1

Application of Flight Mechanics for Bullets

Timo Sailaranta1

Department of Applied Mechanics,

Aalto University School of Engineering, Finland
P.O. Box 14400, FI-00076 Aalto, Finland
[email protected]

Abstract

A numerical study is carried out to investigate the bullet turning at the

trajectory apex, particularly when fired about vertically. The bullet
model includes an aerodynamic model covering angles of attack up to
180º. Computational fluid dynamics is utilized to estimate the
complex high angle of attack properties. The bullet attitude in flight is
computed based on the quaternion in order to avoid equation
singularity problems. Two separate codes with different body-fixed
coordinate systems are used to verify the results obtained. The role of
Magnus-phenomena in turning at apex and at the descending
trajectory part is particularly studied. The objective of the study is to
find out the terminal falling velocity of realistic 7.62 mm bullet. The
results obtained are compared with sparse terminal velocity
experimental data Based on the results it seems that carelessly
upwards fired projectile can be lethal to life in the vicinity of shooter.

Mathematics Subject Classification : 70E15, 76G25

Keywords : Trajectory simulation, Magnus effect, Shooting accident, CFD

1
Corresponding author
 2

Nomenclature
A axial force
a speed of sound
A
CA axial force coefficient
qS
D
CD drag coefficient
qS
CD zero yaw drag coefficient
0

L
CL lift force coefficient
qS
L
Cl rolling moment coefficient
qSd
Cl
Cl spin damping moment coefficient  pd 
p 
 2V 
M
Cm overturning (pitch) moment coefficient
qSd
 2Cm
Cm p Magnus moment coefficient slope  pd 
 2V 
Cm
Cm
 
pitch damping moment coefficient Qd
2V 
q

Cm r p Stability derivate

C m
Cm 

pitch damping moment coefficient

 d
2V

Cm p Stability derivate

N
CN normal force coefficient
qS
CN C N
 normal force coefficient slope

Y
CY side force coefficient
qS
CY p Magnus force coefficient
 2CY
CY p Magnus force coefficient slope
 pd 
 2V 
 2C Z
CZ p
 pd 
Magnus force coefficient slope
 2V 
 3

CZ q Stability derivate

CZ r p Stability derivate

C Z Stability derivate

C Z Stability derivate

CZ p Stability derivate

CG center of gravity
D drag
d projectile diameter
i Imaginary unit 1
Ix ,
S d 2 3

i A , iB Dimensionless moments of inertia

Iy
 
S d 2
3

Ix inertia moment, longitudinal

Iy, Iz inertia moment, transverse
k variation factor
L lift force
L rolling moment
l length
M overturning moment, pitch moment
V
Ma Mach number
a
N yawing moment
N normal force
p projectile spin rate

p̂ dimensionless spin
pd
2V
1
q kinetic pressure V 2
2
S cross section area (reference area) d2 /4
t½ Time-to-half
d
t* Characteristic time t* 
2V
T Temperature
V velocity
v, w transverse velocity components in body mass center
fixed coordinate system (see Fig 5)
W wind velocity
 4

Y side force
 angle of attack, total angle of attack, yaw angle
 Angle of sideslip
 air density

1. Introduction

Newspapers occasionally notify accidents caused by a falling bullet. Also the

scientific literature reports incidences of celebratory firing as a public health
concern internationally. In Kuwait, after the end of the Gulf War, the Kuwaitis
celebrated by firing weapons into the air. Twenty Kuwaitis were reported to die
from falling bullets. In Los Angeles, between the years 1985 and 1992, doctors
treated 118 people for random falling-bullet injuries at King/Drew Medical
Center, and 38 of them died. Practically all of the injuries were due to happy
holiday weekend revellers [1].

In this paper, a representative 7.62 mm bullet computational model was created

and six degrees of freedom (6-dof) simulations were undertaken to find out some
possible trajectories for upwards fired bullets. The results obtained are believed to
representative for the typical military bullet size and mass considered. In Ref. [2]
the issue was earlier investigated with more generic bullet geometry.

