Mecánica Computacional Vol XXVIII, págs.

529-552 (artículo completo)

Cristian García Bauza, Pablo Lotito, Lisandro Parente, Marcelo Vénere (Eds.)
Tandil, Argentina, 3-6 Noviembre 2009

EVALUATING THE EFFECT OF LARGE AND SPREAD EXPLOSIVES

LOADS

Bibiana M. Luccionia,c, Daniel Ambrosinib,c, Steeve Chung Kim Yuend and Gerald N.
Nurickd
a
Instituto de Estructuras, Universidad Nacional de Tucumán, Av. Roca 1800, 4000 S.M. de Tucumán ,
Argentina, [email protected], www.herrera.unt.edu.ar/iest
b
Facultad de Ingeniería, Universidad Nacional de Cuyo, Centro Universitario - Parque Gral. San
Martín - 5500 Mendoza, [email protected] , http://fing.uncu.edu.ar/
c
CONICET, Av Rivadavia 1917, Cdad de Bs As
d
Blast Impact and Survivability Research Unit (BISRU), Department of Mechanical Engineering,
University of Cape Town, Private Bag, Rondebosch 7701, South Africa, www.bisru.uct.ac.za

Keywords Explosion, soil, blast wave, crater, damage, numerical model.

Abstract. Reports on experiments or numerical analysis involving medium to extreme

explosive devices (> 100 kg TNT) are scarce. Blast wavefront parameters and scaling laws
found in the specialized literature have usually been obtained and used for spherical explosive
charges. Similarly, empirical equations proposed for the evaluation of crater dimensions
produced by explosions on the ground level were obtained for compact charges that is
spherical or cylindrical charges, and explosive masses up to 100 kg of TNT.
This paper presents the numerical analysis of the detonation of explosive charges ranging
from 1000 to 26,288 kg of TNT laid on the ground, mainly widespread in a carpet-like form.
The charges consist of different ordnances stacked in different configurations. The effects of
the charge configurations and mass of explosive on the crater dimensions and blast wave
parameters are investigated. While the cube root scaled distance works well for a relatively
compact charge layout the scaled distance parameter has to be modified for cases where
charges are spread in a carpet-like form. Numerical results are compared with experimental
results of crater dimensions and blast wave parameters. Reasonable agreement with the
experiments is obtained.

Copyright © 2009 Asociación Argentina de Mecánica Computacional http://www.amcaonline.org.ar

530 B.M. LUCCIONI, D. AMBROSINI, S. CHUNG KIM YUEN, G.N. NURICK

1 INTRODUCTION
Explosive devices are available in different sizes. Table 1 lists the suggested classification
for explosive devices, in four main groupings based on the size of the charge, by Nurick et al.
(2006). Category S1 is a device of mass up to 0.1 kg of TNT that enables indoor laboratory
blast testing. Category S2 (0.1-10 kg TNT) consists of devices that require outdoor laboratory
experimentation. The explosive devices in the medium (M: 10-100 kg TNT), large (L: 100-
1000kg TNT) and extreme (E: >1000kg TNT) categories consist of different size weapon-
type systems. See Table 1.

SMALL MEDIUM LARGE EXTREME

Category
S1 S2 M L E
Indoor Outdoor Laboratory Test
Facilities Test Range Test Range
Laboratory /Test Range Range
Up to >0.1 Kg to >10 Kg to >100Kg to
Mass >1000Kg
0.1 Kg 10Kg 100Kg 1000Kg
Hand grenade
Oklahoma
Antipersonne Land-Mine Torpedo
Example War Head Nuclear
l Portable Mines Air Bomb
bomb
Mine
Telling-Smith
and Nurick
(1991);
Formby
Nurick and
Guruprasad and and Lok (2005);
Reference Shave (1996)
Mukherjee (2000); Wharton Hirschfelde
s Nurick and
Jacinto et al (2001); (1996); r et al.
Martin
Hanssen et al (2002) Wharton et (1945)
(1989a);
al (2000)
Nurick and
Martin
(1989b)
Table 1: Categorization of size of explosive devices.

Whilst numerous different tests investigating the response of structures, such as beams and
plates, to blast loading conditions in the S1 category have been published in the open
literature, reports describing structural response using medium to extreme explosive devices
are scarce (Chung Kim Yuen et al., 2008).
Empirical equations for the evaluation of blast wave parameters can be found in the
specialized literature. These equations have been, however, obtained for spherical explosives
of less than 1000 kg of TNT. Moreover, these formulas are based on scaling laws that were
proved to work well for that shape of explosives. There are also several studies related to
blast load assessment and the effect of blast loads for spherical explosives of no more than
1000 kg of TNT on structures (Luccioni et al., 2006). Recently a study investigating craters
created by exploding charges ranging from 120 kg to 1900 kg of TNT was presented
(Ambrosini and Luccioni, 2008). The charge consists of different ordnances stacked in
different configurations corresponding to tests performed at Touwsrivier Training Range
(South Africa) (Chung Kim Yuen et al., 2008). The arrangement of the explosive load was
shown to have significant importance in the final dimensions of the crater.
This paper presents the numerical analysis of blast tests in the extreme category (masses of

Copyright © 2009 Asociación Argentina de Mecánica Computacional http://www.amcaonline.org.ar

 Mecánica Computacional Vol XXVIII, págs. 529-552 (2009) 531

explosive greater than 1000 kg of TNT). The test programme was performed at the Vastrap
Weapons Range, South Africa (Chung Kim Yuen et al., 2008). The charge consists of
different ordnances widespread in a carpet-like form. Numerical results are compared with
experimental results of crater dimensions and blast wave parameters. The effects of the
charge configurations and mass of explosive on the crater dimensions and blast wave
parameters are investigated.

