Combined Blast And Fragment Impulse – A New Analytical Approach

M.D. Hutchinson1
1
AWE Aldermaston, Reading, Berkshire, RG7 4PR, UK

Abstract
In his 1953 attempt to match an equation to WWII cased charge blast impulse data, E.M. Fisher was obliged to
modify an earlier equation by U. Fano. However, the author has recently published an alternate equation which
is much more clearly derived from the original energy balance equation of R.W. Gurney. This and further
derivations have opened up the field of blast and fragment impulse to new analysis.
The estimation of the combined impulse from blast and casing fragments is key to the design of structures such
as containment vessels and chambers, where the blast wave duration is small compared to the target (i.e.
containment) response time. Not only must the balance between blast gas and casing fragment momentum be
determined, but also their different modes of target interaction must be understood.
Consequently, experiments by Tan and Held on combined blast and casing fragment impulse, using momentum
sleds and blocks, have been reviewed and re-analysed in the light of the author’s recent derivations. It is
concluded that Held’s original analysis, based on Fano’s equation, led to significant under-estimation of the blast
contribution. So, while further such experiments would be useful, Fano’s equation must not be used in the
analysis. As well as replacements for the Fano and Fisher equations, a new equation for combined blast and
fragment impulse has been derived..

Key words : Impulse – Blast – Fragment – Fano – Held

1. Introduction
Our previous papers [1][2] have shown that an equation for blast impulse form a cased
charge, relative to that from the same charge without a casing, can be derived directly from
the original energy balance equation by R.W. Gurney [3]. Earlier derivations, particularly
that by Fano [4], can be set aside.

1.1. Simplified Derivation

A simplified derivation starts with Gurney’s energy balance equation. Following Gurney, the
charge mass is denoted by C, the casing mass by M and its final radial velocity by V:
1 11 
E .C = M .V 2 +  C.V 2  (1)
2 22 
Note that E is the energy per unit charge mass available to do work, which is normally about
70% of the specific heat of detonation, H. If the casing is removed, the maximum velocity Ve,
at the outer surface of the expanding gases, will from (1) be:
Ve = 4 E (2)
This assumes that the radial velocity distribution in the gases is linear, as proposed by Gurney
for gases expanding within a casing. His equation for initial case fragment velocity, which is
the same as the outermost gas velocity, can be put in the form:
2E C
V= = 4 E (3)
2 + C C + 2M
1 M
By taking the ratio of (2) and (3), we find that the ratio of maximum, i.e. outermost, gas
velocities, cased (V) and uncased (Ve), is simply:
V I C
= = (4)
Ve I 0 C + 2M
Within equation (4), we have also equated the ratio of gas velocities with the ratio of bare
charge gas momentum, I0 and cased charge gas momentum, I and therefore the ratio of the gas
impulses delivered.

1.2 Discussion
It is possible to equate the ratio of gas velocities with the ratio of gas momenta, provided that
Gurney’s assumptions of linear radial distribution of gas velocities can be applied both to the
cased gases and to the uncased gases. It is on this basis that equation (4) can be used to
predict relative blast impulse from cased charges. Equation (4) gives similar predictions to
that derived by Fisher [5], but it has been argued in [1] that it has a sounder basis. The
equation by Fano predicts lower blast impulses, which are not in accordance with the data
analysed by Fisher and which could, therefore, lead to erroneous conclusions.
Equation (4) does not, however, include the impulse due to the casing fragments. A means to
include this will now be shown in the context of a comparison with combined blast and
fragment impulse experimental data.

2. Tan and Held Experimental Data

A paper was received by the author in 2010 from Prof. Dr. M. Held, detailing sled momentum
experiments conducted several years earlier by Tan and Held [6].
2.1 Experimental Set-Up
A rectangular steel plate was mounted at the front of a freely moving sled, which was exposed
to the combined impulse form blast and steel case fragments. The charges were 1.2 kg
cylinders of explosive C4 and were fired both bare and with three thicknesses of casing, 2.5,
5.0 and 10 mm, giving 0.9<M/C<4.1. The experimental layout is shown in Figure 1, taken
from Tan and Held’s paper. Charges were end-initiated and mounted with their axes vertical
and such that the casing fragments, projected at the Taylor angle, would strike the sleds.
Figure 1: Plan view layout for Tan and Held experiments from their paper [6]. Two sleds
were fielded per shot, in this example they are shown at 2 m and 3 m.