A fairly complex, but standard mathematical model, like the one described in Ref.
[3] is needed to accurately enough capture the phenomena of projectile turning at
around and after the apex at about 2 500 meters altitude. Lack of aerodynamic
data typically prevents this kind of simulations. Particularly the aerodynamic
model has to cover the angles of attack α (also known as the yaw angle) from 0 to
180º at small Mach numbers. A separate aerodynamic model appeared more
appealing way to approach the case studied instead of the combined simulation of
fluid and flight mechanics. The long bullet flight time about 1 minute makes it
still beyond the computing capacity typically available.

It is obvious that the very complex interaction between the bullet and the air
cannot be presented completely with a few aerodynamic coefficients and
parameters. Particularly capturing the flow time-history effects is hardly adequate.
However, reasonable estimates for the most important coefficients are included to
the model and the model applicability is extended by varying the Magnus-moment
coefficient. This aerodynamic interaction between the pitch- and yaw-levels was
found to be crucial for the bullet turning and behavior while falling downwards
with decreasing spin velocity. The phenomena were varied also in [2] but it is
carried out in enhanced manner in this study.

The bullet initial trajectory angle was also varied in the simulations and the
turning mechanism at the apex was particularly studied to estimate the model
uncertainties effect on the terminal velocity. Also some frequency-domain
 5

analysis was utilized to find out the oscillation mode characteristics at crucial
trajectory points. Finally, the falling bullet effect on life was estimated based on
the literature [4].

2. Background of the study

The motivation of the study is to find out a simple modification for typical bullet
geometry to decrease falling terminal velocity. In order to do that a strong
subsonic instability with the bullet tumbling followed by large aerodynamic drag
is wished. At the moment one attempting choice to achieve the goal might be an
octagonal base with a large Magnus-moment involved at small angles of attack.
That would at least decrease the nose first falling bullets high kinetic energy.
However, the flight of ordinary bullet geometry is investigated at first in this study
in order to gain knowledge of phenomena.

Some aerodynamic moment is needed to evoke fast spinning bullet (gyro) turning
and make the bullet centre line to follow the velocity vector. It turned out in the
simulations of [2] that the aerodynamic moment and particularly the Magnus
moment is an important factor for projectile turning at the apex and after that
when fired about straight upwards.

The positive Magnus-moment turns the projectile nose to the direction of the
normal coning motion (clockwise seen from behind in case the projectile is also
spinning clockwise) and makes the turning take place easier at small velocities.
Correspondingly negative interaction moment present at high angle of attacks may
prevent the turning entirely.

The Magnus-moment behavior at large yaw angles is particularly difficult to

estimate even for a stationary case not mentioning the time-dependent case
studied. Also modeling the aerodynamics for the trajectory simulations in tabular
form as a function of Mach number etc is probably not adequate. One obviously
needs simultaneous combined solution for fluid equations and the equations of
motion. The software tool development to do the task is and has been going on
intensively in research laboratories around world (for example [5]). However, it
seems that DES-modeling (Detached Eddy Simulations) may be needed to catch
the Magnus-phenomena properly through the subsonic speed region making the
task extremely tedious [6][7].

For the reason the combined simulations needed are typically not possible for time
being due to the bullet long flight time and often limited computing capacity
available. As a compromise the effect of Magnus is looked for in this study by
varying the coefficient behavior as a function time, Mach number and angle of
attack. Ordinary aerodynamic look-up tables were used even though no actual
time history can be included into the simulations via this approach.
 6

3. Bullet Aerodynamic Model

The aerodynamic properties were estimated at first using a simplified engineering

method of Ref. [8] (see also [2]). Also high angle of attack published data was
studied [9][10][11] in order to end up at least qualitatively correct model for the
geometry studied.

In addition a CFD –effort was carried out to shed light on cumbersome high angle
of attack flight of a fast spinning bullet. Two CFD codes were utilized to make it
possible to judge the quality of the results obtained. The diameter based Reynold's
number is subcritical (<30000) in case and a laminar separation was expected to
occur at high angle of attack from both sides of the bullet [10].

