Applied Thermal Engineering 112 (2017) 484–496

Contents lists available at ScienceDirect

Applied Thermal Engineering

journal homepage: www.elsevier.com/locate/apthermeng

Research Paper

Cook-off analysis of a propellant in a 7.62 mm barrel by experimental

and numerical methods
Halil Is
ık ⇑, Fatih Göktas

Turkish Military Academy, Mechanical Engineering Department, Turkey

h i g h l i g h t s

Cook-off is an involuntary self-ignition phenomenon of a propellant.

Temperature distribution in a barrel and the cook-off time was determined.

The outer wall temperature of the barrel was measured with a thermal imager.

Various parametric analyses have been made that affect the cook-off time.

The tests were conducted according to the NATO testing standards.

a r t i c l e i n f o a b s t r a c t

Article history: Cook-off is an involuntary self-ignition phenomenon of a propellant in a cartridge often encountered in
Received 6 January 2016 multiple sequential firings. In this study, an efficient approach has been used for cook-off analysis of a
Revised 26 September 2016 propellant by experimental and numerical methods. Various firing tests have been conducted with a
Accepted 16 October 2016
7.62 mm rifle to determine the temperature distribution in a barrel and the cook-off time. The outer sur-
Available online 17 October 2016
face temperature at the hottest area on the combustion chamber where the cook-off process was
assumed to initiate was measured with a thermal imager as an effective method. For numerical analysis
Keywords:
the model of the combustion chamber combined with the cartridge was created and analyzed using the
Cook-off (self-ignition)
Transient heat transfer
ANSYS 14.5 Academic finite element solver and the temperature distribution of the inner/outer surfaces
7.62 mm barrel of the combustion chamber was determined. The numerical results for temperature distribution were
ANSYS observed to be quite close to the temperature values measured by the thermal imager. The numerical
Finite element method analysis was also validated by comparing the numerically determined cook-off time with the experimen-
tal measurements. At the end, various parametric analyses have been made that affect the cook-off time
of a propellant.
Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction get accurately with a highest velocity. The ability of multiple

Throughout history, the issue of having more powerful and
deterrent weapon systems against threats and attacks has always
been among the priority areas for countries to ensure their securi-
ties. Therefore, countries have tried to take the necessary precau-
tions in order to protect themselves against the enemy and
developed more effective and powerful weapon systems. In this
context, studies have led to the development of the ballistic science
that examines the events occurring in a firing process of a weapon
system. The studies on the development of weapon systems are
particularly focused on firing a projectile with a high firing rate
(number of shots per minute) as far as possible and hitting the tar-
⇑ Corresponding author.
E-mail addresses: halil.isik@yahoo.com, hisik@kho.edu.tr (H. Is
ık).
sequential firing with a high firing rate provides a significant
advantage for a weapon to hit the target effectively. Although,
the multiple firing ability is a very important feature for rifles it
brings together some problems as well. The expansion of the barrel
to high temperatures during firing process causes the combustion
gas pressure in the barrel to fall down and the erosion/abrasion in
the barrel to increase which have negative effects on a barrel
material.
One of the major problems resulting from an overheated com-
bustion chamber of a weapon in multiple firings is the cook-off
(self-ignition) phenomenon. It is generally confined to automatic
guns [1]. When the propellant in the cartridge case starts to burn,
it is quickly converted into combustion gases and very high
pressures and temperatures are generated inside the barrel. The
heating process continues until the projectile leaves the barrel. In

http://dx.doi.org/10.1016/j.applthermaleng.2016.10.104
1359-4311/Ó 2016 Elsevier Ltd. All rights reserved.
 H. Is
ık, F. Göktas
/ Applied Thermal Engineering 112 (2017) 484–496 485