2.2 Diagnostics
The sled velocities were measured on high-speed video cameras and these gave, when
combined with the sled mass, values for the impulses received. Sleds were placed at 1, 1.5, 2
and 3 m stand-off distances. Table 1 and Figure 2 show the impulses derived by Tan and
Held from the measured sled velocities.
Table 1: (Tan and Held’s Table 2) Measured sled momenta from the combined blast and
fragment loads:
Combined Momenta in kg.m/s
Casing 1.0 m 1.5 m 2.0 m 3.0 m
Thickness, mm
0.0 283 137 62 19
2.5 330 185 113 48
5.0 366 237 145 93
10.0 536 350 258 129

600
Combined momentum, kg-m/s

500 0
2.5
400 5
10
300

200

100

0
0 0.5 1 1.5 2 2.5 3 3.5
Stand-off, m

Figure 2. Data from Table 1, plotted to display combined blast and fragment impulse
reduction with increasing stand-off for the bare charge (M=0) and charges with 2.5, 5.0 and
10 mm thick casings (as indicated in the key).

2.3 Observations
The bare charge data in Table 1 indicate that blast impulse from the bare charge fell away
with distance in approximate accordance with an inverse square law. In contrast, the
fragment impulses fell way more slowly with distance. It was concluded by Tan and Held
that the fragments from the cylindrical casing had remained on nearly parallel paths, so the
number of fragments collected by each plate was in simple inverse proportion to its distance
from the charge, rather than in accordance with an inverse square law.
This meant that fragment impulses dominated most of the measurements, making it difficult
to obtain accurate blast impulse values, since the ratio of blast to fragment impulse was not
only a function of M/C, but also of stand-off from the charge.

2.4 Initial Analysis

When analyzing blast impulse data from cased charges, the key data reference points are the
measured bare charge impulses, relative to which the blast impulse components of the
combined blast and fragment impulses can be estimated. Tan and Held assumed that Fano’s
equation could be used to estimate the blast impulse contributions, with the result that their
estimates of the blast impulse contribution were significantly too low. The blast contributions
to the combined blast and fragment impulses from cased charges have therefore been re-
estimated, for the three case thicknesses studied, using equation (4). These estimates are
shown in Table 2:
Table 2: Calculated blast momenta based on Equation (4 ):
H.E. Momenta in kg.m/s
Casing 1.0 m 1.5 m 2.0 m 3.0 m
Thickness, mm
0.0 283 137 62 19
2.5 168 81 37 11
5.0 129 62 28 9
10.0 93 45 20 6

In order to match the combined impulses in Table 1, we now need to estimate the impulse
contributions from the casing fragments, relative to the bare charge impulses.

3. Calculation of Fragment Impulses

As with the above calculation of blast impulse, we will be seeking to derive the momentum
carried by the mass of the casing divided into fragments, towards a set of targets arrayed
around the charge. The values implied by subtracting the blast momenta in Table 2 from the
experimental combined momenta in Table 1 are shown in Table 3:

Table 3: Fragment momenta based on Tables 1 and 2:

Fragment Momenta in kg.m/s
Casing 1.0 m 1.5 m 2.0 m 3.0 m
Thickness, mm
2.5 162 104 76 37
5.0 237 175 117 84
10.0 443 305 238 123