Only one case was analyzed in this study in which the flight altitude was 1000
meters and the falling velocity was taken to be 50 meters per second. The
dimensionless spin is 0.479 with spin 1000 Hz.

Selected Methods for the Numerical Simulation

Two different CFD softwares were used in the simulations: open source and free
OpenFOAM 1.7.0 [12] and commercial ANSYS Fluent 12.1 [13]. The ANSYS
Fluent licenses were provided by CSC - IT Center for Science Ltd.

Based on the theoretical background of the case it was decided to perform the
simulations with purely laminar viscous formulation. The objective was to do both
steady-state and transient simulations of each case. From the start it was assumed
that the steady-state simulations would not be able to capture the Magnus moment
effect correctly but the steady-state results could be used as a starting point for the
transient simulations.

In all simulations the spinning of the bullet was taken into account by placing a
constant angular velocity (rotating wall) boundary condition on the bullet surface.
Since the bullet is axisymmetric this type of boundary condition is totally
sufficient and moving meshes or other more complicated methods need not be
applied.

Simulations were conducted at four different angles of attack: 45º, 90º, 110º and
135º. Other freestream flow parameters were not varied and are listed in table 1.
These are based on the standard atmospheric conditions at an altitude of 1 000
meters.
 7

Table 1 Freestream flow parameters and reference dimensions.

Velocity V = 50 m/s
Pressure p = 89875 Pa
Density ρ = 1.1116 kg/m3
Dynamic viscosity μ = 17.58ˑ10-6 kg/ms
Temperature T = 281.65 K
Reference length d = 7.62ˑ10-3 m
Reference area S = 4.56ˑ10-5 m2
Reynolds number Red = 24 000
Spin rate 6283 rad/s (1 000 rps)

The freestream flow velocity is strictly subsonic and incompressible (Ma = 0.15).
Incompressible pressure-based flow solvers were used in OpenFOAM, but
density-based compressible flow solver was used in Fluent. The density-based
solver in Fluent has succesfully been used in transonic and supersonic simulations
and it can just as well be used in subsonic cases even though compressibility
effects are negligible. The density-based Fluent solver can be used for both
steady-state and transient simulations. In OpenFOAM two different solvers had to
be used. It was decided to to use simpleFoam which is steady-state,
incompressible, laminar and turbulent flow solver for the steady-state simulations.
Transient, incompressible laminar only solver icoFoam was used in the time-
dependent simulations.

Computational Grid and Boundary Conditions

The same computational grid (see Fig. 1) was used with both OpenFOAM and
ANSYS Fluent. The grid was generated with grid generator software Pointwise.
The grid has approximately 2 million hexahedra cells in total. Approximately 10
000 cells form the surface grid. The boundary layer was solved and in order to do
that the mesh was made finer in the proximity of the body. The computational
grid was checked with OpenFOAM checkMesh program and it passed all the cell
geometry tests.
 8

Figure 1. The 7.62 mm projectile computational grid.

Case set-up in OpenFOAM

Freestream boundary condition was used on the incoming flow side of the grid
and fixed pressure and zero gradient velocity condition on the opposite side. The
already mentioned rotating wall velocity boundary condition was used on the
bullet wall. The transient cases were simulated for at least 0.005 seconds which
corresponds to 5 full rotations of the bullet.

Case set-up in Fluent

In Fluent the basic idea of the case set-up was similar to OpenFOAM with some
differences due to the density based solver. The density based solver requires the
energy equation to be enabled and setting of operating pressure and temperature.
Farfield and pressure outlet boundary conditions were used in similar fashion as in
OpenFOAM.
 9

Flowfield Solved

The flowfield obtained at the end of the transient simulation is depicted in Figure
2. The time interval between the frames is 0.00025 s and Figure shows the flow
history around the bullet base during one body rotation.