2 EXPERIMENTAL PROGRAM (Chung Kim Yuen et al., 2008)

The tests were conducted on the Vastrap Weapons Range located 1000 km north west of
Cape Town, a vast test area, which is a fairly flat and sandy. Because of its vast area each test
was carried out on a different location on the range leaving the crater resulting from the blast
untouched. 11 blast tests—ranging from 500 to 26,288 kg of TNT were performed. The list of
ordnance used to make up the charge load is listed in Table 2. The blasts were created using
ordnance such as Projectile AS MK 10, Warhead KC5, Warhead KC9, 84mm HE and 90mm
HE shells. Each test comprised a stack of ammunition as required to configure the
predetermined mass. The ordnance was laid out in a carpet-like way on the flat ground in
different stacking pattern to provide the most favourable packing –labour –time layout. A
typical charge lay-out is shown in Fig. 1(a) (Test 10).

a)

b)
Figure 1: Blast test 10. a) Explosive layout; b) Steel plate

Copyright © 2009 Asociación Argentina de Mecánica Computacional http://www.amcaonline.org.ar

532 B.M. LUCCIONI, D. AMBROSINI, S. CHUNG KIM YUEN, G.N. NURICK

Blas Explos. Plate Stand off Measured Crater dimensions (m)

t Mass W thickness distance deflection a b H2
kg TNT h (mm) R (m) δ (mm)
1 1119.8 6 18.50 0.0 8.7 10.8 0.8
3 18.00 5.0
3 22.50 2.5
3 19.50 4.2
2 1135.2 6 8.80 14.7 ---- ---- ----
3 13.35 28.6
3 12.25 37.3
3 8.65 40.7
3 2250.6 3 13.30 35.5 10.3 12.0 0.9
3 15.00 33.1
6 11.30 37.3
3 18.60 22.0
4 3694.8 3 18.20 36.8 10.7 18.7 1.8
6 16.20 18.1
3 27.90 12.8
3 20.70 34.7
5 6945.4 6 21.25 0.0 11.0 23.0 2.2
3 24.65 7.4
3 25.45 0.0
3 20.95 7.9
6 3395.6 3 21.00 9.1 17.0 22.2 2.3
6 14.50 45.2
3 18.00 35.5
3 16.00 50.2
7 13222 6 19.30 4.2 15.3 21.6 ----
3 27.20 15.7
3 22.85 10.4
3 19.20 27.8
8 22054.9 3 26.60 34.8 15.0 25.5 2.7
3 24.90 31.8
3 20.20 60.6
9 600 3 14.80 23.0 8.1 8.1 2.5
10 27569.3 6 17.00 70.2 20.2 27.2 3.0
3 20.80 103.0
3 24.00 50.9
3 20.40 202.0
11 27223.6 6 22.30 70.3 ----- ----- -----
3 25.50 83.0
Table 2: Vastrap tests (Chung Kim Yuen et al., 2008)

Copyright © 2009 Asociación Argentina de Mecánica Computacional http://www.amcaonline.org.ar

 Mecánica Computacional Vol XXVIII, págs. 529-552 (2009) 533

For almost all the tests the dimensions of the craters produced by the explosive loads were
measured as indicated in Fig.2 and presented in Table 2. Three or four (grey (G), red (R), blue
(B), yellow(Y)) quadrangular mild steel plates of 3 and 6mm thick were placed at different
distances from the explosives loads and were subjected to pressure loads generated by the
blast. A plate-clamping station, 700x700 mm2 in size, shown in Fig 1(b), was used to provide
the quadrangular specimen with suitable support to enable the pressure loadings to result in
large inelastic deformations of the exposed area of 500 x 500 mm2. The mid point deflections
of all the plates were recorded and listed in Table 2.

b a
a
H2

Figure 2: Crater sketch

3 EMPIRICAL EQUATIONS

3.1 Introduction
Historically, the analysis of explosions either has predominantly involved simplified
analytical methods (Baker et al., 1983; Kinney and Graham, 1985; Smith and Hetherington,
1994). Nowadays empirical formulas are still obtained from numerical and experimental
studies are very useful to perform quick prediction of the response of soils and structures to
blast load. A brief description of the empirical formulas that are later compared with
experimental and numerical results of blast wave parameters, plate deflections and craters
dimensions are presented in this section.

3.2 Blast wave parameters

When a condensed high explosive is detonated, a blast wave is formed. It is characterized
by an abrupt pressure increase at the shock front, followed by a quasi-exponential decay back
to ambient pressure and a negative phase in which the pressure is less than environmental
pressure.
The most widely used approach for blast wave scaling is Hopkinson’s law (Baker et al.,
1983) which establishes that similar explosive waves are produced at identical scaled
distances when two different charges of the same explosive and with the same geometry are
detonated in the same atmosphere. Thus, any distance R from an explosive charge W can be
transformed into a characteristic scaled distance Z,
Z = R /W 1 / 3 (1)

Copyright © 2009 Asociación Argentina de Mecánica Computacional http://www.amcaonline.org.ar

534 B.M. LUCCIONI, D. AMBROSINI, S. CHUNG KIM YUEN, G.N. NURICK

where W is the charge mass expressed in kilograms of TNT. The use of Z allows a compact
and efficient representation of blast wave data for a wide range of situations. There are many
solutions for the wave front parameters from both numerical solution and experimental
measurements (Baker et al., 1983; Kinney and Graham, 1985; Smith and Hetherington, 1994).
The results are usually presented in graphics, tables or equations based on experimental or
numerical results, such as the following equations,
PS
=
[
808 1 + (Z / 4.5)
2
] (2)
Po 1 + (Z / 0.048) 1 + (Z / 0.32 ) 1 + (Z / 1.35)
2 2 2

Where Ps is the peak overpressure and Po is the atmospheric pressure.

