See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/320653966

Analysis of air blast effect for explosives in a large scale detonation

Article  in  Korean Journal of Chemical Engineering · October 2017

DOI: 10.1007/s11814-017-0227-6

CITATIONS READS

0 101

6 authors, including:

Kwon Hweeung Sanjeev Maken

Yonsei University Deenbandhu Chhotu Ram University of Science and Technology, Murthal
19 PUBLICATIONS   77 CITATIONS    138 PUBLICATIONS   1,300 CITATIONS   

SEE PROFILE SEE PROFILE

Il Moon
Yonsei University
174 PUBLICATIONS   2,255 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Thermodynamics of Oxygenated Fuel Additive View project

Waste to Energy View project

All content following this page was uploaded by Sanjeev Maken on 29 September 2018.

The user has requested enhancement of the downloaded file.

Korean J. Chem. Eng., 34(12), 3048-3053 (2017) pISSN: 0256-1115
DOI: 10.1007/s11814-017-0227-6 eISSN: 1975-7220
INVITED REVIEW PAPER INVITED REVIEW PAPER

Analysis of air blast effect for explosives in a large scale detonation

Hweeung Kwon*, Kyungjae Tak*, Sanjeev Maken*,**, Hyounsoo Kim***, Jungsu Park***, and Il Moon*,†

*Department of Chemical and Biomolecular Engineering, Yonsei University, 50 Yonsei-ro, Seodaemun-gu, Seoul 03722, Korea
**Department of Chemistry, Deenbandhu Chhotu Ram University of Science and Technology, Murthal-131 039, India
***The 4th R&D Institute-2 Agency for Defense Development, Daejeon 34186, Korea
(Received 11 June 2015 • accepted 15 August 2017)

Abstract−Open Burning/Open Detonation (OB/OD) has been widely used for demilitarization of expired explo-
sives. However, OB/OD effects a variety of hazardous damages to environment. Therefore, using incinerators to treat
expired explosives is required instead of OB/OD. To guarantee the safety of these demilitarization methods, the blast
wave of the explosives should be previously recognized to evaluate the impact of detonations. Although various materi-
als are used to produce explosives, most researches have focused on trinitrotoluene (TNT). Other representative explo-
sives such as research department explosives (RDX) and high melting explosives (HMX) are seldom studied in the
literature. Therefore, our aim was to understand the blast wave of three materials under different geometry throughout
simulations. To improve accuracy and reduce computational time, a zoning technique with Euler-Lagrange coupling
method was used. Due to limitations and difficulties of detonation experiments, simulations were verified by theoreti-
cal models. In case of semi-confined bunker, the simulation results were compared with experimental data, showing a
close match. As a result, cylinder type is the safest incinerator among semi-confined bunker, cylinder, and cube inciner-
ators, in terms of the blast wave.
Keywords: Explosion, Blast Effect, TNT, RDX, Detonation, Scaled Distance

INTRODUCTION erated by various explosives by the TNT equivalent weight, has been
widely used for explosion models [6]. A model with Euler has been
OB/OD techniques have been one of the most popular tools to used to analyze the propagation of the initial shock wave for a spheri-
demilitarize expired explosives for decades. However, OB/OD meth- cal charge of 1 kg TNT equivalent [7]. Mercx [8] analyzed the vapor
ods inherently contain the potential risk of detonation. In addi- cloud explosion with TNT equivalency and the multi-energy method
tion, detonation of explosives can cause massive damage of the in an offshore plant [8]. Liu [9] employed smoothed particle hydro-
environment by the blast wave. Therefore, many countries have dynamics for overpressure prediction [9]. Chapman et al. [10] devel-
developed and used various incinerators instead of OB/OD for oped 1- or 2-dimensional and time-dependent conservation equa-
demilitarization. Since safety is the most important factor in devel- tions [10]. Ning and Tang [11] simulated for a small-scale internal
oping incinerators for demilitarization, it is required to analyze blast overpressure by the use of Livermore Software-DYNA [11]. Jo and
wave under various incinerator geometries. Kim suggest method of explosion analysis based on explosion limit
It is very difficult to analyze a blast wave. Some researchers have concentration in a confined area [12].
conducted detonation experiments to study blast waves under dif- Previous studies have mainly focused on small-scale detonations
ferent geometries and conditions. Baker [1] tested OD of TNT and of TNT. However, recent propellants have used explosives including
analyzed the blast wave by a comparison of experimental and the- RDX and HMX. Therefore, our aim was to predict the overpres-
oretical results [1]. Skacel et al. [2] considered detonation of con- sure of TNT, RDX, and HMX in a large scale open detonation. The
densed explosives inside shock tubes filled with different materials blast wave model embedded in ANSYS AUTODYN was employed
[2]. Chandra et al. [3] focused on dynamic pressure [3]. They ob- and the maximum overpressures were validated with the experi-
tained pressure-time profiles at various locations not only inside but mental data. Detonation simulations were also carried out under
also outside of a shock tube. Rigas and Sklavounos [4] detonated different incinerator types: semi-confined bunker, cylinder, and tube.
dense explosives in a complex tunnel structure [4]. Kowsarinia, This paper consists as follows. In section 2, the basic methodol-
Alizadeh, and Salavati Pour investigated underwater detonation of ogy of blast wave modeling is introduced. Section 3 verifies simula-
hexogen [5]. Because of many limitations to conduct detonation tion results for OD by theoretical models, and section 4 compares
experiments of explosives, simulation becomes a reasonable approach with simulations and experiment dataset in semi-confined bun-
that reduces risks, cost, and time to analyze the blast wave. ker. Section 5 describes the results of different incinerator types.
A TNT equivalent method, which calculates overpressure gen- The final section gives concluding remarks.