Figure 2. Flow velocity at the base of 7.62 mm projectile. Dimensionless spin is

and the angle of attack is 135º. Interval between the frames is 0.00025 s
and the bullet rotates once in Figure.
 10

OpenFOAM based yawing (Magnus) moment coefficient time history is depicted

in Fig 3. The coefficient is given in the coordinate system with positive direction
“nose to left”. The coefficient is seen to oscillate strongly with frequency about
1000 Hz. Corresponding Fluent time history shows oscillatory behavior between
values -1.5…1 with the frequency about 1500 Hz.

Figure 3. The 7.62 mm projectile Magnus moment coefficient Cn versus time (

) at 135º angle of attack.

The CFD-results obtained are revisited more closely in other context.

Constructing the Aerodynamic Model

The Magnus moment is crucial for the bullet behavior at trajectory descending
part and the terminal velocity. The bullet effective shape is non-symmetric in
flight at non-zero angle of attack due to spin and viscous phenomena in the
boundary layer. Because of that the aerodynamic moment vector is not oblique to
the level defined by the bullet symmetry axis and velocity vector. The moment
vector tilt (i.e. Magnus-effect) may cause bullet rapid dynamical instability with
increasing angle of attack and drag associated. Also the phenomena may ease or
prevent the bullet turning at around and after the apex while falling backwards at
about 180º angle of attack.
 11

The Magnus moment coefficient Cn ( ) model created for trajectory

simulations at small velocities is depicted in Fig 4 as a function of yaw angle. The
average moment with the oscillation amplitude limits are given in the coordinate
system used in the trajectory simulations.

The moment oscillation frequency is taken to be 1000 Hz in the trajectory

computations. The frequency value about 180 Hz was also applied for a short time
periods to find out the bullet sensitivity wrt this disturbance. The value was
chosen to match the bullet natural frequency at high altitudes.

1.5

1
Magnus moment Cn

CFD results
0.5
Fit 0-60 deg
0
Fit 60-180 deg
0 50 100 150 200
-0.5 CFD oscillation
CFD oscillation
-1
Oscillation fit
-1.5 Oscillation fit

-2

-2.5
Angle of Attack deg

Figure 4. The 7.62 mm projectile Magnus moment coefficient Cn f(α)-model used in the
trajectory simulations ( ).

The positive Magnus moment coefficient value makes the nose to turn to the right
in case positive nose up angle of attack and the positive spin (clockwise seen from
behind of bullet). The corresponding positive force coefficient CY affects to the
right in case of the positive angle of attack and spin.

Some CFD-results published show negative Magnus moment coefficient values at

small angles of attack (not simulated in this study) at subsonic and transonic
region. However, the moment seems to return back to the positive values at yaw-
angles about 5º [6]. This trend is also in good agreement with the results of Ref.
[7]. The narrow negative moment zone at small angles of attack might be a scale
effect since it does not seem to be present in data published for artillery scale
projectiles.
 12

The reverse moment was estimated based on [6][7] and is seen at left in Fig 5.
The negative values were included to the aerodynamic model at angle of attack
2.5º. The value at 5º was based on the fit (see Appendix A) and a smooth
interpolation was applied at the region 0…5º in the trajectory simulations. The
reverse moment evokes a small amplitude coning motion which was found out to
have a negligible effect on the turning phenomena.

Figure 5. The 7.62 mm projectile Magnus moment coefficient Cn f(α)-model at small

angles of attack used in the trajectory simulations ( ).

The static pitching moment Cm obtained (Fig 7) was compared also with an
experimental result (Fig 6) found in literature [11]. The CFD-result obtained was
as expected and the agreement with the behavior obtained experimentally is good.
 13

Figure 6. A published result for a projectile pitching moment coefficient Cm f(α) (at about
Ma=0.03) [11].

Figure 7. The projectile pitching moment coefficient Cm f(α) (CFD –based result).

