Model for Small Arms Fire Muzzle Blast Wave Propagation in Air

Juan R. Aguilar∗a, Sachi V. Desaib

a
Academia Politécnica Militar. Ejército de Chile. Valenzuela Llanos 623. Santiago 7860251. Chile.
b
U. S. Army, Armament Research, Development and Engineering Center. Picatinny NJ 07608. USA.

ABSTRACT

Accurate modeling of small firearms muzzle blast wave propagation in the far field is critical to predict sound pressure
levels, impulse durations and rise times, as functions of propagation distance. Such a task being relevant to a number of
military applications including the determination of human response to blast noise, gunfire detection and localization,
and gun suppressor design. Herein, a time domain model to predict small arms fire muzzle blast wave propagation is
introduced. The model implements a Friedlander wave with finite rise time which diverges spherically from the gun
muzzle. Additionally, the effects in blast wave form of thermoviscous and molecular relaxational processes, which are
associated with atmospheric absorption of sound were also incorporated in the model. Atmospheric absorption of blast
waves is implemented using a time domain recursive formula obtained from numerical integration of corresponding
differential equations using a Crank-Nicholson finite difference scheme. Theoretical predictions from our model were
compared to previously recorded real world data of muzzle blast wave signatures obtained by shooting a set different
sniper weapons of varying calibers. Recordings containing gunfire acoustical signatures were taken at distances between
100 and 600 meters from the gun muzzle. Results shows that predicted blast wave slope and exponential decay agrees
well with measured data. Analysis also reveals the persistency of an oscillatory phenomenon after blast overpressure in
the recorded wave forms.
Keywords: muzzle blast propagation, small arms fire, atmospheric absorption of sound, muzzle blast wave.

1. INTRODUCTION
Firing a weapon consist of a two phase explosion. The first phase corresponds to the shock-induced chemical reaction of
the ammunition which ignites ammunition propellant and producing the subsequent turbulent flow of heat through the
gun barrel. The second phase represents the re-ignition of unburned propellant and combustion products in the air just
outside the muzzle. Mechanical shock of gun trigger impacting of bullet cartridge starts a shock induced chemical
reaction. As result, a mechanical compression shock wave travels across the propellant at high speed, i.e. the speed of
compression wave in the solid propellant. Once the shock reaches the boundaries of propellant it heats up the
surrounding air and induces an airborne blast wave. The blast wave begins to propagate inside the gun barrel at the speed
of the shock wave in the solid. As the blast wave travels its speed of propagation will decrease asymptotically to the
standard value of speed of sound in air. The muzzle blast flow of the high pressure and high temperature exuding plasma
from the open end of the barrel will introduces a sudden air compression followed by a rarefaction. This is known as the
muzzle blast wave which diverges spherically from the weapon muzzle.
Accurate modeling of small firearms muzzle blast waves is critical to a number of military applications including the
determination of human response to blast noise, gunfire detection/localization, and muzzle based suppressor designs.
Although there is a large amount of research concerning blast waves from small caliber weapons, mostly are studies
oriented to evaluate time averaged blast wave noise in terms of sound pressure levels. On the other hand, models for
muzzle blast waves have been developed, but they are mostly focused to explosion or large caliber muzzle blasts. In this
way, appropriate modeling of blast wave form features, like overpressure, impulse duration, rise time, and of the
variation of these parameters with propagation distance is key to correct modeling of muzzle blast wave propagation in
air.
An acoustic model for predict muzzle blast wave propagation in air of small arms fire is introduced. The model
implements a time domain Friedlander wave with finite rise time which propagates at constant speed and diverges

∗
E-mail: [email protected]

Unmanned/Unattended Sensors and Sensor Networks VIII, edited by Edward M. Carapezza,

Proc. of SPIE Vol. 8184, 81840B · © 2011 SPIE · CCC code: 0277-786X/11/$18 · doi: 10.1117/12.903043