†
To whom correspondence should be addressed. DETONATION OR BLAST WAVE MODELING
E-mail: [email protected]
Copyright by The Korean Institute of Chemical Engineers. To simulate detonation of explosives, two different methods have

3048
 Analysis of air blast effect for explosives in a large scale detonation 3049

Table 1. Physical and chemical properties of TNT, RDX, and HMX [16]
Information
Physical and chemical properties
TNT RDX HMX Reference
Chemical formula C7H5N3O6 C3H6N6O6 C4H8N8O8 Budavari et al., 1989; HSDB, 2009; HSDB, 1995
Molecular weight (g/mol) 227.13 222.26 296.17 Budavari et al., 1989; Merck, 1989; HSDB, 1995
Density at 20 oC (g/cm3) 1.63 1.82 1.89 Budavari et al., 1989; Merck, 1989
Detonation velocity (m/s) 6900 8750 9100 -
Heat of detonation (MJ/kg) 4.55 5.98 6.04 Zhang and Chang, 2013
C1 (Gpa) 3.730×102 7.828×102 7.782×102
C2 (Gpa) 3.747 13.35 7.071
r1 4.15 4.4 4.2
r2 0.9 1.288 1 ANSYS AUDYN
ω 0.35 0.32 0.3
e (Gpa) 6 9.79 10.5
Relative effectiveness factor 1 1.6 1.7

been widely used: Euler and Lagrange. In the Euler region, the (EOS) is used for calculations of overpressures generated by deto-
mesh is fixed in geometry and materials flow through the mesh. nations [13].
Therefore, this method enables treating a large amount of material
ω rν ω r ν ωe
flow and deformation of solid. However, it is not suitable for calcu- P = C1⎛1− -------⎞ e + C2⎛1− -------⎞ e + ------
1 2
(2)
⎝ r1ν⎠ ⎝ r2ν⎠ ν
lating the interface between fluid and solid. Contrary to the Euler
method, the Lagrange method moves mesh and distorts by mate- where P, ω, ν, and e are pressure, specific heat, specific volume, and
rial motion. Thus, it is useful for interface calculation, but is diffi- detonation energy, respectively. C1, C2, r1, and r2 are material param-
cult to handle large deformation of a solid. To take advantage of eters and constants. Table 1 shows properties, parameters, and con-
and to overcome disadvantages of each method, the Euler-Lagrange stants for TNT, HMX, and RDX [14,15]. The parameter and con-
coupling method is employed in this study; Euler method calcu- stant values in Table 1 can be applied for detonation simulations in
lates fluid motion and Lagrange method is used for structural the air.
materials.
The detonation consists of two steps. The first step calculates the OPEN DETONATION
initial expansion of explosives detonation with a 1-dimensional
Euler method that considers radial symmetry to reduce simula- The OD for demilitarization of expired explosives is one of the
tion errors and guarantee accuracies. This result is, then, mapped easiest methods. However, considering that explosives are deto-
onto the 2-dimensional geometry with zoning method in the sec- nated in an open area in OD method, the blast waves generated
ond step. The zoning method refines the mesh in the critical regions. by detonation affect environment significantly. OD is simulated
In general, the Lagrange method is simultaneously used in 2- under various amounts of TNT, RDX, and HMX. It is assumed
dimensional geometry with Euler. Eq. (1) explains the mapping that the explosives exist at the center of the site and are detonated
method to yield high resolution and grid in the first step of the at ground level. The smaller the grid size, the higher accuracy and
blast wave expansion: the more computational cost. Therefore, a zoning method is imple-
mented to improve the simulation accuracy with computational
P=(γ −1)ρe (1)
time reduction; 1 cm for the grid size of Euler and Lagrange method
where γ, e, and ρ are the specific heat ratio, internal energy, and is not applied OD because it do not exist structural materials.
density, respectively. The Jones-Wilkins-Lee (JWL) equation of state The materials used for OD are air and explosives (TNT, RDX,