multiple sequential firings the time interval between two firings is numerical thermal study of the 30 mm MK44 barrel to determine
about 0.1 s. This means the barrel continues to be heated nearly the relationship between the number of rounds fired and the time
without having enough time to cool down. The inner surface tem- to initiate propellant and explosive cook-off in the barrel.
perature of the combustion chamber increases to high values after The previous studies in determining cook-off time of a propel-
numerous firings. As a new cartridge is fed into the barrel and is lant for certain amounts of sequential firings were done mostly
left in place for a while, because of higher temperature the heat by using experimental methods. This is a costly, dangerous and
is transferred in a very short time from the inner surface of the unsafe way which also requires special test setups. The aim of this
combustion chamber to the new cartridge case and from the case study was to develop an effective method for cook-off analysis of a
to the propellant. If the temperature of the propellant rises above propellant by using numerical method. For validation of the
the ignition temperature, the event called cook-off starts to occur numerical analyses experimentally, various firing tests have been
and the projectile is fired as the propellant ignites spontaneously. conducted with a 7.62 mm rifle to determine the temperature dis-
Generally, the firing and heating processes continue for large num- tribution in a barrel and the cook-off time. First of all firing tests of
bers of shots until the ignition temperature is reached which leads single shot, 5 shots in 5 s intervals and multiple sequential firing of
to cook-off. The consequences of this involuntary and sponta- 20 shots were conducted with a 7.62 mm rifle. At the end of each
neously occurring event may be disastrous and the shooter can firing test the temperature values on the outer wall of the combus-
easily be injured. Therefore, the cook-off time and temperature tion chamber were measured by using a thermal imager. This was
for a certain number of rounds fired in periodic intervals should also a new practical and sensitive measurement method compared
always be considered by a weapon designer. to the traditional methods that use thermo-couples. Later, a con-
The studies on heat transfer from the barrel and cook-off pro- stant heat flux value determined iteratively was used in the
cess are as follows. Lee et al. [2] determined the unknown time- numerical analyses as boundary condition instead of gas tempera-
dependent heat flux at the inner surface of a chrome-coated gun ture and convection coefficient which are more difficult to deter-
barrel, in which the interlayer thermal contact resistance between mine. Then, the cook-off times were measured experimentally by
the steel cylinder and the coating was taken into account. They conducting multiple sequential firing tests with a 7.62 mm rifle
estimated the heat flux and the thermal stresses by using the tem- for 180, 200 and 220 shots. The numerical results obtained from
perature values obtained from the solver and confirmed the results ANSYS finite element solver were compared with the experimental
by analytical solutions. Hill and Conner [3] have developed a solver ones. It was observed that the experimental and numerical results
by using Mathematica software to determine the barrel tempera- were confirming each other. Finally, various parametric analyses
ture after multiple firings. They determined the internal ballistic were made to determine the self-ignition time for various number
boundary conditions by using the PRODAS software. It was possi- of shots by changing the firing time interval between two magazi-
ble to calculate the one-dimensional temperature distribution nes, the number of cartridges in each magazine, the thicknesses of
through the barrel thickness by using this solver. Chen et al. [4] the combustion chamber and the cartridge case.
used inverse and input estimation methods to recursively estimate
the transient heat flux and the inner surface temperature of the
2. Theoretical formulation
combustion chamber. Conroy [5] has developed an interior ballistic
code to calculate the one-dimensional two-phase heat transfer in a
In a firing process quite complex events occur in a barrel. All
barrel. Akçay and Yükselen [6] have solved the unsteady tempera-
ture distribution of a machine gun barrel numerically. Chung et al. these events are concerned with interior ballistics of weapons
which is the science of converting the chemical energy of a propel-
[7] derived an empirical equation to get the wear rate for a 40 mm
gun barrel as a function of heat input. They compared the experi- lant into heat and kinetic energy. After initiating the combustion
process of the propellant in the cartridge case by the impact of
mental data with the results obtained from empirical equations.
S
entürk et al. [8] investigated interior ballistics of a 7.62 mm gun the firing pin on the cartridge cap, the propellant starts to burn
quickly and it is converted into combustion gases which generate
barrel using experimental, numerical and analytical methods with
a thermo-mechanical approach. Huang et al. [9] studied the ther- very high pressures and temperatures. Then, the gases push the
projectile forward and as the heat transfer continues from the
mal effects during the firing process with 5.56 mm ceramic gun
barrel by using finite element method to test the usability of
ceramics on various weapons. They compared the experimental
data with the simulation results. Fuller [10] tried to determine
the temperature values on a small-caliber barrel by using a ther-
mal imager. Bin et al. [11] studied cooling of a 155 mm heavy
weapon barrel theoretically by opening channels to solve the
self-ignition problem. They confirmed the theoretically calculated
values by finite element method. Hameed et al. [12] developed
an experimental test setup to determine the self-ignition time for
7.62 mm double base propellant. Mishra et al. [13] performed
numerical transient thermal analyses of midwall-cooled and exter-
nally cooled gun barrels. By using the results they tried to find the
self-ignition temperature and time. Riel et al. [14] identified the
self-ignition time of the propellant for XW-7 rocket by determining
the temperature distribution resulting from combustion of the pro-
pellant. Hasenbein [15] studied the thermal effects in determining
time-temperature relationships for cook-off of XMl23EI Zone 8
propelling charges in the 155 mm XM199 Cannon. Then, he
generated a curve which depicted time to cook-off as a function
of maximum chamber temperature. Jacobs [16] developed a graph
showing the temperature vs. self-ignition time for the fuel of a
ship. Witherell and Pflegl [17] conducted an analytical and Fig. 1. Heat transfer in the first step.
486 H. Is
ık, F. Göktas
/ Applied Thermal Engineering 112 (2017) 484–496

