AME02

The Second TSME International Conference on Mechanical Engineering

19-21 October, 2011, Krabi

A Comparative Study on 6-DOF Trajectory Simulation of a Short Range Rocket

using Aerodynamic Coefficients from Experiments and Missile DATCOM

Pawat Chusilp*, Weerawut Charubhun, and Navapan Nutkumhang

Defence Technology Institute (Public Organization)

47/433 Chang Wattana Road, Pakkred, Nonthaburi, Thailand 11120
*Corresponding Author: [email protected], Tel: 66 2 980 6612 to 15, Fax: 66 2 980 6688 ext 300

Abstract
A comparative study was carried out to investigate the trajectory simulation of a short range solid
propellant rocket using aerodynamic coefficients determined by different methods. The first set of
aerodynamic coefficients was estimated using an aerodynamic prediction code, Missile DATCOM. It was
found that the accuracy of the predicted coefficients was limited due to the limitation of Missile DATCOM
and model simplification. The second coefficient set was obtained from published experimental data and
employed as a benchmark. Then these two sets of coefficients were applied to a 6-DOF rigid body model
for trajectory simulation. The result parameters, such as spin rate, angle of attack, and impact point, were
compared. The comparison suggested that the less accurate coefficients predicted by Missile DATCOM
could be used for predicting velocity and impact point of the selected rocket with moderate errors.
However, significant error was found in the spin rate and angle of attack prediction.
Keywords: Trajectory Simulation, Aerodynamic Coefficients, 6-DOF, Rocket, Missile DATCOM.

List of Symbols Cmq Pitching moment coefficient derivative

Aref Reference area, which is equal to with pitch rate (1/rad)
rocket cross section area (m2) Cnβ Yawing moment coefficient derivative
CA Axial force coefficient with side slip angle (1/rad)
CD Drag coefficient Cnr Yawing moment coefficient derivative
Cl Rolling moment coefficient with yaw rate (1/rad)
Clp Rolling moment coefficient derivative Cnp Yawing moment coefficient derivative
with roll rate (1/rad) with pitch rate (1/rad)
Cmα Pitching moment coefficient derivative CNα Normal force coefficient derivative with
with angle of attack (1/rad) angle of attack (1/rad)
Cmά Pitching moment coefficient derivative CYβ Side force coefficient derivative with side
with angle of attack rate (1/rad) slip angle (1/rad)
Fx,Fy,Fz Aerodynamic forces (N)
 AME02
The Second TSME International Conference on Mechanical Engineering
19-21 October, 2011, Krabi

Mx,My,Mz Aerodynamic moments (N.m) ballistic range, or flight tests. Secondly, the CFD
Fprop Propulsive force (N) methods simulate flow fields and aerodynamic
Mprop Propulsive moment (N.m) forces by solving a set of fundamental
g Acceleration due to Earth’s gravity (m/s2) equations, i.e. Navier-Stokes equations, for fluid
Ixx,Iyy,Izz Rocket moments of inertia (kg.m2) domains divided into discrete cells. Some
Lref Characteristic length, which is equal to advanced CFD techniques also incorporate Rigid
rocket caliber (m) Body Dynamics (RBD) into the simulation [2-6].
m Rocket mass (kg) Finally, the semi-empirical and analytical
p,q,r Rotation components in the body frame methods predict aerodynamic coefficients using
(rad/s) analytical formula and empirical database.
u,v,w Velocity components in the body frame Among these groups, the semi-empirical
(m/s) and analytical methods are arguably the quickest
V Total velocity magnitude (m/s) way to determine aerodynamic coefficients. This
Xcp Center of pressure location measured advantage makes them suitable for the
from nose (m) preliminary design phase, where the rocket
Xcg Center of gravity location measured performance needs to be evaluated quickly.
from nose (m) Aerodynamics softwares that employ semi-
α Angle of attack (rad) empirical and analytical methods are Missile
β Side slip angle (rad) DATCOM, Aeroprediction, PANEL3D, PRODAS,
ρ Atmospheric air density (kg/m3) etc. These softwares have been evaluated for
ø,θ,ψ Rocket attitudes in Earth frame (rad) various geometry configurations. Although some
literatures showed good agreement with
1. Introduction experimental data, it was suggested that these
A six degree of freedom (6-DOF) model softwares cannot give accurate results for every
enables engineers to investigate rocket projectile configuration [7-13]. So caution must
dynamics during the preliminary design phase be taken when using coefficients from these
[1]. The aerodynamic forces and moments softwares in trajectory simulation.
included in the six degree of freedom model are This paper investigates the accuracy of
normally calculated using aerodynamic 6-DOF trajectory simulation using aerodynamic
coefficients. These coefficients can be estimated coefficients that were predicted by Missile
by many methods, which may be categorized DATCOM. The rocket chosen for this study was
into 3 groups [2]: experimental methods, Hydra70, which is a short range solid propellant
computational fluid dynamics (CFD) methods, rocket. Aerodynamics and flight test data of
and semi-empirical and analytical methods. Hydra70 in a published report [14] was used as
Experimental methods obtain a benchmark. These coefficients were applied to
aerodynamic coefficients from wind tunnel, a 6-DOF trajectory model to simulate spin rate,
 AME02
The Second TSME International Conference on Mechanical Engineering
19-21 October, 2011, Krabi