In addition two moment damping coefficients were included into the bullet
aerodynamic model: The spin damping moment coefficient Clp f(Ma) and the
pitch damping moment coefficient Cmq f(Ma) (see Appendix A). A small average
negative pitch damping moment coefficient without f(α) dependency was chosen
for simulations (see for example [10]).
 14

The normal force small angle slope was taken to be 2 at all Mach numbers. The
high angle of attack fit based on the CFD and the engineering method was
adjusted to give reasonable values particularly at low velocities.

The bullet aerodynamic model is given as schematic closed-form equations in

Appendix A to facilitate easy repetition of simulations. The Aerodynamic forces
and moments are made dimensionless based on the projectile diameter d and cross
section area S =πd2/4.

4. Bullet Geometry Model

The simulations were carried out to a 7.62 mm 9.5 g bullet which was fired
upwards with initial velocity 850 m/s. The bullet data is given in Table 2 and
geometry schematics is depicted Figure 8. The weapon rifle makes one spin while
the bullet travels 304.8 mm (12 inches) resulting to the initial spin value 3150
rounds/s. The numerical values used in the simulations for the bullet closely
resemble the ones given at the context of NATO 7.62*51 mm cartridge
description [14]. Inertia properties of table 1 are estimated in this study.

Table 2. Physical and geometrical data used for the 7.62 mm bullet.
Characteristics Value
diameter 7.62 mm
weight 9.5 g
length 28 mm
center of gravity (CG) 17 mm (from the nose)
nose length 14 mm

moment of inertia Ix 6E-8 kgm2 (longitudinal)

moment of inertia Iy=Iz 4E-7 kgm2 (transverse)
 15

Figure 8. The 7.62 mm bullet computational model and true geometry [14].

5. Trajectory Simulation Model

Two different 6-dof simulation codes were written in order to simulate the bullet
flight. The mathematical model needed to accurately enough capture the
phenomena is described in many text books (see for example in Ref. [3]).

The projectile body-fixed coordinate system used is depicted in Figure 9 and the
Figure 10 includes also the earth fixed coordinate frame used. Either the wind-
axis (Lift, Drag) or body-coordinate (Normal force, Axial force) based
aerodynamic model may be used in the simulations. The formulae needed to
transform the data between the systems are also given in Appendix A.

The projectile body-fixed coordinate system was defined in two different ways in
two separate simulation codes used to verify the results obtained. The spinning
mass center fixed coordinate system requires a very short integration time step
compared with the so-called zero-spin coordinate system also used. The
computation procedures with quaternion-based attitude system are explained in
more details in [2] and [3].
 16

Figure 9. The projectile body-fixed coordinate system. The positive moments and angular
velocities are also depicted. The total angle of attack α is the angle between the
xb-axis and the velocity vector V. The applied aerodynamic forces in the Figure
are those in the wind coordinate system (D=Drag, L = Lift and S = side force).

Figure 10. The earth-fixed and the projectile body-fixed coordinate system. The positive
z-coordinate points down inside the globe.

In the trajectory simulations, the aerodynamic forces and moments were at first
obtained based on the total angle of attack value and the body mass center fixed
coordinate system components were obtained taking use of the bullet velocity
components. For example the normal force N lateral components are
 17

 v  W yb 
 
N yb  C N ( , Ma)qS   (1)

 v  W   w  W 
yb
2
zb
2



 
 w  Wzb 
N zb  C N ( , Ma)qS   (2)

 v  W   w  W 
yb
2
zb
2



The cross-coupling (Magnus) body coordinate terms were naturally obtained

using the velocity numerator terms crossed.

The trajectories were integrated numerically and the bullet oscillation eigenvalues
were also solved during the flight path evaluation. The atmosphere model used
was the ISO standard one.