gases toward the chamber wall as shown in Fig. 1, the inner surface
temperature of the combustion chamber wall increases further
until the projectile leaves the barrel. It is not easy to estimate the
amount of heat transfer because the combustion and heat transfer
process occurs within milliseconds. After leaving the barrel the
weapon mechanism throws the empty cartridge case out and feeds
a new cartridge into the combustion chamber. As a new cartridge is
fed into the barrel and is left in place for a while, heat is transferred
in a very short time from the inner combustion chamber wall to
the new cartridge case and from the case to the propellant as
shown in Fig. 2. If the temperature of the propellant rises above
the ignition temperature of the propellant the projectile is fired
spontaneously which is called as cook-off.
The combustion gases expand and fill whole of the combustion
chamber in less than a millisecond. It was assumed that the total
energy released by the gases in the combustion chamber is dis-
tributed homogeneously to the whole chamber volume and the
heat flux through the entire inner surface of the combustion cham- Fig. 3. 7.62  51 mm standard NATO ammunition in the combustion chamber.
ber was taken as constant. The constant heat flux value was deter-
mined iteratively using trial and error method by estimating its
value and comparing the experimental data with the numerical The cross-section of the combined form of the propellant, car-
results obtained by utilizing the estimated heat flux value until tridge case and the combustion chamber wall is shown in Fig. 4.
both results confirmed each other. The heat flux value determined The heat transfer in a cook-off phenomenon is examined in two
for the propellant used in this study is equal to 89 W/mm2. steps. First, after combustion of the propellant the heat transfer
The numerical results showed that the hottest area was near the takes place from the inner surface of the cartridge case toward
narrow end of the cartridge case shown in Fig. 3. This area was the outer surface of the barrel. Then, after feeding the next car-
observed as the thinnest section where the heat transfer was tridge into the chamber the heat transfer takes place from the
assumed to be the fastest and the cook-off process initiated. Then, heated inner surface of the barrel through the contacting surface
all the numerical analyses were repeated at that thinnest section in of the cartridge case toward the propellant.
two dimensions. As the two and three dimensional results were The initial and boundary conditions for the heat transfer pro-
compared to each other it was observed that they both confirm cesses are given as follows:
each other. So, after the confirmation it was decided to make the First step:
parametric analyses in two dimensions.
For t ¼ 0; r1 6 r 6 r3 Tðr; tÞ ¼ T 1 ð2Þ
The general three-dimensional transient heat conduction equa-
tion in cylindrical coordinates is written in the following form [18]: 
@T 
For t > 0; r ¼ r 1 qıı ¼ k  ¼ constant



 @r r¼r1
1 @ @T 1 @ @T @ @T v @T 
kr þ 2 k þ k þ q ¼ qC ð1Þ   @T 
r @r @r r @h @h @z @z @t r ¼ r 3 hair ðT w;r3  T 1 Þ þ er T 4w;r3  T 4surr ¼ k  ð3Þ
@r r¼r3
where r; h and z are the radial, circumferential and axial coordi- Second step:
v
nates, k thermal conductivity (W/m2 K), q the heat generation per
For t ¼ 0; 0 6 r < r 2 Tðr; tÞ ¼ T 1
unit volume (W/m3), q the density, t the time, T the temperature
distribution and C the specific heat (J/kg K). r2 6 r 6 r3 Tðr; tÞ ¼ T prev ;r2 ð4Þ

Fig. 4. Combined cross-section of the burnt gases, cartridge case and combustion
Fig. 2. Heat transfer in the second step. chamber wall.
 H. Is
ık, F. Göktas
/ Applied Thermal Engineering 112 (2017) 484–496 487

@T
For t > 0; r ¼ 0 ¼0
@r 
  @T 
r ¼ r 3 hair ðT w;r3  T 1 Þ þ er T 4w;r3  T 4surr ¼ k  ð5Þ
@r r¼r3

where r 1 is the radius of the cartridge case inner surface, r2 is the

radius of the contact surface between the cartridge case and the
combustion chamber wall, r 3 is the radius of the combustion cham-
ber outer surface, hair is the convection heat transfer coefficient of
air, e is the emissivity of the barrel surface, r is the Stefan-
Boltzmann constant equal to 5.67  108 W/(m2 K4), qıı is the heat
flux, T w is the wall temperature, T 1 is the ambient temperature
nearby the outer surface of the barrel, T prev is the temperature
obtained from the previous iteration and T surr is the absolute tem-
perature (K) of the surrounding surface enclosing the barrel for
radiation heat transfer.
The convection coefficient hair for low speed (1.0–1.5 m/s was
assumed) air flow around the barrel was calculated from Eq. (6) for
cross flow over a cylinder [18]. After calculation the convection
Fig. 6. Thermal imager measurement setup.
coefficient for the air flow around the barrel was obtained as nearly
25 W/m2 K throughout the analysis.

for RePr > 0:2

h Re
5=8 i4=5
Nub ¼ hairk D ¼ 0:3 þ 0:62Re

1=2 1=3

Pr
1 þ 282000 ð6Þ
2=3 1=4
1þ0:4
Pr

where Nub is the Nusselt number for the barrel, D is the barrel outer
diameter, Re is the Reynolds number and Pr is the Prandtl number.
All the properties used in Eq. (6) were evaluated at the film temper-
ature T f ¼ ðT 1 þ T s Þ=2. The emissivity for the barrel surface coated
with manganese phosphate was assumed to be 0.85.