angle of attack, and impact points, etc. Results the body frame to the Earth frame can be done
from simulation using experimental coefficients using a rotation matrix.
were compared to the same simulation using The 3 translations and 3 rotations are
coefficients from Missile DATCOM. calculated by Eqs. (1) – (6).

2. Six Degree of Freedom Model u& =

1
(Fpropulsion + Faxial ) − g sin θ + rv − qw (1)
m
Similar to most literatures, the six 1
v& = Fside + g cosθ sin φ + pw − ru (2)
degrees of freedom in this study comprises 3 m
translational components and 3 rotational 1
w& = Fnormal + g cosθ cosφ + qu − pv (3)
m
components. Following assumptions are made:
1) Rocket structure is rigid. p& =
1
{M propulsion + M x + (I yy − I zz )qr} (4)
I xx
2) Thrust vector is coincident with the q& =
1
{M y + (I zz − I xx ) pr} (5)
rocket longitudinal axis. I yy

3) Product moments of inertia of the rocket r& =

1
{M z + (I xx − I yy )pq} (6)
I zz
are insignificant.
4) Effects from Coriolis, Magnus, and wind
The aerodynamic forces and moments
are neglected.
are determined by Eqs. (7) – (12). Note that the
Two Cartesian coordinate systems are
nomenclatures in the equations below are case
employed: the body frame and the Earth frame.
sensitive.
In the body frame (XYZ), the origin is attached
to the rocket center of gravity but the axes do 1
not roll with the rocket body. All axes are Faxial = ρV 2 Aref (− C A ) (7)
2
Fside = ρV 2 Aref (− CYβ β )
orthogonal to each other. The X axis is 1
(8)
2
coincident with the rocket longitudinal axis and X 1
Fnormal = ρV 2 Aref (− C Nα α ) (9)
is positive forward. The Y axis is directed out the 2
starboard of the rocket. The Z axis points 1 ⎛
M x = ρV 2 Lref ⎜⎜ Cl +
Clp pLref ⎞
⎟⎟ (10)
downward and normal to the X and Y axis, 2 ⎝ 2V ⎠
⎛ qLref ⎞
ρV 2 Lref ⎜⎜ Cmα α + (Cmα& + Cmq )
satisfying the right-hand rule. 1 (11)
My = ⎟
2 ⎝ 2V ⎟⎠
In the Earth frame (NED), the origin is
⎛ ⎞
ρV 2 Lref ⎜⎜ Cnβ β + (Cnr r + Cnp p ) ref
1 L
(12)
attached to the launch site. The N and E axis Mz = ⎟⎟
2 ⎝ 2V ⎠
pointed to the geographic north and east
respectively. The D axis pointed downward, 3. Aerodynamic Coefficients
following the right-hand rule. The orientation of 3.1. Hydra70 Rocket
body frame relative to Earth frame is defined by
The Hydra70 rocket is a short range
3 Euler angles (ø, θ, ψ). Transformation from unguided solid propellant rocket that is
aerodynamically stabilized. Besides air-to-
 AME02
The Second TSME International Conference on Mechanical Engineering
19-21 October, 2011, Krabi