It is important to note that Eq.(2) and most empirical formula found in the specialized
literature are based on the assumption that the blast originates from a spherical charge.
Moreover, the accuracy of predictions and measurements in the near field is lower than in the
far field, probably due to the complexity of blast phenomena (Smith and Hetherington, 1994).

3.3 Determination of the mid-point deflection

Dimensionless analysis provides a useful insight into scaling to enable a better
understanding of the characteristic response of geometrically similar plates subjected to
impulsive loading. Nurick and Martin (1989a; b) presented an empirical relationship to
predict the mid-point deflection–thickness ratio of thin quadrangular plates subjected to
uniform blast load. From this relationship, the mid-point deflection of a 500x500 mm2 and
3mm thick steel plate (static yield stress=250MPa, density=7850 kg/m3) can be calculated
(Chung Kim Yuen et al. 2008) as
δ [mm] = 0.114 I [N s] (3)

where I is the total impulse on the plate that can be approximated as

I = is A (4)

where A is the exposed area of the plate and i s the specific impulse.

3.4 Crater formation

A crater is always formed when an explosive load is detonated on the soil surface. The
crater dimensions defined by Kinney and Graham (1985) are used in this work (Figure 3): D
is the apparent crater diameter, Dr is the actual crater diameter and H 2 is the apparent depth
of the crater.

Copyright © 2009 Asociación Argentina de Mecánica Computacional http://www.amcaonline.org.ar

 Mecánica Computacional Vol XXVIII, págs. 529-552 (2009) 535

Figure 3: Definitions of the crater dimensions.

Studies concerned with the characteristics of craters caused by explosions usually resort to
dimensional analysis and statistics. The scaling law establishes that any linear dimension L of
the crater can be expressed as a constant multiplied by W α divided by the distance of the
charge from the ground, where α is a coefficient that is dependent on whether the
gravitational effects can be neglected or not. When the gravitational effects can be neglected
the cubic root law is applicable α =0.33 and in the other cases the functional dependence can
be quite complex.
There is not much information about explosions at ground level. Statistical studies of about
200 accidental above-ground explosions of relatively large magnitude are presented by
Kinney and Graham (1985). The results exhibit a variation coefficient in the crater diameter
of about 30%. From these results, the following empirical equation for the crater diameter
was proposed (Kinney and Graham, 1985)
D (m) = 0.8W 1 / 3 (kg ) ± 30% (5)
The authors have conducted a series of tests performed with different amounts of explosive
at short distances above and below ground level, as well as on the soil surface (Ambrosini et
al., 2002). They also presented (Ambrosini and Luccioni, 2006) a numerical study on craters
formed by explosive loads located on the soil surface. From these results, the following
equation has been proposed for the evaluation of the apparent diameter of the crater formed
by spherical blast loads laid on the ground,
D(m) = 0.51W 1 / 3 (kg ) ± 5% (6)
The variation of ±5% accounts for the differences between soil properties that could be
found in different sites.

4 NUMERICAL MODELS

4.1 Introduction
All the numerical analysis is performed with a hydrocode (AUTODYN v11.0, 2007). In
order to carry out a comparable analysis, the mass of the explosive is defined by TNT masses.
The corresponding masses for other explosives can be obtained through the concept of TNT

Copyright © 2009 Asociación Argentina de Mecánica Computacional http://www.amcaonline.org.ar

536 B.M. LUCCIONI, D. AMBROSINI, S. CHUNG KIM YUEN, G.N. NURICK

equivalence (Formby, 1996).

An Euler Godunov multi material with strength higher order processor (Alia and Souli,
2006) is used to model the problems including the air, the explosive charge and the soil.

4.2 Material models

4.2.1 Air
The ideal gas equation of state is used for the air. In an ideal gas, the internal energy is a
function of the temperature alone and if the gas is polytropic the internal energy is simply
proportional to temperature. It follows that the equation of state for a gas, which has uniform
initial conditions, may be written as,
p = (γ − 1)ρe (7)
in which p is the hydrostatic pressure, ρ is the density and e is the specific internal energy.
γ is the adiabatic exponent, it is a constant (equal to 1 + R/cv) where constant R may be taken
to be the universal gas constant R0 divided by the effective molecular weight of the particular
gas and cv is the specific heat at constant volume. The values of the constants used for air are
presented in Table 3.