Proc. of SPIE Vol. 8184 81840B-1

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 06/15/2016 Terms of Use: http://spiedigitallibrary.org/ss/TermsOfUse.aspx

 spherically from the gun muzzle. In addition to the inverse square law, the effects in blast wave of atmospheric
absorption of sound were also incorporated into the model. Differential model to predict atmospheric absorption as
function of distance consider both thermoviscous and molecular relaxational processes of nitrogen and oxygen. Instead
of traditional frequency based algorithms, in this work atmospheric absorption of blast waves was implemented in the
time domain. A recursive formula derived from numerical integration of corresponding differential equations was
implemented. Solutions were obtained by means of a Crank-Nicholson finite difference scheme. Theoretical predictions
obtained from our model will be compared to data base of previously recorded real world muzzle blast wave signatures.
Recordings containing gunfire acoustical signatures were taken outdoors by shooting a set of different sniper weapons of
varying calibers and placing the microphones at distances between 100 and 600 meters from the gun muzzle, in steps of
approximately 100 meters. Real world signature analysis yields curves of muzzle blast wave overpressures, durations
and rise times against propagation distance. Finally, time series plots containing simulated blast wave forms with added
thermoviscous absorption were compared to the plots of measured signatures of muzzle blast waves from a weapon type.
The recorded wave forms also reveal the existence of an oscillatory phenomenon that modulates the amplitude of the
exponential decay of after blast overpressure. This phenomenon has been reported previously in the literature and
appears to be highly dependent on the ground surface nature. There are several approaches trying to explain the causes of
this effect including ground acoustical impedance effects and the occurrence of creeping waves. Although providing a
full explanation of this phenomenon, a better characterization of this phenomenon is tried using autocorrelation analysis
and the Wigner-Ville distribution technique applied to the recorded blast wave time series.

2. THEORETICAL BACKGROUND
2.1 Friedlander wave model
Friedlander wave model is assumed to be the appropriate way for predict pressure wave forms produced by impulsive
acoustic sources [1]. Instantaneous acoustic pressure produced by a Friedlander wave with finite rise time is given by:

⎧ 0 −∞ < t < 0
⎪
p(t ) = ⎨ Ps t / b 0≤t <b (2.1)
⎪P (1 − (t − b) / c)e −t / c b≤t <∞
⎩ s

here Ps corresponds to blast overpressure, b is the rise time, and c is the impulse duration. Plot of instantaneous acoustic
pressure produced by a Friedlander wave is illustrated in figure 1.

Figure 1. An ideal Friedlander wave with finite rise-time.

2.2 Atmospheric absorption of blast wave

Besides the effects of geometrical divergence, this approach also considers atmospheric absorption of blast waves.
Atmospheric absorption is a dissipative process, which comprises the effects of viscosity, thermal conductivity,

Proc. of SPIE Vol. 8184 81840B-2

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 06/15/2016 Terms of Use: http://spiedigitallibrary.org/ss/TermsOfUse.aspx

 molecular relaxation, and dispersion. Partial differential equation model of atmospheric absorption considering
thermoviscous process only can be written as
∂p δ ∂2 p (2.2)
=
∂ r 2co 3 ∂t 2

where δ is the diffusivity of sound [2]. Using a Crank-Nicholson finite difference scheme the solution for equation 2.2
yields the following discrete-time recurrence formula
δ dr (2.3)
p (( r + 1), t ) = p (( r − 1, t ) + ( p ( r , (t + 1)) − 2 p ( r , t ) + p ( r , ( t − 1)))
c o3 ts 2

If molecular relaxation processes are included in the dissipative model, equation (2.2) becomes
∂p δ ∂2 p m ∂2 p ∂ p −1 (2.4)
= +
∂ r 2c o 3 ∂t 2 2c o ∂t 2
∑ [2π
j =1, 2
fri +
∂t
]

here m = (cinf co ) 2 − 1 , co is the low frequency speed of sound, cinf is the high frequency speed of sound, and
f r i correspond to the i-th relaxation frequency (i=1, for Nitrogen; i=2 for Oxygen) [3]. Relaxation frequencies for
Nitrogen and for Oxygen are 202 Hz and 18417 Hz, respectively.