Since the fragment impulses dominated many of the sled impulses measured, to demonstrate a
predictive capability, we must attempt to match these values, using equation (4) to estimate
the fragment impulses, based on the bare charge impulses and the casing/charge mass ratio,
M/C.
 3.1 Fragment Impulse Derivation
In a cylindrical system, with a linear radial dependency of gas velocity, the combined scalar
radial momentum Q of both gases and fragments will, by integration, be:
 2 
Q =  M + C V (5)
 3 
Therefore, based on the measured bare charge impulse at 1 m, treated as I0, the varying case
fragment impulse contributions IM can be estimated as a fraction of I0. Taking from equation
(5) the ratio of casing and cased gas momenta, and applying it to equation (4), we can derive
the casing momentum IM as a fraction of the bare gases momentum, I0:

IM 3 M  C 
=   (6)
I0 2 C  C + 2M 
The total impulse would therefore be, from (4) and (6):

IC + I M  3 M   C 
= 1 +    (7)
I0  2 C   C + 2M 
In deriving equations (6) and (7) from equation (5), it has been assumed that both blast and
fragments interact with the targets in the same manner. It will be shown in the next sub-
section that this assumption is not quite consistent with the experimental data.

3.2 Fragment Impulses for the Tan and Held Example

The estimated case fragment momenta for the Tan and Held experiments are shown in Table 3
above. In attempting to match the momenta in Table 3, using equation (7), it has been
necessary, in calculating the values in Table 4, to adjust the coefficient of M/C in equation (7)
downwards, from 3/2 to 3.0/2.9, as in equation (8).

I C + I M  3.0 M   C 
= 1 +    (8)
I0  2.9 C   C + 2M 
An argument can be made for this adjustment, based on the fact that, while blast waves reflect
quite elastically from solid surfaces, casing fragments tend to penetrate, making their impacts
more plastic. The coefficient 3/2 in (7) springs from the linear radial distribution of gas
velocities, but we propose that this has to be multiplied by 1.0/1.45, to represent the ratio of
fragment and gas reflection coefficients. This is further discussed in the next sub-section.
Table 4 Calculated fragment momenta, based on Equation (7), but with adjustment of the
coefficient of M/C in (7) to 3.0/2.9, as in equation (8):
Fragment Momenta in kg.m/s
Casing 1.0 m 1.5 m 2.0 m 3.0 m
Thickness, mm
2.5 160 107 80 27
5.0 256 171 128 43
10.0 398 265 199 66
The calculated fragment impulses in Table 4 are near to those estimated in Table 3, but only
for lower stand-off distances and lower casing masses. Part of the explanation could be that,
given the stochastic nature of natural fragmentation, there was significant scatter in the mass
of fragments stopped by each sled, especially for the thicker casings and larger sled stand-off
distances.

3.3 Consideration of Reflection Coefficients

The lower value (3.0/2.9) of the coefficient in equation (8) would be consistent with a higher
reflection coefficient, e.g. 1.8, for blast waves undergoing nearly elastic collisions with their
targets and a much lower reflection coefficient, e.g. 1.24 (=1.8/1.45), for casing fragments,
which at high impact velocities undergo mostly plastic collisions with their targets. The
fragments from the 10mm casings will, however, have impacted at lower velocity. This
would have allowed them to impact more elastically, and could accounted for the relevant
impulses in Table 3 being significantly higher than those calculated in Table 4.
A reliable figure for fragment reflection will be necessary to obtain good predictions in future
and will need to be confirmed by further theory and/or experimental studies. It will certainly
depend on all the factors affecting the elasticity of the fragment impact, such as its incoming
velocity and pitch, its own yield stress and that of the target.

3.4 Combined Impulses Calculated for the Tan and Held Example
It is now possible to add the predicted blast and fragment momenta for the cased charges at
the four distances, as shown in Tables 2 and 4, with the intention of comparing these with Tan
and Held’s measured values in Table 1.
Table 5: Calculated combined blast and fragment momenta, based on Equations (5) and (7):
Combined Momenta in kg.m/s
Casing 1.0 m 1.5 m 2.0 m 3.0 m
Thickness, mm
2.5 328 188 117 38
5.0 384 233 156 51
10.0 491 310 219 73
 Case thickness 2.5mm Case thickness 10mm
350 600
Case Fragments Case Fragments
Predicted momentum, kg-m/s

