 i=1

s2 σ3a
σ3 = I3 =
s0 σ3b
xCG = 0.657 − zCG = 0.035   
a→b

N
90 105 120 135 150 165 180 s2 = Ai · ωi2 .
Angle of attack (deg) i=1

Figure 10: The aerodynamic coefficients of the payload (pitching The results show that the range for natural frequency of
moment dynamic stability) [10]. the short period response is matched by both simulation
models as demonstrated by the fact that I2 and I3 are
close to unity. Other than that, the averaged amplitude of
during the descent, accounting for covers, deployment bags, the spectral response is less precisely reproduced by the
and parachutes release. As a matter of fact, minor parametric simulation models, as experimental data represent a modal
changes (α, xcg , and zcg ) shift the attitude for the rotational response that is overexcited by atmospheric disturbances
equilibrium of the payload and alter its dynamic stability. and suspension system asymmetries, mainly at parachute
This point is also a concern for the validation of simulation deployment.
models by comparison with flight test data [23], in which a
transition between different equilibrium states and levels of 6. Conclusions
dynamic stability can be externally forced by the atmospheric
perturbations, exciting the combined parachute-payload The present work outlines a comparative analysis of two
system. Another contributing element is the dynamics of the simulation models (SM1 and SM2) of a parachute-payload
parachute that in the higher fidelity model (SM2) allows system with different levels of complexity. An example of
for attitude changes decoupled from the suspended body, limited complexity reconfigurable simulation models for a
14 International Journal of Aerospace Engineering

30 30

25 25

20 20

Amplitude (deg/s)
Amplitude (deg/s)

15 15

10 10

5 5

0 0
1 10 1 10
Frequency (rad/s) Frequency (rad/s)

Flight test Flight test

SM1 SM2
(a) (b)

Figure 11: The effect of model complexity on pitch rate spectral response (drogue parachute).

30 30

25 25

20 20
Amplitude (deg/s)

Amplitude (deg/s)

15 15

10 10

5 5

0 0
1 10
1 10
Frequency (rad/s)
Frequency (rad/s)
Flight test Flight test
SM1 SM2
(a) (b)

Figure 12: The effect of model complexity on pitch rate spectral response (main parachute).
International Journal of Aerospace Engineering 15

payload decelerated by one or more parachutes is given, Abbreviations

including details and implementation features usually omit-
ted as the focus of the research in this field is typically on the CG: Center of gravity
investigation of mission design issues, rather than addressing SM1: Simulation model 1
general implementation guidelines for the development of a SM2: Simulation model 2.
reconfigurable simulation tool. The dynamics of the system
are modeled through a simple multibody model that repre-
Symbols
sents the expected behavior of an entry vehicle during the
final deceleration phase by means of a system of aerodynamic
decelerators. The simulators are designed according to a c: Damping coefficient
comprehensive vision that enforces the simplification of the CBE : Body-to-Earth axis transformation
coupling mechanism between the payload and the parachute, matrix
with an adequate level of physical insight still available in CBW : Body-to-wind axis transformation
terms of sensitivity of system performance metrics to design matrix
parametric changes. CD : Parachute drag coefficient
A reference mission described was used to evaluate CGW : Auxiliary transformation matrix
the two simulation models. The aerodynamic data for the Cl : Roll moment coefficient
payload—which the simulation uses as input—were ob- Cl p : Roll moment coefficient due to roll rate
tained with wind tunnel static and forced oscillation tests CM : Pitching moment coefficient
performed at the Politecnico di Torino. CMq : Pitching moment coefficient due to
The trajectory is reproduced by the two simulators as the pitch rate
profiles, mainly shaped by the parachute opening sequence, CN : Yaw moment coefficient
are matched by both SM1 and SM2 models, even if the use of CNr : Yaw moment coefficient due to yaw rate
total angle of attack for the interpolation and reconstruction CWB : Wind-to-Body axis transformation
of payload aerodynamic coefficients (as in SM2 model) matrix
provides a more accurate fit. The riser and parachute loading CX : Force coefficient in x-axis direction
are estimated in a way that is coherent with the deceleration CXq : Force coefficient in x-axis direction due
profiles. to pitch rate
The major discrepancy between the two models is the CY : Force coefficient in y-axis direction
level of dynamic stability of the modal response (short- CY r : Force coefficient in y-axis direction due
period dynamics). This can be explained with the extreme to yaw rate
sensitivity of the aerodynamic coefficients to center of gravity CZ : Force coefficient in z-axis direction
location and angle of attack. Another contributing element is CZq : Force coefficient in z-axis direction due
the dynamics of the parachute which in the higher fidelity to pitch rate
model (SM2) allows for attitude changes decoupled from d: Payload reference length
the suspended body, providing a more realistic behavior D0 : Canopy nominal diameter
of the parachute-payload system. Differently, the parachute DP : Parachute drag
trajectory in the SM1 model is aligned instantly with e0 · · · e3 : Quaternion parameters
the velocity vector of the payload, probably a too crude FR : Riser force (tension)
approximation if the purpose of the analysis is the attitude g: Acceleration of gravity
dynamics of the suspended vehicle. Furthermore, a multilink h: Canopy height
suspension system (as modeled in SM2 implementation) h: Altitude
induces a steering effect on the payload, neglected by the li : Index of similarity
single-point suspension of SM1 model. IXX : Moment of inertia
The spectral results show that the range for the natural IXY : Product of inertia
frequency of the short-period response—as measured in IXZ : Product of inertia
flight—is matched by both simulation models as demon- IY Y : Moment of inertia
strated by the fact that the indices of similarity I2 and I3 IY Z : Product of inertia
are close to unity. Differently, the averaged amplitude of IZZ : Moment of inertia
the spectral response is less precisely reproduced by the k: Stiffness
simulation models, as the experimental data represent a ka : Added mass coefficient
modal response that is over-excited by atmospheric distur- kb : Bridle stiffness
bances and suspension system asymmetries, at parachute kr : Riser stiffness
deployment mainly. L: Rolling moment
As a general conclusion, the lower fidelity simulation tool l0 : Offset length
SM1 exhibits a limited advantage in terms of computational lr : Riser length
workload, mainly providing a simplified approach for pre- m: Payload mass
liminary trajectory estimation and parachute sizing. M: Pitching moment
M∗: Parachute total mass
16 International Journal of Aerospace Engineering