3. EXPERIMENTAL PROCEDURES
3.1 Muzzle blast wave simulations
Computer simulations of muzzle blast wave propagation in air were obtained by implementing the Friedlander wave
model given in Eq. (2.1) and adding the effect of thermoviscous absorption of sound given by Eq. (2.2). Sound
absorption is calculated in the time domain implementing Eq. (2.3) in a marching scheme. Although absorption due to
molecular relaxation was evaluated in the model it was observed that produced model instability and consequently
molecular relaxation was discarded and only thermoviscous absorption was considered.

100
Friedlander
Friedlander + Thermoviscous

80

80
a

40

20

2:

20 0 20 40 80 80 100 120 140 180

Time

Figure 2. Muzzle blast waves obtained using proposed model: Solid lined corresponds to ideal Friedlander wave
with finite rise time. Dashed line shows the effect of thermoviscous absorption in blast wave form after
propagating 600 meters.

Proc. of SPIE Vol. 8184 81840B-3

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 06/15/2016 Terms of Use: http://spiedigitallibrary.org/ss/TermsOfUse.aspx

 Muzzle blast wave forms obtained with this approach are shown in the figure 2: solid line plots an ideal Friedlander
wave with finite rise time, and the dashed line plot the same blast after propagating 600 meters. Simulation results show
the model is able to reduce the blast peak overpressure and increasing the blast impulse duration and the rise time. Model
performance will be evaluated using a database of real world data of small arm fire muzzle blast waves.
3.2 Analysis of measured data of muzzle blast wave
Database of previously recorded real world muzzle blast wave signatures was available for analysis purposes.
Recordings containing gunfire acoustical signatures were taken outdoors by shooting a set different set of sniper
weapons of varying calibers and placing the microphone at distances between 100 and 600 meters from the weapon
muzzle, in steps of approximately 100 meters. Real world signature analysis of these recordings yields curves of muzzle
blast wave overpressures, durations and rise times against propagation distance. Figure 3 depicts the attenuation of
muzzle blast peak overpressure and the variation of blast impulse duration as functions of propagation distance
corresponding to three different types of weapons.

IOU
---4--- Type A
Type B
-+ Type C
Bp henical spreading

10

UI
IOU 200 300 400 BOO BOO
Barge Em]

Figure 3. Attenuation of the muzzle blast overpressure (left) and blast impulse duration (right) as functions of
distance for various weapon types.

IOU 200 300 400 600 800

Range Em]

Figure 4. Plot of muzzle blast rise time as function of distance for various weapon types.
While blast peak overpressure is easily calculated as the highest value of instantaneous acoustic pressure, blast impulse
duration is measured as the time interval between the blast start and the first zero crossing after it . For measuring blast
rise time, the definition given by Ligthill, as the time the impulse takes to rise from 5% to 95% of its peak overpressure,

Proc. of SPIE Vol. 8184 81840B-4

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 06/15/2016 Terms of Use: http://spiedigitallibrary.org/ss/TermsOfUse.aspx

 was used. Figure 4 shows the variation of blast rise time as function of propagation distance for the three types of
weapons employed in this experiment.
The recorded wave forms also reveal the existence of an oscillatory phenomenon which modulates the amplitude of the
exponential decay after blast overpressure, see figure 6a. This result has been reported previously in the literature [4] and
appears to be highly dependent on the ground surface nature, e.g. grassland, snow [5]. There are some approaches trying
to explain the causes of this effect including ground acoustical impedance effects and the occurrence of creeping waves
or surface acoustics waves (SAF). The phenomena would be explained by the effect of ground absorption in the reflected
blast wave. In this regard, the paper of Albert and Orcutt [6] briefly describes this phenomenon. Although providing a
full explanation of this phenomenon is beyond the scope of this paper, a contribution to better characterization of this
phenomenon is advanced through signal processing techniques.

Figure 5. Auto Correlation analysis of sniper weapons Type A and Type C. Time series (left), Auto Correlation
(right).

Figure 6. Muzzle blast wave signature of a Type A weapon recorded at 100 m from the muzzle. a) Time series; b)
Wigner-Ville Distribution.