Predicted momentum, kg-m/s

300
Gases 500 Gases
250 Combined
Combined 400
200
300
150
200
100

50 100

0 0
0 1 2 3 4 0 1 2 3 4
Stand-off, m Stand-off, m

Figure 3: Showing the predictions, normalized to the experimental bare charge impulses, for
the casing and explosive impulses with the smallest and largest casing thicknesses.
In Figure 3, we can see how the predicted impulses from the casing fragments come to
dominate as the casing thickness is increased. However, this predicted dominance is not as
great as if the Fano equation for blast impulse had been used.
Table 6: Showing the ratios between blast impulse data values obtained by the above
reanalysis of Tan and Held’s combined impulse data.
Ratios of measured and predicted cased
charge momenta
Casing 1.0 m 1.5 m 2.0 m 3.0 m
Thickness, mm
2.5 0.99 1.02 1.03 0.79
5.0 1.05 0.98 1.08 0.55
10.0 0.92 0.89 0.85 0.56
In Table 6, we can now see directly how the impulses predicted in this paper compare with
those measured by Tan and Held. Fair agreement can be seen at up to 2 m stand-off, but
beyond this the predictions do not reach the values of the measured impulses. As expected,
the agreement is poorer for the thicker casings and larger stand-offs.

4. Conclusions
A new equation (8) has been derived to provide an estimate of the combined case fragment
and blast impulse to a target, compared to the impulse I0 from the same charge without a
casing. For most situations, where the fragments impact the target more or less plastically,
equation (8) will need to be modified, e.g. as in equation (9). Thus, useful predictions could
have been made for the sled momenta measured by Tan and Held at up to 2 m stand-off,
based on the measured values for I0 at 1 m, provided that the above ratio of blast and fragment
reflection coefficients had been estimated.
A degree of understanding has been reached, regarding the physics behind Tan and Held’s
data, using the approach above. A further investigation, somehow measuring the fragment
impulses separately, could provide actual values for the fragment reflection coefficient
required for the given conditions. One possibility would be to use a gun to fire fragment
simulating projectiles into a target and measure its velocity.
Given the complexity of interpretation required, the Tan and Held sled method is not easily
applicable to a broad parameter study of blast and fragment impulse. However, it could be a
valuable method for confirming the combined blast and fragment impulses to be expected
under a specific set of relevant conditions.

ACKNOWLEDGEMENTS
The kindness of the late Prof. Dr. M. Held in offering the above data for reanalysis is much
appreciated. A first draft of the above analysis was discussed with Held at a meeting in July
2010 and it was argued by him that the author should seek publication, in Propellants,
Pyrotechnics, Explosives, of his alternate equation to those of Fano and Fisher. The relevant
papers [1][2] have since been kindly accepted by the journal.
This work was sponsored by the UK Ministry of Defence.

REFERENCES
[1] Hutchinson, M.D. (2009) The Escape of Blast From Fragmenting Munitions Casings Int. J. Impact
Eng. Volume 36, pp185-192.
[2] Hutchinson, M.D. (2011) Replacing the Equations of Fano and Fisher for Cased Charge Blast
Equivalence –I Ductile Casings, Propellants, Explosives, Pyrotechnics, Volume 36, pp310-313.
[3] Gurney, R.W. (1943) The Initial Velocities of Fragments from Bombs, Shells and Grenades, Army
Ballistic Research Laboratory, Report BRL 405, Aberdeen Proving Ground, Maryland USA.
[4] Fano, U. (1944) Methods for Computing Data on the Terminal Ballistics of Bombs - II Estimation
of the Air Blast, Army Ballistic Research Laboratory, Report BRL 524, Aberdeen Proving Ground,
Maryland USA.
[5] Fisher, E.M., (1953) The Effect of the Steel Case on the Air Blast from High Explosives, Naval
Ordnance Laboratory, White Oak MD, USA.
[6] Held, M. & Tan, G.E.B., (2003) Radial Blast Loads of Confined Cylindrical Charges, International
Symposium “Interaktion der Wirkung von Munitionen mit Bauwerken”.