3. Experimental work

The firing tests for the cook off process were conducted with a
7.62 mm G3 Rifle on the firing test setup shown in Fig. 5. The tests
were conducted by using spherical double based 7.62  51 mm
standard NATO ammunition according to the NATO testing stan- Fig. 7. Thermal measurement point of the location of the loaded cartridge in the
barrel.
dards [19]. In multiple sequential firing tests the G3 Rifle fired
whole of the 20 cartridges in a magazine in 2 s. Accordingly, the
time period between two shots was taken as 0.1 s. This time period
was divided into two parts. First, the time period that elapsed from
the beginning of combustion of the propellant until the projectile
leaves the muzzle was taken as approximately 1.2 ms. During this
very short combustion time period there will be heat flux in both
positive and negative directions. But, the negative heat flux was

Fig. 5. The shooting test setup with a G3 rifle. Fig. 8. 2-D cross-section at the thinnest section of the cartridge case.
488 H. Is
ık, F. Göktas
/ Applied Thermal Engineering 112 (2017) 484–496

neglected in the analyses as its magnitude was observed to be
much smaller than the positive heat flux, which is assumed to be
constant as 89 W/mm2. The reasonably close match obtained
between the calculated and measured outer barrel temperatures
validates the assumption of neglecting negative heat transfer. In
the remaining 98.8 ms the weapon mechanism threw the cartridge
case out and fed a new cartridge from the magazine into the barrel.
In multiple firing tests the time period measured for replacing a
magazine was approximately 5 s. During the replacement time
period of a new cartridge case and a magazine the constant heat
flux value resulting from the combustion gases is set to zero
because of no combustion. But, the heat transfer from the barrel
to the cartridge continues because of the temperature difference
between the cartridge case and the heated barrel wall. Before start-
ing the numerical analyses all the time-dependent initial and
boundary conditions were put into the ANSYS program. According
to this information, ANSYS solves the problem using the proper
time steps.
To determine the cook off time of a weapon various amounts of
multiple sequential firing tests were conducted. At the end of all
firing tests a new cartridge was fed immediately into the barrel
and waited for nearly 30 min. If the propellant in the cartridge
was ignited during this time, it was defined as the cook-off time
of that number of shots.
The transient temperature distribution of the outer wall of the
barrel was measured with a FLIR thermal imager as shown in
Fig. 6. The data of temperature readings were obtained by using
the ‘‘FLIR-ResearchIR max” program. As the cartridge was fed into
the barrel, the hottest area of the cartridge case where the cook-off
process was assumed to occur corresponds to the point shown in
Fig. 7. Therefore, the temperature values for all the tests were mea-
sured around that point. This method for measuring the barrel wall
temperature distribution was quite effective and the results came
out to be reasonable.
4. Numerical analysis
The three-dimensional model of the system which consists of
the propellant, cartridge case and combustion chamber wall of
the barrel shown in Fig. 8 was created in ANSYS 14.5 Design Mode-
ler as seen in Fig. 9. One of the most significant steps in a numerical
solution method is to create an appropriate mesh in the solution
domain to get accurate results. It is also very important that the
results are independent of the mesh. To check for the mesh inde-

Fig. 9. 3-D transparent model of system consisting of the propellant, cartridge case and combustion chamber wall of the barrel in, (a) 3-D view, (b) 2-D view.

Fig. 10. Mesh of the 3-D solid model of the system in, (a) full view, (b) half view.
 H. Is
ık, F. Göktas
/ Applied Thermal Engineering 112 (2017) 484–496 489

Table 1
Meshes with various number of elements.

Direction Mesh 1 Mesh 2 Mesh 3 Mesh 4

r 5 18 26 18
h 28 28 28 28
z 60 60 60 60
Ts,o (°C) 17.403 18.575 18.575 18.575

Fig. 11. The mesh formation around the barrel.

Table 2
Time steps used in the analysis.

Analysis Initial time Minimum Maximum Ts,o (°C)

step (s) time step (s) time step (s)
1 105 106 104 18.7
2 104 105 103 18.7
3 103 104 102 18.7
4 102 103 101 18.7

Table 3
Material properties used in ANSYS.

Material Density Specific heat (J/kg K) Thermal conduction

(kg/m3) coefficient (W/m K)
AISI 4140 Steel 7850 150  C  200  C ! 473 100  C ! 42:6
350  C  400  C ! 519 200  C ! 42:2
550  C  600  C ! 561 400  C ! 37:7
600  C ! 33
Cartridge case 8600 162 111
(Brass)
Propellant 1600 1465 0.18