In order to find out the eigenvalues i.e. natural oscillation modes the equations
were Laplace-transformed and the characteristic polynomial was solved
numerically to obtain the roots in frequency domain [15]. The equation used is

   3
 2  i B  C Z i B  i   pˆ 0 C Z i B    s
    p  
 C Z i B  2 C m  C Z C m 
  q  q 
 2 2 
  pˆ 0 C Z  i A  pˆ 0 C Z  C m 
 p p rp

 2C  2
  2  C m
 C C
Z q m
 ˆ
p 0 Z rp m 
C s
p
 
   pˆ C i B  2  pˆ 0 i A  2  pˆ 0 C m  C Z pˆ 0 i A  C Z pˆ 0 C m  
  0 Z p rp   rp  
  i   pˆ C C m  2  p0 C m
ˆ  C Z p0 C m
ˆ  p0 C Z C m
ˆ

  0 Z 

  
p q 
p q 
p rp

C C  pˆ 02 C i A  pˆ 02 C C m  2 C m  C Z C 
 Z m q Z
p
Z
p rp  q m 

 
 2  
   pˆ 0 C Z C m  i  C Z pˆ 0 i A  C Z pˆ 0 C m  pˆ 0 C Z C m    s  0
 rp p    rp p q 

  2  pˆ C  C ˆ C
p  ˆ C
p C 
 0 m p Zq 0 m p 0 Z rp m 
  (3)
 18

6. Results and Discussion

The simulated bullet terminal velocities (TV) are depicted in Fig 11 as a function
of launch elevation angle. The initial trajectory angle (elevation) and aerodynamic
model were varied in the study as explained in the previous chapter.

The bullet turning nose down will take place at the launch angle 80º despite the
yawing moment oscillating (see Fig 12) frequency. At higher launch angles the
bullet will fall either the base first or about sideways depending on the moment
oscillation. The CFD-based aerodynamic model oscillation obtained (1000 Hz) is
seen as a white noise by the bullet and the bullet will land the base first with
velocity about 80 m/s.

The bullet fast oscillation mode period time was found out to be about 0.0055 s
and the bullet/fluid resonance was obtained by increasing the Magnus-moment
period time up to about value 0.0055 s. A fairly short time resonance, say 10
oscillation periods is enough to change the bullet flight entirely. The resonance
does not appear anymore at launch angles smaller or equal to 80º.

160

140
Terminal velocity m/s

120

100

80
No resonance
60
Resonance
40

20

0
0 10 20 30 40 50 60 70 80 90
Launch angle deg

Figure 11 The bullet terminal velocity as a function of initial elevation (trajectory) angle. The
Magnus-moment oscillation frequency was varied in simulations.
 19

Figure 12 The bullet Magnus-moment behavior vs. angle of attack (an example obtained
from the trajectory simulations).

The bullet velocity, angle of attack and Euler angle θ histories are depicted in Fig
13, 14 and 15 in case 86º initial elevation angle. The bullet oscillation is seen to
increase rapidly with increasing speed and air density at the late part of
descending phase.
 20

900

800

700
Velocity m/s

600

500

400 No resonance

300 Resonance

200

100

0
0 10 20 30 40 50 60 70
Time s
Figure 13 The bullet velocity as a function of flight time with the initial elevation (trajectory)
angle 86º. The Magnus-phenomena are also varied in Figure.

200
180
160
Angle of Attack deg

140
120
100
No resonance
80
Resonance
60
40
20
0
0 10 20 30 40 50 60 70
Time s
Figure 14 The bullet total angle of attack as a function of flight time with the initial elevation
(trajectory) angle 86º. The Magnus-phenomena are also varied in Figure.
 21

100
90
80
70
Theta deg

60
50
No resonance
40
Resonance
30
20
10
0
0 10 20 30 40 50 60 70
Time s
Figure 15 The bullet vertical Euler angle θ as a function of flight time with the initial
elevation (trajectory) angle 86º. The Magnus-phenomena are also varied in Figure.

The bullet flight history will change dramatically if frequencies are made to match
for a short time after the apex. Now the fast mode oscillation is excited with
increasing angular velocity and the bullet will land almost sideways with high
drag and low velocity associated. The matching of the frequencies with the
resonance as a result might be possible in reality also since the CFD-analysis
carried out in this study was extremely limited. For example only one
dimensionless spin value was used in the flow simulations.

Some variation of the pitching moment coefficient Cm f(α) -trend was also
performed in [2] with minor effect on the bullet terminal velocity. The coefficient
was not varied in this study since its effect is considered negligible.