Equation of State: Ideal gas

γ = 1.4
Reference density: ρa = 1.225 10-3 g/cm3
Reference temperature: To= 288.2 K
Specific heat: cv = 717.3 J/kgK
Table 3: Air properties

4.2.2. TNT

Lee-Tarver equation of state (Lee and Tarver, 1980) is used to model both the detonation
and expansion of TNT in conjunction with “Jones - Wilkins - Lee” (JWL EOS) to model the
unreacted explosive.
The (JWL) equation of state can be written as,

⎛ ω ⎞ −r1v ⎛ ω ⎞ − r2v ωe
p = C1 ⎜⎜1 − ⎟⎟e + C 2 ⎜⎜1 − ⎟⎟e + (8)
⎝ r1v ⎠ ⎝ r2 v ⎠ v

Where p is the hydrostatic pressure, v = 1 ρ is the specific volume, ρ is the density, C1,
r1, C2, r2 and ω (adiabatic constant) are constants and their values have been determined
from dynamic experiments and are available in the literature for many common explosives.
The values used for TNT are presented in Table 4.

Copyright © 2009 Asociación Argentina de Mecánica Computacional http://www.amcaonline.org.ar

 Mecánica Computacional Vol XXVIII, págs. 529-552 (2009) 537

Equation of State: JWL

Reference density ρ = 1.658 g/cm3
C1 = 3.7377 108 kPa
C2 = 3.73471 106 kPa
R1 = 4.15
R2 = 0.9
ω = 0.35
C-J detonation velocity: 6.93 103 m/s
C-J energy/unit volumen: 6 106 KJ/m3
C-J pressure: 2.1 107 kPa
Table 4: TNT properties

4.2.3. Soil
A shock equation of state combined with an elastoplastic strength model based on Drucker
Prager criterion and a hydro tensile limit are used for the soil. The initial density is taken as ρ
= 2.2 g/cm3 (wet density). The wet density is obtained considering a mean dry density of
2100 kg/m3 and a moisture content of 5%.
The experimental fact is that for most solids and many liquids, that do not undergo a phase
change, the values on the shock Hugoniot for shock velocity U and material velocity behind
the shock up can be adequately fitted to a straight line

U = co + su p (9)

Where c0 is sound speed.

The Mie-Gruneisen form of equation of state based on the shock Hugoniot is used:

p = p h + Γρ ( e − e h ) (10)

Where Γ is the Gruneisen Gamma G, defined as:

⎛ ∂p ⎞
Γ = v⎜ ⎟ (11)
⎝ ∂v ⎠ v

It is assumed that Γ ρ = Γ0 ρ0 = constant and

Copyright © 2009 Asociación Argentina de Mecánica Computacional http://www.amcaonline.org.ar

538 B.M. LUCCIONI, D. AMBROSINI, S. CHUNG KIM YUEN, G.N. NURICK

ρ o co2 μ (1 + μ ) 1 ph ⎛ μ ⎞
ph = eh = ⎜ ⎟ (12)
[1 − (s − 1) μ ]2 2 p o ⎜⎝ 1 + μ ⎟⎠

The assumption of constant Γ ρ is probably not valid. Furthermore, the assumption of a linear
variation between the shock velocity U and the particle velocity up does not hold for too large a
compression. At high shock strengths some nonlinearity in this relationship is apparent,
particularly for non-metallic materials. This non linearity is covered by a smooth interpolation
between two linear relationships.
A Drucker Prager criterion with standard values is adopted for the strength model. The
yield stress is a piecewise linear function of pressure.
A summary of soil properties used for soil is presented in Table 5.

Equation of State: Shock Strength: Drucker Prager

Reference density ρ = 2.2 g/cm3
Gruneisen Gamma Γ = 0.11
co = 1.614 103 m/s
S =1.5
Shear Modulus G = 2.4 105 kPa
Pressure 1 =-1.149 103 kPa Yield stress 1 = 0 kPa
Pressure 2 = 6.88 103 kPa Yield stress 2 = 6.2 103 kPa
Pressure 3 = 1.0 1010 kPa Yield stress 3 = 6.2 103 kPa
Hydro tensile limit pmin = -100 kPa
Table 5: Soil properties

4.3 Boundary conditions

In order to fulfill the radiation condition, a transmitting boundary is defined for soil
subgrids external limits. The transmit boundary condition allows a stress wave to continue
“through” the physical boundary of the subgrid without reflection. The size of the numerical
mesh can be reduced using this type of boundary condition. The transmit boundary is only
active for flow out of a grid. The effectiveness of this boundary condition is checked in some
of the examples presented in this paper.

5 CRATER FORMATION

5.1 Introduction
The simulation of craters produced by explosive loads widespread in a carpet-like form is
presented in this section. First three blast tests described in section 2 are numerically
reproduced and the results are compared with experimental ones. Once the ability of the

Copyright © 2009 Asociación Argentina de Mecánica Computacional http://www.amcaonline.org.ar

 Mecánica Computacional Vol XXVIII, págs. 529-552 (2009) 539

numerical model has been checked, further numerical analysis is carried out in order to study
the effects of the charge configurations and mass of explosive on the crater dimensions.

5.2 Numerical simulation of Vapstrap tests

Three typical tests covering the range of 1119.8-27569.3 kg TNT were numerically
simulated. These tests correspond to blast tests 1, 5 and 10 in Table 2. In order to carry out a
comparable analysis, the mass of the explosive is defined by TNT masses. Using symmetry
conditions, only a quarter of the problem was simulated. The numerical models used are
presented in Fig. 4. In each model soil, air and TNT were modeled. For clarity air is not
represented in the models shown in Fig.4. The explosive was widespread in the same area as
in the experiment. In the case of test 1, three explosive strips were modeled to represent the
experiments. The mesh was refined in coincidence with the explosive load. Detonation lines
were defined in correspondence with detonators in each test. The simulation was carried out
until the craters remain unchanged.
Fig.5 shows the craters produced by the explosive tests and those obtained with the
numerical models. The experimental and numerical results for the crater dimensions are
presented in Table 6. There is a reasonable agreement between numerical and experimental
results. The differences are in the order of the variability in experimental measures for this
type of tests. The craters simulated are always smaller and more stretched than actual craters
and a good agreement in crater depth indicated as H2 in Fig. 3 is achieved.