surface applications, Hydra70 is also used for 3.2 Aerodynamic Data from Experiments
surface-to-surface applications in some trainings This study employs the aerodynamics in
and experimental works due to its low cost. The the unclassified report published by Dahlke and
configuration selected for this study is the MK66 Batiuk at US Army Missile Command [14] as a
Mod1 rocket motor mated to M261 warhead, as benchmark. This literature provides experimental
shown in Fig. 1. It is important to note that M261 data of CD, CNα , Cmq , Cl , Clp , and Xcp for
warhead is chosen due to the availability of test Hydra70 MK66/M151 and MK66/M261
data [14]. It does not imply that the Royal Thai configurations. The literature also presents the
Armed Forces employs such a submunition spin rate, Mach number obtained from a flight
warhead. test. It was described that CD for power-off
The rocket is 70 mm in caliber and 1.7 period was obtained almost entirely from the
m long. The rocket motor produces a total actual flight test and CD for power-on period was
impulse of 1500 lb-sec approximately, which determined by adjusting the base pressure drag.
enables the maximum speed up to Mach 1.8 CNα ,Cmq , Cl , and Clp were derived based on
and maximum range more than 10 km. Fig. 2 wind tunnel test data. All coefficients were
shows the nozzle with unfolding wrap-around presented as a function of Mach number only. In
fins. All fins are beveled at the leading edge and addition, CA and Cmα are not presented in the
partially at the trailing edge to produce desirable Dahlke and Batiuk [14]. So they are derived from
aerodynamic rolling moment characteristic. It is Eqs. (13) and (14).
difficult to model these features by semi-
empirical or analytical methods. In addition, the CA =
(C D − C Nα α sin α ) (13)
nozzle outlet is fluted to providing torque during cosα
C Nα (X cg − X cp )
the power-on period. Cmα = (14)
Lref

Furthermore, by assuming 90º rotational

symmetry, we have CYβ = CNα , Cnβ = -Cmα , and
Fig. 1 Hydra70 rocket [15] Cnr = Cmq. Although the Hydra70 is actually 120º
rotational symmetry, this approximation should
be reasonable for a spinning rocket.

3.3. Aerodynamic Data from Missile DATCOM

Missile DATCOM is a semi-empirical
code for preliminary missile aerodynamic design.
Fig. 2 Wrap-around fins of MK66 rocket motor It can predict aerodynamics coefficients of both
finned and non-finned projectiles based on
empirical data and analytical formula. The
 AME02
The Second TSME International Conference on Mechanical Engineering
19-21 October, 2011, Krabi

original FORTRAN 77 version of Missile static coefficients. Predicted values of dynamic

DATCOM was developed by the McDonnell coefficients might have errors up to 30% [7,8].
Douglas Corporation [16]. The 1997 version was Despite the fact that the wrap-around
used in this study. fins of MK66 Mod1 rocket motor are curved, the
The curvature of wrap-around fins is rocket is simply a Body-Tail configuration. So
neglected and the fins are treated as a planar fin Missile DATCOM should be capable of
for simplification. The diameter step at the fin predicting such aerodynamic characteristics
hinge is not modeled. The beveled leading edge accurately. However, it was found that there are
and partially beveled trailing edge of the fin, as large errors in the predicted roll moment
shown in Fig. 2, are also neglected. These coefficients because some geometric features
simplifications could results in prediction errors that affect aerodynamic characteristics could not
but the purpose of this study is to investigate the be input into Missile DATCOM as discussed
effects of errors in coefficients on trajectory previously.
simulation, not to obtain accurate aerodynamic Overall, both power-on and power-off CA
coefficients from Missile DATCOM. are under predicted in supersonic region but
Appendix A shows the aerodynamic over predicted in subsonic region. Fig. 3
coefficients predicted by Missile DATCOM compares CA predicted by Missile DATCOM to
throughout the range of Mach 0.4 to 2.5 and CA derived by Eq. 13 at zero angle of attack. It
angle of attack -12º to 12º. could be seen that the error of power-on CA is
3.4. Prediction Accuracy higher than the error of power-off CA in most
The ability of Missile DATCOM in range of Mach number. Furthermore, the
predicting aerodynamics for various rocket magnitudes of CNα, Cmq, Cl, and Clp are largely
configurations has been studied in many under predicted.
research works [7-9,12,13]. Some studies were
done at the angle of attack up to 90º and speed
ranging from subsonic to supersonic. It was
suggested that Missile DATCOM was reasonably
accurate for simple missile configurations such
as Body alone and Body-Tail with low sweep fin
angle. For Body-Wing-Tail configurations, errors
of predicted CA , CNα , Cmα were less than 4% at
angle of attack 0º to 20º and 22% at angle of Fig. 3 CA at zero angle of attack
attack 20º to 45º [7,8]. Both over prediction and
under prediction were observed. Furthermore,
prediction of dynamic derivative coefficients such
as Cmq, Cmά, Clp, was less accurate than that of
 AME02
The Second TSME International Conference on Mechanical Engineering
19-21 October, 2011, Krabi