The turning process at around the apex was also investigated in [2] by letting wind
turbulence velocity fluctuations to affect on bullet flight. The results obtained
indicated that at least the conceptual bullet of [2] was fairly insensitive to
disturbances.
 22

Bullet Terminal Velocity Effect

At launch angles ≤ 80º the nose down landing bullet possess at least the estimated
minimum lethal energy 40 J ([4]) after falling down from the altitude of about 2.5
km. The energy corresponds now to TV of about 92 m/s. At large launch angles the
skull penetrating speed 60 m/s [16] is mostly clearly exceeded without velocity retarding
resonance present. With oscillation initiated the skull penetrating speed is not always
reached.

Only experimental comparison result found by authors for 0.3 cal (7.62 mm)
bullet gives the terminal velocity for upwards fired bullet about 300 ft/s (about 90
m/s) [16]. Nearest corresponding result of this study would be about 80 m/s in
case landing base first without the resonance initiated oscillation.

7. Conclusions

A relatively realistic computational model was created for a 7.62 mm military

bullet and its turning around the apex was investigated particularly in case it was
fired about straight up to the air. The terminal velocities were determined as a
function of the elevation angle with the Magnus phenomena varied. The
computational study undertaken shows that the bullet terminal velocity may be
120…135 m/s (~ 455 km/h ~ 285 mph) when the launch angle is ≤ 80º. In the
launch angle region of 80…90º the terminal velocity might vary between values
40…80 m/s. The result depends on possible Magnus-moment caused bullet
oscillations. Obviously one needs to study more closely this particular case. Some
more variations of aerodynamics etc are needed to figure out the all relevant
factors for the terminal velocity.

The terminal velocities obtained when shooting about straight upwards may not
always be lethal but the careless shooting can at least be considered dangerous
despite the firing elevation angle. Although the bullets falling at terminal velocity
are traveling slowly, they do travel fast enough to cause significant injury and
death. In most cases of this study the bullet possessed the estimated minimum
lethal energy 40 J at the end of trajectory. The skull penetrating speed 60 m/s is
mostly clearly exceeded.

The danger caused to the shooter was estimated to be fairly small because of
minimal hit probability and the simulation indicated bullet smaller landing
velocities when fired about straight upwards.

Recent in-house studies show that drastic subsonic instability might be evoked by
properly redesigning the bullet base area. The base form causing that and not
affecting remarkably the bullet ballistics at supersonic speeds might be for
example octagonal shape. The use of new geometry might at least limit nose down
 23

falling bullet caused damages. Some of the falling energy would be wasted away
by additional drag caused by the bullet tumbling. However, an unstable rotating
bullet caused possible extra damages are also to be considered at this context.