ma : Parachute added mass [3] R. C. Maydew and C. W. Peterson, “Design and testing of high
mP : Parachute mass performance parachutes,” AGARD AG-319, 1991.
N: Yawing moment [4] NASA Langley Research Center, “Deployable aerodynamic
p: Canopy porosity deceleration systems,” NASA SP-8066, 1971.
p: Roll rate [5] M. L. Accorsi, J. W. Leonard, R. J. Benney, and K. R. Stein,
q: Pitch rate “Structural modeling of parachute dynamics,” AIAA Journal,
r: Yaw rate vol. 38, no. 1, pp. 139–146, 2000.
R: Canopy radius [6] M. L. Accorsi, R. J. Benney, V. Kalro, J. W. Leonard, K. R. Stein,
and T. E. Tezduyar, “Parachute fluid-structure interactions: 3-
rk : Numerical residual
D computation,” Computer Methods in Applied Mechanics and
s: Payload displacement Engineering, vol. 190, no. 3-4, pp. 373–386, 2000.
sP : Parachute displacement [7] M. L. Accorsi, R. J. Benney, J. W. Leonard, K. R. Stein, and T. E.
SP : Parachute reference area Tezduyar, “Fluid-structure interactions of a round parachute:
t: Time modelling and simulation techniques,” Journal of Aircraft, vol.
u: Velocity in x body direction 38, no. 5, pp. 800–808, 2001.
v: Velocity in y body direction [8] J. R. Cruz and J. S. Lingard, “Aerodynamic decelerators for
V: Payload-free stream velocity planetary exploration: past, present, and future,” in Proceed-
Vc : Payload velocity ings of the AIAA Guidance, Navigation, and Control Conference,
VP : Parachute velocity pp. 5342–5361, August 2006.
Vs : Parachute velocity (at line stretch) [9] M. Ballion, “Blunt bodies dynamic derivatives,” AGARD R808,
w: Velocity in z body direction 1995.
x: North displacement (trajectory) [10] G. Guglieri and F. Quagliotti, “Low speed dynamic tests on a
capsule configuration,” Aerospace Science and Technology, vol.
X: Force in x body direction
4, no. 6, pp. 383–390, 2000.
xcg : x-wise CG location
[11] L. E. Ericsson and J. P. Reding, “Re-entry capsule dynamics,”
xp: x-wise suspension point location Journal of Spacecraft and Rockets, vol. 8, no. 6, pp. 579–586,
xref : x-wise CG location (reference) 1971.
y: East displacement (trajectory) [12] R. G. Hume, A Two Dimensional Mathematical Model of a
Y: Force in y body direction Parachute in Steady Descent, Aeronautical Research Council,
ycg : y-wise CG location 1973, C.P. No. 1260.
yp: y-wise suspension point location [13] M. Neustadt, R. E. Ericksen, J. J. Guiteras, and J. A. Larrivee,
z: Down displacement (trajectory) “A parachute recovery system dynamic analysis,” Journal of
Z: Force in z body direction Spacecraft and Rockets, vol. 4, no. 3, pp. 321–326, 1967.
zcg : z-wise CG location [14] F. M. White and D. F. Wolf, “A theory of three-dimensional
zp: z-wise suspension point location parachute dynamic stability,” Journal of Aircraft, vol. 5, no. 1,
zref : z-wise CG location (reference) pp. 86–92, 1968.
α: Angle of attack [15] D. Wolf, “Dynamic stability of a nonrigid parachute and
β: Angle of sideslip payload system,” Journal of Aircraft, vol. 8, no. 8, pp. 603–609,
1971.
γ: Flight path angle
[16] K. F. Doherr and H. Schilling, “Nine degree-of-freedom
ε: Strain simulation of rotating parachute systems,” Journal of Aircraft,
ε̇: Strain rate vol. 29, no. 5, pp. 774–781, 1992.
η: Wake penalty factor [17] C. Tory and R. Ayres, “Computer model of a fully deployed
θ: Pitch angle parachute,” Journal of Aircraft, vol. 14, no. 7, pp. 675–679,
ξ: Damping ratio 1977.
ρ: Air density [18] J. A. Eaton, “Added masses and dynamic stability of par-
σ: Measurement error achutes,” Journal of Aircraft, vol. 19, no. 5, pp. 414–416, 1982.
φ: Roll angle [19] J. A. Eaton, “Added fluid mass and the equations of motion of
ψ: Yaw angle a parachute,” Aeronautical Quarterly, vol. 34, no. 3, pp. 226–
ωi : Frequency. 242, 1983.
[20] P. A. Cuthbert, “A software simulation of cargo drop test,” in
Proceedings of the 17th AIAA Aerodynamic Decelerator Systems
Acknowledgment Technology Conference, Monterey, Calif, USA, 2003, AIAA
Paper 2003-2132.
The author wishes to acknowledge the support given by Mr. [21] P. A. Cuthbert and K. J. Desabrais, “Validation of a cargo air-
Michael Gordon. drop simulator,” in Proceedings of the 17th AIAA Aerodynamic
Decelerator Systems Technology Conference, Monterey, Calif,
References USA, 2003, AIAA Paper 2003-2133.
[22] P. C. Klimas, “Fluid mass associated with an axisymmetric
[1] D. J. Cockrell and A. D. Young, “The aerodynamics of canopy,” Journal of Aircraft, vol. 14, no. 6, pp. 577–580, 1977.
parachutes,” AGARD AG-295, 1987. [23] G. Guglieri and F. Quagliotti, “Validation of a simulation
[2] T. W. Knacke, Parachute Recovery Systems Design Manual, Para model for a planetary entry capsule,” Journal of Aircraft, vol.
Publishing, 1st edition, 1992. 40, no. 1, pp. 127–136, 2003.
International Journal of Aerospace Engineering 17