Table 2. Maximum overpressure in the OD cases

Simulation results
Pressure (kPa)
Distance TNT RDX HMX
1 kg 2 kg 3 kg 1 kg 2 kg 3 kg 1 kg 2 kg 3 kg
2m 604 762 918 778 979 1,184 887 1,110 1,330
4m 103 147 189 257 366 0,499 287 0,389 0,517
6m 042 058 098 109 159 0,222 117 0,175 0,245
8m 028 038 046 059 085 0,120 062 0,088 0,132

Korean J. Chem. Eng.(Vol. 34, No. 12)

3050 H. Kwon et al.

Fig. 1. Pressure-time profiles for OD of TNT.

and HMX). The OD simulation for of explosives has a mesh size

of approximately 4,500,000. The indices for detonation performance
are overpressure and impulse as dangerous factors [17]. We focused
mainly on overpressure. The overpressures are measured at four
sampling points in the geometry: two, four, six, and eight meters
from the detonation point. The result is shown in Table 2. The order
of maximum pressures measured at a distance of two meters is the
same with the cases of detonation velocity and heat of detonation:
from TNT (9,180 kPa) to RDX (11,840 kPa) and HMX (13,300 kPa).
Fig. 1 shows the overpressures as a function of time at the four
sampling points for open detonation of TNT. The peaks at the
points are attenuated rapidly. Moreover, the arrival time gap of the
blast wave between neighbor sampling points becomes larger along
with the distance, showing that blast propagation velocity becomes
slower.
In the literature, there is a theoretical model that represents the
relationship between scaled overpressure and scaled distance for
TNT detonation. The definition of scaled overpressure is expressed
in Eq. (5) and calculated by Eq. (6) [18,19]. Kinney and Graham
detonated TNT and established Eq. (6) based on the result of TNT
detonations.
P
Ps = -----o (5)
Pa
Z 2
1616 1+ ⎛⎝ ------e-⎞⎠
p 4.5
-----o = ----------------------------------------------------------------------------------------------- (6)
pa Ze ⎞ 2
Ze ⎞ 2 Ze ⎞ 2
1+ ⎛⎝ ------------ 1+ ⎛ ---------- 1+ ⎛ ---------
-
0.048⎠ ⎝ 0.32⎠ ⎝ 1.35⎠

Ps, Po, and Pa in Eq. (5) are the scaled overpressure (unitless), the Fig. 2. The relationship between scaled overpressure and scaled dis-
peak side-on overpressure, and ambient pressure (1 atm), respec- tance for explosives.
tively. Ze in Eq. (6) is the scaled distance and is defined with TNT
mass and the distance, as shown in Eq. (7).
r - is applied for RDX and HMX before using the theoretical model.
Ze = -----------
1/3
, (7) The conversion factors of TNT equivalent mass for HMX and RDX
mTNT
are shown in Table 1. The drawing procedure for points of Fig. 2 is
Fig. 2 illustrates the relationship not only for TNT but also for as follows: (1) Compute the overpressure of each explosives using
RDX and HMX. Substituting TNT mass to TNT equivalent mass the simulation, (2) Calculate scaled distance using the scaling law,

December, 2017
 Analysis of air blast effect for explosives in a large scale detonation 3051

Fig. 3. Geometry of semi-confined bunker type with explosives.