pendency the mesh around the barrel was analyzed with various
grid numbers and the mesh was created as seen in Fig. 10. For ver-
ification of the numerical analysis various mesh configurations
were tested as shown in Table 1. The optimum mesh size that gave Fig. 12. Three dimensional temperature distribution in the system for, (a) single
the converged solution was taken as Mesh 2 as seen in Fig. 11. shot, (b) 5 shots in 5 s intervals, (c) multiple sequential firings of 20 shots in 2 s.
Then, the required time step was checked as seen in Table 2. The
ANSYS program takes the time steps in analysis 2 as default. Ini-
tially, the program chooses automatically the initial time step perature Ts,o (°C) of the barrel. In this study, the time steps of anal-
and if required it changes the time step to min or max values until ysis 2 were chosen. In the numerical analysis the number of
the solution will be converged. As it is shown in Table 2 the time iterations was taken as 10,000 and the residual (Res) between
steps for all analyses gave the same result of the outer surface tem- the initial ðT t Þ and last ðT tþ1 Þ temperature was taken as 104 and
490 H. Is
ık, F. Göktas
/ Applied Thermal Engineering 112 (2017) 484–496

Table 4
The comparison of the barrel outer surface temperatures.

Firing test type Outer Outer Outer surface

surface surface temperature
temperature temperature (Experimental)
(2-D) (°C) (3-D) (°C) (°C)
Single shot 18.65 18.7 18.7
5 shots with 5 s intervals 30.4 30.4 30.4
Sequential firing of 20 shots 64.6 64.5 64.2
within 2 s

Fig. 15. Temperature (°C)-Time (s) curve for multiple firings of 20 cartridges within
2 s obtained by the ANSYS analysis and the thermal imager.

Table 5
Confidence intervals of numerical and experimental temperatures.

Analysis Confidence Interval Decision

Single shot 0:1275 < C int < 0:2375 No Difference
5 shots with 5 s intervals 0:5188 < C int < 0:9256 No Difference
20 shots within 2 s 10:2832 < C int < 11:7743 No Difference

Fig. 13. Temperature-Time variation for a single shot obtained by the ANSYS
Table 6
analysis and the thermal imager.
The cook-off times of the firing tests and numerical analysis.

Number of Cook-off time measured Cook-off time obtained from

shots in firing test numerical analysis
180 + 1 No cook-off Not rose to cook-off temperature
200 + 1 Cook-off in 18.8 s Rose to cook-off temperature of
173 °C in 18.3 s
220 + 1 Cook-off in 5.1 s Rose to cook-off temperature of
173 °C in 5 s

yses were repeated in two-dimensions. It was observed that the

results of 2-D and 3-D models were confirming each other. So,
the rest of the analyses were made at the maximum temperature
point which corresponds to the thinnest part of the cartridge case
instead of examining the entire chamber. In the numerical analysis
the variable material properties were put into the ANSYS finite ele-
ment solver as given in Table 3.

5. Results and discussion

In this section, first the results of the numerical analyses were

validated and discussed by comparing them with the experimental
Fig. 14. Temperature-Time variation for multiple firings of 5 cartridges with 5 s results of various firing tests. This was followed by several para-
intervals obtained by the ANSYS analysis and the thermal imager.
metric analyses causing and affecting cook-off process.

checked by the convergence criteria as jT tþ1  T t j 6 Res. The solu-
tion was assumed to be converged when the convergence criteria
were assured or the number of iterations exceeded 10,000.
As it was explained in Section 2, after various firing tests and
numerical analyses the hottest area of the inner combustion cham-
ber surface was determined where the heat transfer was assumed
to be the fastest and the cook-off initiates. Then, the two-
dimensional model and mesh of the cross-section of the maximum
temperature point were created as shown in Fig. 11 and the anal-
5.1. Validation of the numerical analysis
The numerical analysis was validated in two steps by compar-
ing the numerical and experimental results of various firing tests
to each other.
In the first step, three firing tests of single shot, 5 shots in 5 s
intervals and multiple sequential firings of 20 shots within 2 s
were conducted and after completion of these tests the outer sur-
face temperature values of the combustion chamber were mea-
 H. Is
ık, F. Göktas
/ Applied Thermal Engineering 112 (2017) 484–496 491

sured by using the thermal imager. Then, three-dimensional Fig. 9. As explained in Section 2 the constant heat flux value used
numerical analyses for all the three firing tests were made using in the numerical analysis was determined iteratively as
ANSYS 14.5 Transient Thermal Solver on the model shown in 89 W/mm2. The convection coefficient for the outside air was taken

Fig. 16. The temperature variations at the inner and outer barrel surfaces obtained from the numerical analysis for multiple firings of, (a) 180 shots, (b) 200 shots (c) 220
shots.
492 H. Is
ık, F. Göktas
/ Applied Thermal Engineering 112 (2017) 484–496

as 25 W/m2 K. The numerical analysis results of the combustion
chamber temperature distribution for the three types of firing tests
are shown in Fig. 12. It was observed from that figure that the hot-
test area of the inner combustion chamber surface was for all the
three types of firing analyses at the same section near the narrow
end of the cartridge case. Then the numerical analyses were
repeated at that section in two dimensions. The maximum outer
combustion chamber surface temperature values obtained from
2-D/3-D numerical analyses and the experimental measurements
are compared to each other in Table 4. Also, the experimental
and numerical results of the outer combustion chamber surface
temperature variation with respect to time are compared to each
other as seen in Figs. 13–15. It was observed that the results of
2-D/3-D models and experimental measurements were confirming