References
[1] A. N. Incorvaia, D. M. Poulos, R. N. Jones and J. M. Tschirhart, Can a
Falling Bullet Be Lethal at Terminal Velocity? Cardiac Injury Caused by a
Celebratory Bullet, The Annals of Thoracic Surgery, Ann Thorac Surg
2007(83):283-284, 2007,
http://ats.ctsnetjournals.org/cgi/content/full/83/1/283
[2] T. Sailaranta, A. Pankkonen and A. Siltavuori, Upwards Fired Bullet
Turning at the Trajectory Apex, Applied Mathematical Sciences, pp 1245-
1262, Vol. 5, 2011, no. 25-28, Hikari Ltd.
[3] P. H. Zipfel, Modeling and Simulation of Aerospace Vehicle Dynamics,
AIAA Education Series, 2000.
[4] Jaro Hokkanen, Putoavan luodin lentomekaniikka ja iskuvaikutukset (A
falling bullet flight mechanics and shock effects), Bachelor`s Thesis, Aalto
University, 2011 (in Finnish).
[5] Lauri Vesaoja, Coupling of Flow Computation and Flight Mechanics,
Master Thesis, pages 119, Aalto University, Department of Applied
Mechanics, 2009.
[6] S. Doraiswamy and G.V. Candler, Detached Eddy Simulations and
Reynolds-averaged Navier-Stokes Calculations of a Spinning Projectile,
AIAA Journal of Spacecrafts and Rockets, Vol. 45 No. 5, September-
October 2008.
[7] S. I. Silton, Navier-Stokes Computations for a Spinning Projectile from
Subsonic to Supersonic Speeds, AIAA Journal of Spacecrafts and Rockets,
Vol. 42 No. 2, March-April 2005.
[8] T. Sailaranta and A. Siltavuori, AeroFi – Technical report, HUT
Laboratory of Aerodynamics, report T-255, 2008 (unpublished).
[9] USAF Stability and Control DATCOM.
Published by Flight Control Division, Air Force Flight Dynamics
Laboratory. Revised edition. Wright-Patterson Air Force Base, Ohio, 1978.
[10] F. G., Moore,
Approximate Methods for Weapon Aerodynamics.
Progress in Astronautics and Aeronautics Series, vol. 186, AIAA, 2000.
[11] Takashi Yoshinaga, Kenji Inoue and Atsushi Tate, Determination of the
Pitching Characteristics of Tumbling Bodies by the Free Rotation Method,
Journal of Spacecraft, Vol. 21, No. 1, Jan.-Feb., 1984, pages 21-28
[12] OpenFOAM 1.7.0, http://www.openfoam.com/
[13] ANSYS Fluent 12.1, http://www.ansys.com/
[14] http://en.wikipedia.org/wiki/7.62x51mm_NATO
 24

[15] T. Sailaranta, A. Siltavuori, S. Laine and B. Fagerström, On projectile

Stability and Firing Accuracy. 20th International Symposium on Ballistics,
Orlando FL, USA, 23-27 September 2002, NDIA.
[16] http://en.wikipedia.org/wiki/Celebratory_gunfire
 25

Appendix A : Bullet aerodynamic model

The trajectory integration was carried out in body-fixed coordinate (see Fig 5) system.
The aerodynamic forces and moments are made dimensionless using reference length d,
ref. area πd2/4 and the kinetic pressure. The Angle of attack is in radians in the formulae
below.
Table 1. Bullet Aerodynamic Properties.

Coefficient Formulae Limits Eq

zero drag force coefficient CDo  0.2 Ma<1
CDo  0.5  (Ma  1) / 10 Ma>=1
axial force coefficient C A ( )  cos( ) C Do α = 0...90 deg
2

 sin(2 ) / 5
C A ( )  0.5cos( )  sin(2 ) / 4 α = 90..180 deg

Cm ( )  2.8 sin(2 ) / 2 α = 0…90 deg

pitching moment coefficient
Cm ( )  1.4 sin(2 ) / 2 α = 90..180 deg

normal force coefficient C N ( )  2 sin   0.8(sin  ) 2 α = 0…180 deg

Magnus moment coefficient Cnp ( )  0.1(sin(3 ) α = 0...60 deg

(pd/(2V)=1)
Magnus moment coefficient Cnp ( )  0.45 sin(1.5(  60 / 180)) α = 60...180
(pd/(2V)=1) deg
Magnus moment coefficient Cnp  0.1Ma  0.1 @ 2.5 deg
(pd/(2V)=1)

Magnus force coefficient CYp ( )  3(sin  ) 5 α = 0…90 deg

(pd/(2V)=1)
CYp ( )  3 sin  α = 90..180 deg

spin damping moment coeff Clp  0.035  Ma / 150

pitch damping moment Cmq  2  8Ma 2 Ma<1
coefficient
Cmq  10 Ma>=1
 26

The often used wind-coordinate aerodynamic force-system (L, D and S in Fig 9) is

replaced in this study by aero-ballistic force system (N = normal force in the xbV-level
oblique to the xb-axis and A = axial force parallel with xb). The aerodynamic force
coefficients in the wind coordinate system can be obtained from

C D ( )  C N sin   C A cos  (4)

C L ( )  C N cos   C A sin  (5)