Table 3. Maximum overpressure in the semi-confined bunker cases

Pressure (kPa)
Experiment Simulation results
Distance TNT Distance TNT RDX HMX
1 kg 2 kg 3 kg 1 kg 2 kg 3 kg 1 kg 2 kg 3 kg 1 kg 2 kg 3 kg
5.37 m 181.28 261.52 424.75
5.37 m 179.93 267.83 492.23 220.85 332.86 621.64 234.65 352.36 649.60
5.37 m 176.71 240.83 387.97
5.47 m 159.48 263.99 382.80
5.47 m 166.21 259.01 382.92 198.32 310.45 473.40 214.51 336.30 498.22
5.47 m 158.23 253.04 326.67

(3) Put a point for overpressure values with scaled distance in 1.8 m, as shown in Fig. 3. TNT, RDX, and HMX are located at the
graph (Fig. 2), and (4) Compare to simulation results and theoreti- center of the semi-confined bunker. One, two, and three kg of explo-
cal model. According to Fig. 2, these simulations are ground-burst sives are used as the open detonation cases. Park and Lee studied
with ground reflection effect, and it can be concluded that the explosion pressure effects of different ignition points by experi-
trend relationship between scaled overpressure and scaled dis- ment [20]. In this study, two sampling points to measure overpres-
tance shows a close match with the theoretical model. The over- sure exist at 5.37 m and 5.47 m away from the detonation point.
pressure of RDX and HMX is bigger approximately 1.18 and 1.23 Table 3 shows the experiment and simulation results. The exper-
times than TNT. iment was tested twice for each case of TNT detonation. Half of
the results for TNT detonation simulation have an overpressure value
DETONATION IN VARIOUS GEOMETRIES in the range between two experimental results. The other over-
pressure values in the simulation such as 5.37 m for 2 kg TNT and
To analyze the geometry effect on the blast wave, three semi- 5.47 m for 1 kg TNT overestimate the experimental results. How-
confined geometries are considered in this study: bunker, cylinder, ever, the experiment shows a large gap between twice-measured
and cube incinerators. overpressure values; these seem to be acceptable.
1. Semi-confined Bunker Type In Table 3, although the increasing ratio of overpressure from
The semi-confined bunker has its length of 20 m and radius of 1 kg to 2 kg is different from that of overpressure from 2 kg to 3 kg

Korean J. Chem. Eng.(Vol. 34, No. 12)

3052 H. Kwon et al.

for all the explosives cases, there is little effect of explosives type on 5.47 m for 1 kg explosives, 5% decreased for 2 kg explosives, and
the trend. For instance, an overpressure at 5.37 m in the semi-con- 23% for 3 kg explosives.
fined bunker case is increased approximately 50% from 1 kg to 2. Cylinder and Cube Types
2 kg and 85% from 2 kg to 3 kg regardless of explosives types. This Fig. 4 shows geometries for the cylinder and cube types. The
trend is similar to the experimental result. The overpressure in the length for these cases is 20 m, which is the same with the semi-
semi-bunker type is approximately 9% decreased from 5.37 m to confined bunker case. Radius of 1.8 m was used for the cylinder

Fig. 4. Geometry of cylinder and cube types with explosives.

Table 4. Maximum overpressure in the cylinder cases

Simulation results
Pressure (kPa)
Distance TNT RDX HMX
1 kg 2 kg 3 kg 1 kg 2 kg 3 kg 1 kg 2 kg 3 kg
5.37 m 98.31 159.91 240.37 116.93 190.85 289.17 124.19 205.20 312.10
5.47 m 60.13 132.14 166.24 070.97 144.86 198.54 075.71 169.11 214.80

Table 5. Maximum overpressure in the cube cases

Simulation results
Pressure (kPa)
Distance TNT RDX HMX
1 kg 2 kg 3 kg 1 kg 2 kg 3 kg 1 kg 2 kg 3 kg
5.37 m 170.54 237.47 369.31 207.94 293.99 459.42 219.69 308.28 482.84
5.47 m 144.61 232.14 266.82 173.24 279.26 322.85 184.31 299.14 345.35

December, 2017
 Analysis of air blast effect for explosives in a large scale detonation 3053