Fig. 17. Temperature variation and the cook-off time of the propellant for different amounts of shots for time intervals of Dt = 5–30 s between two magazines.
 H. Is
ık, F. Göktas
/ Applied Thermal Engineering 112 (2017) 484–496
each other. So, it was decided to continue for the rest of parametric
cook-off analyses at the section of maximum temperature only in
two dimensions.
At first, the firing test for a single shot was conducted and the
outer surface temperature of the barrel was measured by using
the thermal imager. After analyzing the problem numerically with
the same conditions the experimental and numerical results of the
outer surface temperature were observed that they showed nearly
similar behavior and rose up to 18.7 °C from an initial temperature
of 16 °C as shown in Fig. 13.
The second test was the firing of 5 shots with 5 s intervals.
Before starting with the test the barrel was cooled with air till
the barrel temperature fell to 17.7 °C. It was observed that the
experimental and numerical results of the outer surface tempera-
ture showed again similar behavior and both of the temperatures
rose nearly up to 30.4 °C as shown in Fig. 14.
The third test was the multiple sequential firing test of 20 shots
within 2 s with a barrel having an initial temperature of 12 °C. As
the experimental and numerical results of the outer surface tem-
perature were compared to each other it was observed that the
final temperature measured by using the thermal imager was
64.2 °C whereas the numerically obtained value was 64.5 °C as
shown in Fig. 15. Consequently, it was observed that the numerical
and experimental results were reasonably close and confirm each
other.
Fig. 18. Temperature variation and the cook-off time of the propellant for different amounts of shots for 10–30 cartridges in each magazine.
493
To check the accuracy of results ‘‘Paired t-test analysis” was
made for the experimental and numerical results compared in
Figs. 13–15 [20]. In this analysis first, the differences ðDi Þ between
the numerical and experimental values of temperatures were cal-
culated. After calculating the mean value ðDmean Þ of the tempera-
ture differences, the variance ðVarðDi ÞÞ of the differences was
obtained. At last for confidence level of 95%, the confidence interval
ðC int Þ for the temperature differences was determined as
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Dmean  SND VarðDi Þ 6 C int 6 Dmean þ SND VarðDi Þ ð7Þ
where SND is the standard normal distribution value [21]. The
Paired t-test analysis does not give any significant difference
between the experimental and numerical temperature data as
shown in Table 5.
In the second step, the numerical analyses of the multiple
sequential firing tests of 180, 200 and 220 shots were validated
by comparing the results with the experimental cook-off time
measurements.
With all the necessary data entered into ANSYS, the inner and
outer temperature variations for the whole barrel were analyzed
for the multiple firing tests. Analyses were performed with magazi-
nes each containing 20 cartridges with 5 s intervals between them.
As an input value the initial temperature has been taken as 26 °C
and the convection coefficient for the ambient air was taken as
25 W/m2 K. The temperature variations at the inner and outer bar-
494 H. Is
ık, F. Göktas
/ Applied Thermal Engineering 112 (2017) 484–496

rel surfaces obtained from the numerical analysis for all the tests
are shown in Fig. 16. In the next step, at the end of each numerical
analysis 1 more cartridge was loaded into the barrel and waited for
a while. Meanwhile the numerical analysis was continued and the
temperature variation of the propellant in the new cartridge was
observed whether or when it reached the cook-off temperature.
The analysis results for 180 cartridges show that the highest tem-
perature rose up to 168.67 °C in 720 s and fell down nearly to
160 °C as the analysis run for 1000 s. So, it could not reach the
cook-off temperature of 173 °C. In the analysis of 200 cartridges
it was observed that the propellant temperature reached 173 °C
in 18.3 s whereas for 220 cartridges in 5 s. So, for both cases the
cook-off process occured.
Later on, with the same conditions and number of cartridges
experimental multiple sequential firing tests were conducted to
compare and validate the numerical analysis. First, nine magazines
were prepared containing twenty cartridges in each and tenth
magazine with only one cartridge. The firing test was conducted
with these nine magazines (180 cartridges) in five second intervals.
After the test was finished the single cartridge in the tenth maga-
zine was loaded into the combustion chamber and waited for
30 min. During this time it was observed that no ignition took
place. The same tests were conducted for 200 and 220 cartridges
and the cook-off times were measured. For 200 cartridges the self
ignition time was measured as 18.8 s and for 220 cartridges 5.1 s.
The comparison of self ignition times for numerical analysis and
experimental firing tests is summarized in Table 6. The results
were observed to be reasonably close and confirm each other.
5.2. Various parametric cook-off analyses
The cook-off time is directly related to the temperature of the
combustion chamber. Any changes to the structure of the barrel
and cartridge case or the firing time interval affect also the
temperature distribution in the combustion chamber. In this
section, to observe the effects on cook-off various numerical
analyses were made using ANSYS by changing the firing period
between two magazines, the number of cartridges in each maga-
zine, the thickness of the combustion chamber and the cartridge
case.
5.2.1. Analysis made by changing the time interval between two
magazines
In multiple sequential firings the time interval between remov-
ing and loading a new magazine is approximately 5 s. In this time
interval for changing the magazine cooling process takes place in
the barrel. So, the time interval between two magazines is a critical
parameter for a cook-off analysis. In this part of the study, the
cook-off time of the propellant for different amounts of shots by
changing the time interval (Dt) between two magazines (contain-