4. Trajectory Simulation 4.2 Results and Discussion

4.1 Analysis Setup 4.2.1 Range and Drift
Four combinations of aerodynamic The predicted range and drift of impact
coefficients, namely EXP, COMBINE1, points from 16 runs are presented in Table 2
COMBINE2, and COMBINE3, were applied to and the impact points are illustrated in Fig. 4.
the trajectory simulation. EXP consists of Predicted range from COMBINE1,
aerodynamic coefficients from experiment only. COMBINE2, COMBINE3 runs are about 18% to
COMBINE1 consists of aerodynamic coefficients 25% greater than EXP runs as shown in Table
from Missile DATCOM but uses Cl and Clp from 2. Moreover, it could be seen that the larger
experimental data. COMBINE2 is similar to elevation angle, the more error. The over
COMBINE1 except that it includes only Clp. predicted range could be caused by the under
COMBINE3 consists of all coefficients from predicted predicted CA in subsonic region hence
Missile DATCOM. Table 1 summarizes the less drag. Although CA is over predicted in
combinations of aerodynamic coefficients used supersonic region, most of the flight is in
in simulation. subsonic region as shown in Fig. 3.
Trajectory simulations were performed at Table. 2 Range and drift of impact points
launching elevation angles 30º, 40º, 50º, and 60º
for each set of aerodynamic coefficients. Totally,
sixteen simulation runs were performed. The
fourth order Runge-Kutta method was employed
for integration using time step 0.0001 sec.

Table. 1 Aerodynamic coefficient combinations

Fig. 4 Predicted impact points

 AME02
The Second TSME International Conference on Mechanical Engineering
19-21 October, 2011, Krabi

As shown in Table 2, the drift error hence higher predicted spin rate during the
percentage is more than 700% in some runs. power-on period.
However, the drift error measured in percentage
might not be very meaningful. The drift, which is
a denominator, is very small so the drift error
percentage becomes very large. From the
design point of view, small drift error is
acceptable because it does not affect rocket
sizing and range performance prediction.
Overall, the predicted impact points in
Fig. 5 Spin rate at elevation 60º
COMBINE1 runs are always drifted most
leftward and the predicted impact points in
4.2.3 Velocity
COMBINE3 runs are drifted most rightward as
Fig. 6 compares the Mach number
shown in Fig. 4. All COMBINE runs uses the
measured from the flight test [14] to COMBINE1
same coefficients from Missile DATCOM except
runs. The results from COMBINE2, COMBINE3,
Cl and Clp. So the discrepancies could be
and COMBINE4 runs, which also employ CA
caused by Cl and Clp.
from Missile DATCOM, are almost the same as
4.2.2 Spin rate
COMBINE1, so they are not presented here. For
Fig. 5 compares the spin rate measured
all curves, the Mach number increases sharply
from the flight test [14] and all runs at elevation
to supersonic at the first 1 sec then decreases
angle 60º. The simulated spin rates at other
to subsonic after the first 8 sec.
elevation angles are almost equal so they are
It could be seen that the simulated Mach
not presented here. Note that the elevation
number is close to the flight test data but the
angle of the flight is not specified in [14] so this
predicted maximum Mach number during the
comparison is not conclusive.
first two seconds is lower than the flight test
It could be seen that EXP, COMBINE1
data. However, as previously stated, the
runs, which applied Cl and Clp from wind tunnel
comparison is not conclusive due to the fact that
experiments, predicts the spin rate very closed
the elevation angle of the flight test is not
to the flight test data. The COMBINE2 run does
specified.
not include Cl so there is no induced roll moment
and the spin rate ceases quickly after the power-
on period. The COMBINE3 run uses Clp from
Missile DATCOM, in which the magnitude is
much lower than Clp from experiments. Less roll
damping force is estimated in COMBINE3 run
Fig. 6 Mach number from COMBINE1 runs
 AME02
The Second TSME International Conference on Mechanical Engineering
19-21 October, 2011, Krabi