Tests Results a(m) b(m) b/a H2(m)

Blast 1 Exper. 8.7 10.8 1.2 0.8
1119.8 kg Numer. 6.3 10.6 1.7 0.7
TNT Difer. % 28% 1.6% -- 16%
Blast 5 Exper. 11 23 2.09 2.2
6945.4 kg Numer. 8.2 22.6 2.75 2.5
TNT Difer. % 25% 1.7% -- -13.6%
Blast 10 Exper. 20.2 27.2 1.34 3
27569.3 kg Numer. 15.76 23 1.45 3.30
TNT Difer.% 21,9% 15.4 -10%
Table 6: Crater dimensions

Copyright © 2009 Asociación Argentina de Mecánica Computacional http://www.amcaonline.org.ar

540 B.M. LUCCIONI, D. AMBROSINI, S. CHUNG KIM YUEN, G.N. NURICK

a)

b)

c)

Figure 4: Numerical models for Vastrap tests. a) Test 1 (1119.8 kg TNT); b) Test 5 (6945.4 kg TNT); c) Test 10
(27569.3 kg TNT)

Copyright © 2009 Asociación Argentina de Mecánica Computacional http://www.amcaonline.org.ar

 Mecánica Computacional Vol XXVIII, págs. 529-552 (2009) 541

a)

b)

Copyright © 2009 Asociación Argentina de Mecánica Computacional http://www.amcaonline.org.ar

542 B.M. LUCCIONI, D. AMBROSINI, S. CHUNG KIM YUEN, G.N. NURICK

c)

Figure 5: Experimental and numerical craters for Vastrap tests. a) Test 1 (1119.8 kg TNT); b) Test 5 (6945.4 kg
TNT); c) Test 10 (27569.3 kg TNT)

5.3 Craters produced by axial symmetric blast loads

In order to study the effect of explosive charge layout on crater dimensions, the craters
produced by the same explosive charges but with cylindrical shape were simulated. Two
different layouts were modeled for each explosive mass: (C) a cylindrical carpet like layout
with the same area in plan than the tests and (M) a cylindrical compact layout with diameter
equal to height. Typical models for both cases are shown in Fig.6. The dimensions of the
explosive cylinders used for each test are presented in Table 7. Due to symmetry conditions,
these problems were simulated with axial symmetric models with a considerable save in
computer time in comparison with actual shape numerical tests presented in section 5.2.

d1
d1 d3=d1
d3

a) b)
Figure 6: Numerical models for axial symmetric numerical tests (6945.4 kg TNT). a) Carpet like explosive (C);
b) Compact layout (M)

Copyright © 2009 Asociación Argentina de Mecánica Computacional http://www.amcaonline.org.ar

 Mecánica Computacional Vol XXVIII, págs. 529-552 (2009) 543

Carpet like explosive layout (C) Compact explosive (M)

W (kg TNT)
d1 (mm) d3 (mm) d1 (mm)= d3 (mm)
1119.8 7440 15.8 814.1
6945.4 7720 91 1757
27569.3 13930 11.1 2780
Table 7: Explosive dimensions for axial symmetric numerical tests

The craters numerically obtained for 1119.8 kg TNT with both (C) and (M) explosive
layouts are shown in Fig.7. The diameters of the carters are presented in Table 8 for
comparison with experimental and numerical equivalent diameter (diameter of the circle with
equal area) of the craters produced by actual shape explosives. It can be seen that the
equivalent diameter of craters produced by cylindrical explosive loads is always smaller than
that obtained for the rectangular layout used in the tests. Moreover, when the explosive is
concentrated in a compact cylinder (M), even smaller craters are obtained.

Rectangular layout Cylindrical layout

Equivalent crater diameter Crater diameter
W (kg TNT) D (m) D (m)
Numer - Carpet Numer.-Compact
Exper. Numer.
like explosive (C) explosive (M)
1119.8 10.9 9.2 8.4 6.0
6945.4 17.9 15.4 11.8 8.5
27569.3 26.4 21.5 18.4 11.1
Table 8: Crater diameters for axial symmetric numerical tests

a)

b)

Figure 7: Craters obtained for axial symmetic numerical tests (1119.8 kg TNT). a) Carpet like explosive (C); b)
Compact layout (M)

Copyright © 2009 Asociación Argentina de Mecánica Computacional http://www.amcaonline.org.ar

544 B.M. LUCCIONI, D. AMBROSINI, S. CHUNG KIM YUEN, G.N. NURICK

5.4 Analysis of results

All the results obtained in previous sections are plotted on Fig.8 representing the
equivalent apparent crater diameter as a function of the cubic root of the equivalent TNT
explosive mass for comparison. The lines representing Eqs. (5) and (6) together with points
corresponding to experimental and numerical results previously obtained by the authors are
also plotted in Fig.8. The points correspond to experimental results from crater tests with
spherical explosive loads of 1-10 kg TNT lying on the ground (Ambrosini et al. 2002),
numerical crater tests with spherical explosive loads of 50-500 kg TNT lying on the ground
(Ambrosini and Luccioni 2006) and numerical crater tests for compact (not cylindrical)
explosive charge layouts of 120-1900 kg TNT reported by Ambrosini and Luccioni (2008).
40