Proc. of SPIE Vol. 8184 81840B-5

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 06/15/2016 Terms of Use: http://spiedigitallibrary.org/ss/TermsOfUse.aspx

 In order to detect oscillatory components from exponential decay, autocorrelation analysis was applied to the muzzle
blast signals. The results of autocorrelation analysis for two different types of weapons, namely Type A and Type C, are
shown in figure 5. In both cases, the technique removes high frequency noise present in the blast signal and helps to
better visualization of the oscillations after blast. Attempting to measure the frequency content of the after blast signal,
high resolution time-frequency analysis using the Wigner-Ville Distribution was applied to the recorded time series.
Resultant time-frequency representation are shown in figure 6 and exhibits the presence of spurious and pseudo
harmonic low frequency component with fundamental frequency in the range of about 200 Hz.
3.3 Model validation
Plots of time series containing simulated blast wave forms with added thermoviscous absorption were then compared to
the plots of measured signatures of muzzle blast waves from a weapon type. In figure 7, solid line corresponds to the real
world blast wave of a Type A weapon recorded at 300 m from the weapon muzzle, the dashed line corresponds to the
signature obtained Friedlander wave model plus thermoviscous absorption.

10
Type A
Friedlander + Thermoviscous

200 400 800 800 1000 1200

Samples ]Fs = 130719 Hz]

Figure 7. Comparison of muzzle blast wave signature of a Type A weapon recorded at 300 m (solid line) and the
corresponding simulation result (dashed line).

Certainly, adjusting conveniently the values of blast rise time and blast duration, the proposed model is able to simulated
real blast wave form. On the other hand, the model gives no account of ground reflection or the SAF phenomenon.

5. CONCLUSIONS
Acoustical model for prediction muzzle blast wave propagation in air has been introduced. The model implements a
Friedlander wave model and considers the effects of thermoviscous atmospheric sound absorption in blast wave
propagations. Molecular relaxation processes, when applied to dissipative model introduces instability and thus was
disregarded.
When compared with real world blast wave forms, the simulation results indicate that the proposed model is able to
produce blast wave forms that qualitatively agree with measured data. It is observed that the model can strongly predict
well the rise slope and the mean of the exponential decay. On the other hand, the model tends to overestimates peak
overpressure, which could be attributable to the low pass filtering effect imposed by the atmospheric absorption model
and is capable of rounding the sharp peak of the Friedlander wave.
Signal analysis of blast wave forms as function of propagation distance indicates that blast duration and rise time can be
well described of propagation distances up to 300 m, beyond this limit the variation of these parameters becomes
unstable. In addition analyses of oscillatory phenomenon after blast impulse reveal some features characteristics of this
occurrence.

Proc. of SPIE Vol. 8184 81840B-6

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 06/15/2016 Terms of Use: http://spiedigitallibrary.org/ss/TermsOfUse.aspx

 REFERENCES

[1] Hamernik, R., and Shueh, K., “Impulse sound: some definitions, physical acoustics and other considerations,” J.
Acoust. Soc. Am. 90 (1), p. 189-196 (1991).
[2] Cleveland, R., Chambers, J., Bass, H., Raspet, R., Blackstock, D., and Hamilton, M., “Comparison of computer
codes for the propagation of sonic boom waveforms through isothermal atmospheres,” J. Acoust. Soc. Am. 100,
3017-3027 (1996).
[3] Pierce, A., [Acoustics: An introduction to its physical principles and applications], The Acoustical Society of
America, New York (1989).
[4] Talmadge, C., Waxler, R., Gilbert, K., and Kulichkov, S., “Observation of low frequency acoustic surface waves in
the nocturnal boundary layer,” J. Acoust. Soc. Am. 124, 1956-1962 (2008).
[5] Cramond, A., Don, C., “Impulse propagation over grassland in presence of temperature and wind gradients,” J.
Acoust. Soc. Am. 78, 1029-1037 (1985).
[6] Albert, D., Orcutt, J., “Acoustic pulse propagation above grassland and snow: Comparison of theoretical and
experimental waveforms,” J. Acoust. Soc. Am. 87, 93-100 (1990).

Proc. of SPIE Vol. 8184 81840B-7

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 06/15/2016 Terms of Use: http://spiedigitallibrary.org/ss/TermsOfUse.aspx