Fig. 19. Temperature variation and the cook-off time of the propellant for different amounts of shots by increasing the combustion chamber thickness in increments of
0.5 mm.
 H. Is
ık, F. Göktas
/ Applied Thermal Engineering 112 (2017) 484–496 495

ing 20 cartridges in each) was analyzed. The analyses were made
for the time intervals of 5, 10, 15, 20, 25 and 30 s beginning with
200 shots each time. Then, the self-ignition time was obtained after
feeding one more cartridge into the combustion chamber and wait-
ing for a while until the propellant temperature reaches 173 °C.
The cook-off of the propellant in the cartridge case within the first
10 s after a new cartridge is feeded into the barrel can cause seri-
ous accidents. Therefore, safe time before cook-off was taken as
10 s. If the cook-off time was obtained as greater than 10 s the
analysis was repeated by adding 20 shots (1 magazine) each time
to the total number of shots. So, for each time interval the amount
of shots that do not cause cook-off was determined.

Fig. 20. Temperature variation and the cook-off time of the propellant for different amounts of shots by increasing the cartridge case thickness in increments of 0.2 mm.
496 H. Is
ık, F. Göktas
/ Applied Thermal Engineering 112 (2017) 484–496
The results of the analyses are shown in Fig. 17. When examin-
ing the results, it was observed that for time intervals of 5 and 10 s
the cook-off process started with 221 shots. As the time interval
was increased to 15 and 20 s the cook-off process started with
241 shots. For the time intervals of 25 and 30 s the beginning of
the cook-off process was observed with 261 and 281 shots. Conse-
quently, it was determined that more shots can be conducted by
increasing the time interval between two magazines.
5.2.2. Analysis made by changing the number of cartridges in each
magazine
This time, the number of shots in each magazine was changed
to 10, 15, 25 and 30. The analysis was repeated beginning with
201 shots and increasing one magazine each time until the cook-
off time dropped below 10 s. The results of the analyses are shown
in Fig. 18. When examining the results, it was observed that for the
time intervals of 5 and 10 s the cook-off process started with 221
shots. As the time interval was increased to 15 and 20 s the
cook-off process started with 241 shots. For the time intervals of
25 and 30 s the beginning of the cook-off process was observed
with 261 and 281 shots. Consequently, it was determined that
more shots can be conducted by increasing the time interval
between two magazines.
5.2.3. Analysis made by increasing the thickness of the combustion
chamber
In this part, the variation of temperature and cook-off time of
the propellant was analyzed by changing the combustion chamber
thickness in increments of 0.5 mm between parametric tests. The
analysis was repeated again beginning with 201 shots and increas-
ing one magazine each time until the cook-off time dropped below
10 s. The results of the analysis are shown in Fig. 19. When exam-
ining the results, it was observed that by increasing the combus-
tion chamber wall thickness 0.5 and 1.0 mm the cook-off process
started with 241 shots. As the combustion chamber wall thickness
was increased by 1.5 and 2.0 mm the cook-off process was
observed to begin with 261 shots. Consequently, it was determined
that the available shot number also increased by increasing the
thickness of the combustion chamber.
5.2.4. Analysis made by increasing the thickness of the combustion
chamber
Finally, the variation of temperature and cook-off time of the
propellant was analyzed by changing the thickness of the cartridge
case in increments of 0.2 mm between parametric tests. The anal-
ysis was also repeated beginning with 201 shots and increasing
one magazine each time until the cook-off time dropped below
10 s. The results of the analysis are shown in Fig. 20. When exam-
ining the results, it was observed that by increasing the thickness
of cartridge case wall by 0.2 and 0.4 mm the cook-off process
started with 221 shots. As the thickness of the cartridge case wall
was increased by 0.6 to 1.0 mm the cook-off process was observed
to begin with 241 shots. Consequently, it was determined that the
available shot number increased by increasing the thickness of the
cartridge case.
6. Conclusion
The cook-off problem was investigated using experimental and
numerical methods. Throughout the study various firing tests have
been conducted with a 7.62 mm rifle for validation of the numeri-
cal analysis. First, firing tests of a single shot, 5 shots in 5 s intervals
and multiple sequential firings of 20 shots were conducted. At the
end of each firing test the outer surface temperatures on the com-
bustion chamber wall were measured with a thermal imager. Later,
the constant heat flux value used in ANSYS finite element solver
was determined iteratively using trial and error method. This heat
flux value was used in the analyses to determine the cook-off time.
Then, the self-ignition time was determined experimentally by
conducting multiple sequential firing tests for 180, 200 and 220
shots. The analyses made with ANSYS were compared with the
experimental results. It was observed that the experimental and
numerical results were closely conforming to each other. Finally,
to observe the effects on cook-off and to determine the cook-off
time various numerical analyses were made in ANSYS by changing
the firing time interval between two magazines, the number of car-
tridges in each magazine, the thicknesses of the combustion cham-
ber and the cartridge case.
In the future, the cartridge case can be modeled three dimen-
sionally including the thermal stresses with a thermo-mechanical
approach.
Acknowledgment
The authors would like to express their thanks to MKE Gazi
Fis
ek Fabrikasi for its valuable support.
References
[1] AMCP 706-252, Research and Development of Materiel, Engineering Design
Handbook, Gun series, Gun Tubes, U.S. Army Materiel Command (1964).
[2] H.L. Lee, Y.C. Yang, W.J. Chang, T.S. Wu, Estimation of heat flux and thermal
stresses in multilayer gun barrel with thermal contact resistance, Appl. Math.
Comput. 209 (2009) 211–221.
[3] R.D. Hill, J.M. Conner, Transient heat transfer model of machine gun barrels,
Mater. Manuf. Processes (2012) 840–845.
[4] T.C. Chen, C.C. Liu, H.Y. Jang, P.C. Tuan, Inverse estimation of heat flux and
temperature in multi-layer gun barrel, Int. J. Heat Mass Transfer (2007) 2060–
2068.
[5] P. Conroy, Gun Tube Heating, Ballistic Research Laboratory, 1991.
[6] M. Akçay, M.A. Yükselen, Unsteady thermal studies of gun barrels during the
interior ballistic cycle with non-homogenous gun barrel material thermal
characteristics, J. Therm. Sci. Technol. 34 (2) (2014) 75–81.
[7] D.Y. Chung, N. Shin, M. Oh, S.H. Yooand, S.H. Nam, Prediction of erosion from
heat transfer measurements of 40 mm gun tubes, Wear (2007) 246–250.