35

30

25

D[m] 20

15

10

0
0 5 10 15 20 25 30 35
W[Kg]1/3
Eq.(5) Eq.(5) (Inf.)
Eq.(5) (Sup.) Eq.(6)
Eq.(6) (Inf.) Eq.(6) (Sup.)
Experim. results (Ambrosini et al. 2002) Numer. results (Ambrosini et al. 2006)
Numer. (Ambrosini et al. 2008) Vastrap tests (Table 7)
Vastrap Numer. (Table 7) Cylind. Numer. (C)
Cylind. Numer. (M)

Figure 8: Apparent crater diameter for explosive load on the ground

The tendency remarked in previous section relating craters produced by different explosive
layouts is clear in Fig.8. While craters produced by carpet like explosives are better
represented by Eq.(5), crater diameters obtained for compact explosives are better represented
by Eq.(6). In both cases, it seems that the linear approximation is only valid for explosive
loads up to the large category (L) (less than 1000 kg TNT). In order to represent the complete
range of explosive masses simulated, the following equations are proposed and represented in
Fig.9 together with experimental and numerical results.
(C) D(m) = 1.7463W 1 / 4 (13)

(M) D(m) = 0.8338W 1 / 4 (14)

Copyright © 2009 Asociación Argentina de Mecánica Computacional http://www.amcaonline

 Mecánica Computacional Vol XXVIII, págs. 529-552 (2009) 545

30

20

D[m]

10

0
0 5 10 15
W[Kg]1/4

Experim. results (Ambrosini et al. 2002) Numer. results (Ambrosini et al. 2006)
Vastrap tests (Table 7) Vastrap Numer. (Table 7)
Cylind. Numer. (C) Cylind. Numer. (M)
Numer. (Ambrosini et al. 2008) Lineal (Eq.13)
Lineal (Eq. 14)

Figure 8: Proposed relationship for apparent crater diameter for explosive load on the ground.

6 BLAST WAVE PARAMETERS AND PLATE DEFLECTIONS

6.1 Introduction
In order to assess the parameters of the blast wave originated from different explosive
layouts, the pressure and impulse time history at points situated at different distances from the
explosive charge center were registered for all the cylindrical blast tests simulated. The gauge
points were located at a height of 350mm in coincidence with the steel plates’ centers in all
cases.

6.2 Pressure
The resulting peak overpressure values as a function of the scaled distance are represented
in Fig.10. Distances are measured from the explosive center. Some points corresponding to
the numerical simulation of Vastrap tests 5 and 10 are also included in Fig.10. These points
are coincident with some of the steel plates in the tests.
For the cases of cylindrical explosives, the same models were run but avoiding blast wave
reflection on the ground and the corresponding results are also plotted on Figs.10. In this way,
the effect of ground reflection can be evaluated. The effect of blast wave reflections on the
ground is important in the case of compact explosives but it is almost negligible in case of
widespread explosives.
The case of blast Test 1 but with cylindrical compact explosive was simulated with a finer
mesh. Results corresponding to the refined mesh are almost coincident with those obtained
with the coarser mesh used in the rest of the numerical models. This result proves that the
refinement used is enough for this type of problems.

Copyright © 2009 Asociación Argentina de Mecánica Computacional http://www.amcaonline.org.ar

546 B.M. LUCCIONI, D. AMBROSINI, S. CHUNG KIM YUEN, G.N. NURICK

a)
80
Empirical (Eq.2)
Numer. Cylind. (M) (Z)
Numer. Cylind. (M) without ground reflection(Z)
60
Numer. Cylind. (M) without ground reflection, refined mesh (Z)
Numer. Cylind. (C) (Z)
ps=(Ps-po)/po

Numer. Cylind. (C) without ground reflection (Z)

40 Numer. Cylind. (C) (Z')
Numer. Cylind. (C) without ground reflection (Z')

20

0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
Z (m/kg^(1/3)) or Z' (m/kg^(1/4))

b)
80

Empirical Eq.2
Numer. Cylind. (M) (Z)
Numer. Cylind. (M) without ground reflection (Z)
60
Numer. Cylind. (C) (Z)
Numer. Cyilind. (C) without ground reflection
ps=(Ps-po)/po

Numer. Cylind. (C) (Z')

40 Numer. Cylind. (C) without ground reflection (Z')
Numer. Rectang. (Z)
Numer. Rectang. (Z')

20

0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
Z (m/Kg^(1/3)) or Z'(m/Kg^(1/4))

Copyright © 2009 Asociación Argentina de Mecánica Computacional http://www.amcaonline.org.ar

 Mecánica Computacional Vol XXVIII, págs. 529-552 (2009) 547

c)

120

Empirical Eq.2
100 Numer. Cylind. (M) (Z)
Numer. Cylind. (M) without ground reflection (Z)
Numer. Cylind. (C) (Z)
80
Numer. Cylind (C) without ground reflection (Z)
ps=(Ps-po)/po

Numer. Cylind (C) (Z')

Numer. Cylind (C) without ground reflection (Z')
60
Numer. Rectang. (Z)
Numer. Rectang. (Z')
40

20

0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
Z (m/Kg^(1/3)) or Z'(m/Kg^(1/4))

Figure 10: Peak side on overpressure vs scaled distance. a) Test 1 (1119.8 kg TNT); b) Test 5 (6945.4 kg TNT);
c) Test 10 (27569.3 kg TNT)

The curves corresponding to empirical Eq.(2) are also included in Figs.10. The comparison
with numerical results shows that while the cubic scale law works well for free field compact
explosions, it is not appropriate for carpet like explosions. Following the results presented by
Chung et al. (2008), a modified scaled distance is defined as
Z ′ = R /W 1 / 4 [m/kg1/4] (15)
Peak overpressure values obtained for the carpet like explosives are also represented as a
function of Z ′ in Figs.10. The resulting points are almost coincident with the empirical curve
corresponding to Eq.(2).