4.2.4 Angle of attack simulated angle of attack, side slip angle, and
Fig. 7 shows the simulated angle of spin rate are inaccurate.
attack and side slip angle of all runs at elevation So it could be recommended for the
angle 60º. The data at other launching elevation selected rocket that the aerodynamic coefficients
angles, which is not presented, follows the same predicted by Missile DATCOM are used for
trend but smaller in magnitude. impact range and Mach number only.
The EXP run predictes much smaller
angle of attack and side slip angle than other 6. References
COMBINE runs. This result was expected [1] Design of Aerodynamically Stabilized Free
because the magnitude of Cmq from experiments Rocket (1990). MIL-HDBK-762, US Army Missile
is much higher than those of Cmα , Cmά , Cmq Command, AL, USA.
from Missile DATCOM. So the actual damping [2] Bartowitz, M. E. (2008). Determination of
force is higher than the simulated ones. static and dynamic stability coefficients using
Beggar, Master Thesis, Air Force Institute of
Technology, Wright-Patterson Air Force Base,
OH, USA.
[3] Sahu, J. (2005). Time-accurate numerical
prediction of free flight aerodynamics of a finned
projectiles, AIAA Atmospheric Flight Mechanics
Conference and Exhibit, AIAA 2005-3817, San
Francisco, CA, USA.
[4] Kokes, J., Costello, M., and Sahu, J. (2006).
Generating an aerodynamic model for projectile
flight simulation using unsteady time accurate
Fig. 7 Angle of attack at elevation 60º computational fluid dynamic results, ARL-CR-
577, Army Research Laboratory, Aberdeen
5. Conclusion Proving Ground, MD, USA.
In summary, the prediction of impact [5] Sahu, J. (2007). Unsteady flow computations
range using coefficients from Missile DATCOM of a finned body in supersonic flight, ARL-TR-
could give errors up to 25%. Impact drift errors 4230, Army Research Laboratory, Aberdeen
are much smaller than impact range errors if Proving Ground, MD, USA.
considered in the net distance. The greater [6] Costello, M., Gatto, S., and Sahu, J. (2007).
launching elevation, the more errors in theh Using computational fluid dynamics-rigid body
predicted range and drift. The predicted Mach dynamic (CFD-RBD) results to generate
number is close to the flight test data. The aerodynamic models for projectile flight
simulation, ARL-TR-4270, Army Research
 AME02
The Second TSME International Conference on Mechanical Engineering
19-21 October, 2011, Krabi

Laboratory, Aberdeen Proving Ground, MD, effects in Missile DATCOM, AIAA 2005-4833,
USA. the 23rd AIAA Applied Aerodynamics Conference,
[7] Sooy, T. J. and Schmidt, R. Z. (2005). Toronto, Canada.
Aerodynamic predictions, comparisons, [14] Dahlke, C. W. and Batiuk, G. (1990). Hydra
validations using Missile DATCOM (97) and 70 MK 66 aerodynamics and roll analysis, TR-
Aeroprediction 98 (AP98), Journal of Spacecraft RD-SS-90-6, US Army Missile Command,
and Rockets, vol. 42(2), 2005, pp 257-265. Redstone Arsenal, AL, USA.
[8] Abney, E. J. and McDaniel, M. A. (2005). [15] Field Manual 1-140: Helicopter Gunnery
High angle of attack aerodynamic predictions (1996). United States Army, USA.
using Missile DATCOM, the 23rd AIAA Applied [16] Missile DATCOM User’s Manual (1998).
Aerodynamics Conference, Toronto, Ontario, AFRL-VA-WP-TR-1998-3009, US Air Force
Canada. Research Laboratory, USA.
[9] Sigal, A. (2007). Comparative aerodynamic
analysis of a missile with an eight-fin tail, the Appendix A: Aerodynamic Coefficients from
23rd International Symposium on Ballistics, Missile DATCOM
Tarragona, Spain.
[10] Lesieutre, D. J., Love, J. F., Dillenius, M. F.
E., and Blair Jr., A. B. (2002). Recent
applications and improvements to the
engineering-level aerodynamic prediction
software MISL3, the 40th AIAA Aerospace
Sciences Meeting, Reno, NV, USA.
[11] Atik, H., Erdem, B., Ilgaz, M., Kargancioglu,
I. M., Katirci, A., Mahmutyazicioglu, E., and
Yalcin, L. (2008). Prediction capabilities and
comparison of panel, semi-empiric and CFD
codes for missile aerodynamic analysis, the 26th
AIAA Applied Aerodynamics Conference,
Honolulu, HI, USA.
[12] Teo, H. H. (2008). Aerodynamic predictions,
comparisons, and validations using Missile Lab
and Missile DATCOM (97), Master Thesis, Naval
Postgraduate School, CA, USA.
[13] Horton, A. N. and McDaniel, M. A. (2005).
Identification and correction of axial force
prediction discrepancies due to angle of attack
 AME02
The Second TSME International Conference on Mechanical Engineering
19-21 October, 2011, Krabi