case. For the cube case, each side of 3.6 m was employed. ACKNOWLEDGEMENTS
The results for cylinder and cube types are in Tables 4 and 5,
respectively. Among three geometries, the cylinder type is the saf- The authors acknowledge the financial support of The Next-
est in terms of overpressure in comparison with the other two Generation Converged Energy Materials Research Center and
types. In general, the overpressure ratios are changed along with Agency for Defense Development.
different geometries: about 63% increase from 1 kg to 2 kg and
51% from 2 kg to 3 kg for the cylinder type and 40% and 56% for REFERENCES
the cube type. Moreover, the amount of explosives significantly
affects decreasing ratio of overpressure at 5.37 m to overpressure at 1. W. E. Baker, Explosion in air, University of Texas, Austin, 150-163
5.48 m. As with the increasing ratio of overpressure by the amount (1973).
of explosives, the decreasing ratio of overpressure by the distance 2. R. Skacel, B. Janovsky, L. Dostal and J. Svihovsky, J. Loss Prev. Pro-
has little relation to explosives type. Therefore, it is concluded that cess Ind., 26, 1590 (2013).
overpressure generated by RDX or HMX can reasonably well pre- 3. N. Chandra, S. Ganpule, N. N. Kleinschmit, R. Feng, A. D. Holm-
dict overpressure generated by TNT under the same amount of berg, A. Sundaramurthy, N. Selvan and A. Alai, Shock Waves, 22,
explosives and geometry. 403 (2012).
4. F. Rigas and S. Sklavounos, J. Hazard. Mater., A 121, 23 (2005).
CONCLUSION 5. E. Kowsarinia, Y. Alizadeh and H. S. Salavati Pour, IJE Transac-
tions B: Applications, 25, 65 (2012).
To select a safe incinerator type, we focused on blast wave anal- 6. F. D. Alonso, E. G. Ferradas, J. F. S. Perez, A. M. Aznar, J. R. Gimeno
ysis for detonation of OD and various types. We analyzed blast and J. M. Alonso, J. Loss Prev. Process Ind., 19, 724 (2006).
wave to OD of TNT, RDX, and HMX, as compared with theoreti- 7. I. G. Cullis, J R Army Med. Corps., 147, 16 (2001).
cal results. The R-squared values of the simulation results are 0.96- 8. W. P. M. Mercx, A. C. Van den Berg, C. J. Hayhurst, N. J. Robert-
0.97 and it is concluded that the simulation shows a good perfor- son and K. C. Moran, J. Hazard. Mater., 71, 301 (2001).
mance. In addition, the model of the semi-confined bunker was 9. M. B. Liu, G. R. Liu, Z. Zong and K. Y. Lam, Comput. Fluids, 32,
simulated for explosives detonation and Euler-Lagrange coupling 305 (2003).
method was used. To take advantage of them simultaneously, the 10. T. C. Chapman, T. A. Rose and P. D. Smith, Int. J. Impact Eng., 16,
calculation results show a close match with experimental data. The 777 (1997).
blast wave ratios of TNT to RDX and TNT to HMX are 1 : 1.18- 11. P. F. Ning and D. G. Tang, J. Chongging Unversity, 11, 119 (2012).
1.19 and 1 : 1.23-1.30, respectively. Furthermore, the detonation of 12. Y. D. Jo and J. Y. Kim, Korean J. Chem. Eng., 18(3), 292 (2001).
three explosives was also tested for two other geometries: a cylin- 13. G. Baudin and R. Serradeill, EPJ Web of Conferences 10, 00021
der and a cube. In general, the overpressures of semi-confined bun- (2010).
ker, cylinder, and cube types are much higher than OD results due 14. X. Y. Wei, Z. Y. Zhao and J. Gu, Int. J. Rock Mech. Min. Sci., 46,
to geometry shapes. The blast wave in the cylinder is lower than 1206 (2009).
semi-confined bunker and cube due to shock wave distribution. 15. R. Jeremić and Z. Bajić, Scientific-Technical Review, LVI, 58 (2005).
As a result, the incinerator of cylinder type is more safe than semi- 16. J. Lee, J-H. Han, Y. Lee and H. Lee, Int’I J. Aeronautical Space. Sci.,
confined bunker and cube types. 16, 50 (2009).
Finally, through the validation, the Euler-Lagrange coupling- 17. S. G. Cho, K. T. No, E. M. Goh, J. K. Kim, J. H. Shin, Y. D. Joo and
based simulation produced reliable results and this method can be S. Seong, Bull. Korean Chem. Soc., 26, 399 (2005).
utilized to evaluate the safety of incinerator-based demilitarization. 18. F. D. Alonso, E. G. Ferradas, J. F. S. Perez, A. M. Aznar, J. R. Gimeno
In the future work, the detonation of complex explosives will be and J. M. Alonso, J. Hazard. Mater., A137, 734 (2006).
simulated with the proposed geometries and identical environment. 19. G. F. Kinney and K. J. Graham, Explosive shocks in air, Springer,
Berlin (1985).
20. D. J. Park and Y. S. Lee, Korean J. Chem. Eng., 29(2), 139 (2012).

Korean J. Chem. Eng.(Vol. 34, No. 12)

View publication stats