6.3 Impulse
The peak impulse values as a function of the scaled distance are represented in Fig.11.
Like in the case of overpressure values, the impulse values for compact blast loads are greater
than those for carpet like explosives. Nevertheless, the tendency of results is not so clear like
in the case of overpressure values. Impulse values for compact explosive loads follow with
some scattering the empiric curve presented by Kinney and Graham (1985). Points
corresponding to impulse values are closer to that curve when they are represented as a
function of the modified scaled distance Z ′ defined in Eq.(15) but they tend to a constant
value, even greater than that predicted by empirical equations for high scaled distances
( Z ′ > 3 ).

Copyright © 2009 Asociación Argentina de Mecánica Computacional http://www.amcaonline.org.ar

548 B.M. LUCCIONI, D. AMBROSINI, S. CHUNG KIM YUEN, G.N. NURICK

a)

4000
Empirical (Kinney and Graham 1985)
Numer. Cylind. (M) (Z)
Numer. Cylind. (M) without ground reflection (Z)
3000 Numer. Cylind. (M) without ground reflection refined mesh (Z)
Numer. Cylind. (C) (Z)
Numer. Cylind. (C) without ground reflection (Z)
is (Pas)

Numer. Cylind. (C) (Z')

2000 Numer. Cylind. (C) without ground reflection (Z')

1000

0
0 1 2 3 4 5
Z(m/Kg^(1/3)) or Z(m/Kg^(1/4))

b)

4000
Empirical Eq.2
Numer. Cylind. (M) (Z)
Numer. Cylind. (M) without ground reflection (Z)
Numer. Cylind. (C) (Z)
3000 Numer. Cylind. (C) without ground reflection (Z)
Numer. Cylind. (C) (Z')
Numer. Cylind. (C) without ground reflection (Z')
Numer. Rectang. (Z)
Numer. Rectang. (Z')
i s (Pas)

2000

1000

0
0 1 2 3 4 5
Z (m/kg^(1/3)) or Z (m/kg^(1/4))

Copyright © 2009 Asociación Argentina de Mecánica Computacional http://www.amcaonline.org.ar

 Mecánica Computacional Vol XXVIII, págs. 529-552 (2009) 549

c)

4000

3000
i s (Pas)

2000

Empirical (Kinney and Graham 1985)

Numer. Cylind. (M) (Z)
Numer. Cylind. (M) without ground reflection (Z)
1000 Numer. Cylind. (C) (Z)
Numer. Cylind. (C) (Z')
Numer. Cylind. (C) without ground reflection (Z)
Numer. Cylind. (C) without ground reflection (Z')
Numer. Rectang. (Z)
Numer. Rectang. (Z')
0
0 1 2 3 4 5
Z(m/kg^(1/3)) or Z(m/kg^(1/4))

Figure 11. Peak side on impulse vs scaled distance. a) Test 1 (1119.8 kg TNT); b) Test 5 (6945.4 kg TNT); c)
Test 10 (27569.3 kg TNT)

6.4 Plate deflections

With the impulse values empirically (Kinney and Graham, 1985) and numerically
obtained, the plates deflections can be estimated using Eq.(3). Numerical values of impulse
are directly obtained for the different models. Empirical values are calculated using the
modified scaled distance Z ′ . The corresponding mid point deflection values are represented
as a function of the measured mid point deflection in Fig.12.
In all cases, greater plate deflections are predicted for concentrated (M) than for carpet like
explosives (C). Empirical results lie between those corresponding to the two types of
explosive layout simulated: concentrated and spread. Although the points do not lie on the
line representing the coincidence of evaluated and measured deflections, they are close to it. It
seems that the use of Z ′ combined with Eq.(3) works better for smaller blast charges than for
greater charges. Registered deflection is much greater than calculated values for greater
charges.

Copyright © 2009 Asociación Argentina de Mecánica Computacional http://www.amcaonline.org.ar

550 B.M. LUCCIONI, D. AMBROSINI, S. CHUNG KIM YUEN, G.N. NURICK

250
d calc=d exper
Empir. (Z')
200 Numer. Rectang.
Numer. Cylind. (C)
Numer. Cylind. (M)
calc (mm) 150

100

50

0
0 50 100 150 200 250
δ exper (mm)

Figure 12. Plates mid point deflection

7 CONCLUSIONS
The numerical results presented in this paper provide an insight into the effect of large-
scale explosions. The loading condition resulting from the detonation of large amount of
ordnance widespread on the ground in a carpet like fashion has proven to be different to that
originated from the detonation of compact explosives.
Reasonable agreement of numerical results with the experiment was obtained for crater
dimensions. The shape and the dimensions of the crater formed in the underlying soil strongly
depend on the explosive layout. The equivalent crater diameter for carpet like explosives is
always grater than that for compact explosives. Moreover, for carpet like explosives, the
equivalent diameter is greater for rectangular layouts than for circular layouts.
It was also proved that existing empirical formula for the prediction of crater diameter are
not adequate for explosive masses greater than 3500 kg and new expressions covering all the
range of explosives masses, from small to extreme cases, are proposed.
While the cube root scaled distance works well for evaluating the pressure and impulse
values originated from a compact charge layout, the scaled distance parameter has to be
modified to a fourth root for cases where charges are spread in a carpet-like fashion. The
effect of blast wave reflections on the ground are almost negligible for this type of explosive
layout.