entürk, H. Is
ık, C. Evci, Thermo-mechanically coupled thermal and stress
[8] A. S
analysis of interior ballistics problem, Int. J. Therm. Sci. (2016) 39–53.
[9] X. Huang, P. Conroy, R. Carter, 5.56 mm ceramic gun barrel thermal analyses
with cycled ammunition, in: 23rd International Symposium on Ballistics
Tarragona, Spain, 16–20 April, 2007.
[10] P.W.W. Fuller, Gun barrel temperature study using short time scan thermal
imaging, Instrum. Aerospace Simul. Facil. (1989) 159–167.
[11] W. Bin, C. Gang, X. Wei, Heat transfer in a 155 mm compound gun barrel with
full length integral midwall cooling channels, Appl. Therm. Eng. 28 (2008)
881–888.
[12] A. Hameed, M. Azavedo, P. Pitcher, Experimental investigation of a cook-off
temperature in a hot barrel, Defence Technol. (2014) 86–91.
[13] A. Mishra, A. Hameed, B. Lawton, Transient thermal analyses of midwall
cooling and external cooling methods for a gun barrel, J. Heat Transfer 132
(2010).
[14] G. Riel, R.H.F. Stresau, W.M. Slie, Cook-Off Studies of the Booster XW-7 and
Variants, US Naval Ordnance Laboratory, 1957.
[15] R.G. Hasenbein, The Effects of Temperature on Time to Propellant Cook-Off,
Benet Weapons Laboratory, 1976, 13 May.
[16] G. Jacobs, Naval Weapon Cook-Off Improvement Concepts and Development,
Naval Weapons Center, China Lake, CA, 1981.
[17] M. Witherell, G. Pflegl, Prediction of Propellant and Explosive Cook-Off for the
30 Mm HEI-T and Raufoss MPLD-T Rounds Chambered in a Hot MK44 Barrel,
US Army Armament Research, 2001.
[18] Y.A. Cengel, Heat Transfer: A Practical Approach, second ed., McGraw-Hill,
2003.
[19] AEP-97 Multi Calibre Manual of Proof and Inspection (M-C MOPI) for 5.56 mm,
7.62 mm, 9 mm and 12.7 mm Ammunition, NATO Standard, NATO Army
Armaments Group (NAAG), NATO Standardization Agency (NSA).
[20] E. Değirmenci, C. Evci, H. Is
ık, M. Macar, N. Yılmaz, M.H. Dirikolu, V. Celik,
Thermo-mechanical analysis of double base propellant combustion in a barrel,
Appl. Therm. Eng. 102 (2016) 1287–1299.
[21] A.M. Law, W.D. Kelton, Simulation Modelling & Analysis, McGraw-Hill,
Singapore, 1991.