[24] U. M. Ascher and L. R. Petzold, Computer Methods for

Ordinary Differential Equations and Differential Algebraic
Equations, SIAM, Philadelphia, Pa, USA, 1999.
[25] K. E. Brenan, S. L. Campbell, and L. R. Petzold, Numerical
Solution of Initial Value Problems in Differential Algebraic
Equations, SIAM, Philadelphia, Pa, USA, 1996.
[26] L. R. Petzold, “Description of DASSL: a differential alge-
braic system solver,” in Proceedings of the 10th International
Mathematics and Computers Simulation Congress on Systems
Simulation and Scientific Computation, Montreal, Canada,
1982.
[27] R. Celi, “Solution of rotary-wing aeromechanical problems
using differential-algebraic equation solvers,” Journal of the
American Helicopter Society, vol. 45, no. 4, pp. 253–263, 2000.
[28] S. Portigliotti, J. C. Paulat, and Rives J., “ARD descent and
recovery subsystem dynamics and trajectory flight data eval-
uation,” in Proceedings of the 1st AAAF International Sympo-
sium on Atmospheric Reentry Vehicles and Systems, Arcachon,
France, 1999.
[29] S. Portigliotti, “Parachute flight dynamics modeling and
simulation: code validation through ARD post-flight analysis,”
in Proceedings of the 2nd AAAF International Symposium on
Atmospheric Reentry Vehicles and Systems, Arcachon, France,
2001.
[30] E. S. Hanff, “Direct forced oscillation techniques for the de-
termination of stability derivatives in wind tunnel,” AGARD
LS-114, 1981.
 International Journal of

Rotating
Machinery

International Journal of
The Scientific
Engineering Distributed
Journal of
Journal of

Hindawi Publishing Corporation

World Journal
Hindawi Publishing Corporation Hindawi Publishing Corporation
Sensors
Hindawi Publishing Corporation
Sensor Networks
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014

Journal of

Control Science
and Engineering

Advances in
Civil Engineering
Hindawi Publishing Corporation Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014

Submit your manuscripts at

http://www.hindawi.com

Journal of
Journal of Electrical and Computer
Robotics
Hindawi Publishing Corporation
Engineering
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014

VLSI Design
Advances in
OptoElectronics
International Journal of

International Journal of
Modelling &
Simulation
Aerospace
Hindawi Publishing Corporation Volume 2014
Navigation and
Observation
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014
in Engineering
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014
Engineering
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2010
Hindawi Publishing Corporation
http://www.hindawi.com
http://www.hindawi.com Volume 2014

International Journal of
International Journal of Antennas and Active and Passive Advances in
Chemical Engineering Propagation Electronic Components Shock and Vibration Acoustics and Vibration
Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014