8 AKNOWLEDGEMENTS
The financial support of the CONICET (Argentina) and CIUNT (National University of
Tucumán) is gratefully acknowledged.

REFERENCES
Alia, A. and Souli, M., High explosive simulation using multi-material formulations. Applied
Thermal Engineering, 26:1032–1042, 2006
Ambrosini, R.D., Luccioni, B., Danesi, R., Riera, J. and Rocha, M.. Size of Craters Produced
by Explosive Charges on or Above the Ground Surface. Shock Waves, 12(1):69-78, 2002.

Copyright © 2009 Asociación Argentina de Mecánica Computacional http://www.amcaonline.org.ar

 Mecánica Computacional Vol XXVIII, págs. 529-552 (2009) 551

Ambrosini, R.D. and Luccioni, B. Craters produced by explosions on the soil surface. Journal
of Applied Mechanics, ASME, 73(6):890-900, 2006
Ambrosini, D and Luccioni, B. Craters Produced by Large-Scale Explosions, Mecánica
Computacional, XXVI:1801-1822, 2008
AUTODYN, Explicit Software for Non-Linear Dynamics, Version 11.0, User’s Manual.
Century Dynamics Inc, 2007.
Baker, W.E., Cox, P.A., Westine, P.S., Kulesz, JJ, Strehlow, RA. Explosion hazards and
evaluation. Elsevier; 1983.
Chung Kim Yuen, S., Nurick, G.N., Verster, W. Jacob, N., Vara, A.R., Baldena, V.H. ,
Bwalya, D., Govender, R.A., Pittermann, R.A. Deformation of mild steel plates subjected to
large-scale explosions. International Journal of Impact Engineering, 35:684–703, 2008.
Formby, S.A. and Wharton, R.K. Blast characteristics and TNT equivalence values for some
commercial explosives detonated at ground level. Journal of Hazardous Materials, 50:183-
198, 1996.
Guruprasad, S. and Mukherjee, A. Layered sacrificial claddings under blast loading Part II -
experimental studies. Int J Impact Eng, 24(9):975-984, 2000.
Hanssen, A.G., Enstock, L, Langseth, M. Close-range blast loadin loading of aluminium foam
panels. International Journal of Impact Engineering, 27(6):593–618, 2002.
Hirschfelder, J.O., Littler, D.J., and Sheard, H., Estimated blast pressures from TNT charges
of 2 to 10000 tons, Los Alamos National Laboratory Library, LA Report 316, 3 9338 00350
4049:1-12., 1945.
Jacinto, A.C., Ambrosini, R.D., Danesi, R.F. Experimental and computational analysis of
plates under air blast loading. International Journal of Impact Engineering, 25(10):927–47,
2001.
Kinney, G.F., Graham, K.J. Explosive shocks in air. 2nd ed. Springer Verlag; 1985.
Luccioni, B., Ambrosini D.and Danesi, R. Blast load assessment using Hydrocodes”,
Engineering Structures 28(12), (2006) 1736-1744.
Lee, E.L. and Tarver, C.M. Phenomenological model of shock initiation in heterogeneous
explosives. Physics of Fluids, 23(12):2362-2372, 1980.
Lok, T.S. Response of solid circular ductile cantilever subjected to air-blast, Private
communication, 2005.
Nurick, G.N., Martin, J.B. Deformation of thin plates subjected to impulsive loading—a
review. Part I: Theoretical considerations. International Journal of Impact Engineering,
8(2):159–69, 1989a.
Nurick, G.N., Martin, J.B. Deformation of thin plates subjected to impulsive loading—a
review. Part II: Experimental studies. International Journal of Impact Engineering,
8(2):171–86 1989b.
Nurick, G.N. and Shave, G.C. The deformation and tearing of thin square plates subjected to
impulsive loads-an experimental study. Int J Impact Eng, 18(1):99-116, 1996.
Nurick, G.N., Chung Kim Yuen, S., Jacob, N., Verster, W., Bwalya, D.,Vara, A.R. Response
of quadrangular mild-steel plates to large explosive load. Second international conference
on design analysis of protective structures (DAPS). Nanyang Technological University, 30–
44, 2006.
Smith, P.D., Hetherington, J.G. Blast and ballistic loading of structures. Butterworth-
Heinemann Ltd; 1994.

Copyright © 2009 Asociación Argentina de Mecánica Computacional http://www.amcaonline.org.ar

552 B.M. LUCCIONI, D. AMBROSINI, S. CHUNG KIM YUEN, G.N. NURICK

Teeling-Smith, R.G. and Nurick, G.N., The deformation and tearing of thin circular plates
subjected to impulsive loads, Int J Impact Eng,11(1):77-91, 1991.
Wharton, R.K., Formby, S.A., and Merrifield, R. Airblast TNT equivalence for a range of
commercial blasting explosives. Journal of Hazardous Materials, 79(1-2):31-39, 2000.

Copyright © 2009 Asociación Argentina de Mecánica Computacional http://www.amcaonline.org.ar