Shock Waves (2011) 21:111–119

DOI 10.1007/s00193-011-0308-0

ORIGINAL ARTICLE

Prediction of clearing effects in far-field blast loading of finite

targets
A. Tyas · J. A. Warren · T. Bennett · S. Fay

Received: 14 July 2010 / Revised: 7 September 2010 / Accepted: 22 October 2010 / Published online: 26 February 2011
© Springer-Verlag 2011

Abstract It is well known that when a blast wave strikes the List of symbols
face of a target, the duration of the loading, and hence the total a Sonic velocity in air
impulse imparted to the target may be influenced by the prop- a0 Sonic velocity in air at ambient conditions
agation of a rarefaction, or “clearing” wave along the loaded CD Drag coefficient
face of the target adjacent to free edges. Simple methods of ii Incident specific impulse (impulse per unit area)
predicting the effect of clearing on reducing the blast loading ir Reflected specific impulse (impulse per unit area)
impulse have been available for many years, but recent stud- pi Incident (side-on) overpressure
ies have questioned the accuracy and physical basis of these pr Reflected overpressure
approaches. Consequently, several authors have used numer- q Dynamic pressure
ical modelling and/or experimental techniques to determine S Stand-off (distance from detonation point to target)
empirical predictive methods for the clearing effect. In fact, t Time
the problem had been addressed more than 50 years ago in ta Arrival time of blast wave (t = 0 is time of detonation)
a study which appears to have been since overlooked by the tc Clearing time
blast research fraternity. This article presents the results of td Positive duration of blast wave
that earlier study, and provides experimental validation. The U Velocity of incident shock
analytical predictions are very simple to determine, and are u Particle velocity of air behind shock front
shown to be in excellent agreement with experimental results. W Mass of explosive charge
x Shortest distance from a point on a target to
Keywords Blast loading · Clearing · Diffraction · nearest free edge 
Rarefaction Z Scaled distance S/W 1/3
η Hudson’s clearing length scale
λ Spatial length of positive phase of blast wave
θ Decay coefficient of blast wave

1 Introduction

Communicated by C. Needham. The nature of the loading which results when a shock wave
A. Tyas (B) · J. A. Warren · T. Bennett
from the detonation of a high explosive in air arrives at
Department of Civil and Structural Engineering, and interacts with a target structure has been the subject of
University of Sheffield, Mappin Street, Sheffield S1 3JD, UK experimental, numerical and theoretical research for many
e-mail: [email protected] decades. For geometrically simple scenarios, such as a spher-
ical explosive charge detonated in free air, the key parameters
S. Fay
Blastech Ltd., The Bioincubator, 40 Leavygreave Road, of the loading are well established for many common types
Sheffield S3 7RD, UK of explosive. Theoretical predictions for such scenarios were

123
112 A. Tyas et al.

developed many decades ago for cases where the distance

from the detonation to the target is relatively large and the
shock can be assumed to be weak. Non-linearities associ-
ated with stronger shocks preclude the development of sim-

Overpressure (p)
ple closed form solutions of the shock equations, and in
such cases, numerical solutions or experimental results are p
MAX i
required. Kingery and Bulmash [1] produced a database of
experimental records of blast loading parameters from a wide
range of explosive charge masses and stand-offs, and this
database formed the core of the widely used blast loading ta td Time from detonation (t)
predictive tool ConWep [2].
A blast wave travelling in free air has the characteris-
tic “Friedlander” form, comprising a sudden step change in
Fig. 1 Friedlander blast wave
pressure, followed by an exponential decay and subsequent
negative phase due to the over-expansion of the air. This
is shown schematically in Fig. 1. The positive phase of the an early reduction in reflected pressure, over and above the
Friedlander curve is given by: normal Friedlander decay and thus reducing both the pos-
  itive duration and the magnitude of the reflected impulse.
t − ta − t−ta
p (t) = pmax 1 − e θ (1) The time between impingement of the shock and the onset
td
of this “clearing” effect of the rarefaction wave will obvi-
As the blast wave propagates through free air, or in a case ously increase with the distance between a point of interest
where the wave travels parallel to a rigid surface, the wave is on a target and the nearest free edge. Since the magnitude
called an incident wave, and the pressure and specific impulse of the reflected impulse is frequently a key parameter, which
(impulse per unit area) of the Friedlander function are given influences structural response, it is clearly of importance to
the subscript “i” (or sometimes “s” for “side-on”). If the blast be able to predict the effect of the rarefaction wave.
wave impinges normally on a rigid target, it reflects back Diffraction of a shock wave travelling parallel to a rigid
towards the detonation point, resulting in an amplification of surface and meeting a free edge has been extensively studied
the pressure (and consequent increase in specific impulse), for many decades, but theoretical studies of the case where
and the parameters of the consequent reflected wave are given the shock wave reflects from a finite length surface normal to
the subscript “r”. In weak shock conditions, the maximum direction of travel of the shock front are rare in peer-reviewed
pressure associated with the incident wave is low (typically literature. A notable exception to this is the study of weak
<1bar = 100 kPa); in this situation, the magnitude of the shock diffraction by Hunter and Keller [3]. Whilst this anal-
reflected pressure is approximately twice that of the inci- ysis gives a useful qualitative framework for addressing our
dent wave, and both the incident shock velocity (U ) and current problem it assumes that the conditions behind the
the sonic velocity in the shocked air (a) is not significantly incident shock front are constant, rather than decaying with
greater than that at ambient conditions (a0 ). ConWep pro- time as in the Friedlander blast wave and consequently the
vides predictions for the arrival time (ta ), positive duration results are not immediately applicable. Perhaps as a result of
(td ), peak overpressure ( pmax(i,r) ), decay coefficient (θi,r ), the paucity of theoretical analysis, published studies of the
positive impulse (i i,r ), shock velocity (U ) and peak particle problem at hand, and related problems have tended to focus
velocity in the air behind the shock (u) for several common on numerical and/or experimental approaches e.g. [4–10].
high explosive materials for both free field detonations of Earlier empirical predictions of the clearing effect were
spherical explosive charges, and groundburst detonation of presented by Kinney and Graham [11], and the US Tri-
hemispherical charges. Services Manuals TM5-1300[12] and TM5-855-1[13]. These
The Kingery and Bulmash/ConWep predictions for approaches aim to find the total impulse experienced as a blast
reflected pressure and impulse are based on the assumption wave interacts with a target of finite frontal area; all assume
that the target surface is effectively infinite, which in practice that clearing begins immediately on the arrival of the blast
means that td is sufficiently small in magnitude for the entire wave t = ta , and is fully completed at t = ta , +tc , where
positive loading phase to be completed before the blast load- tc is a time delay known as the “clearing time”. The aver-
ing at a point of interest on the face of the target is altered age pressure on the target is assumed to fall linearly from
by the effect of diffraction of the blast wave around free the peak reflected pressure ( pr,max ) at t = ta to the sum of
edges of the target. Where this is not the case, we might the incident pressure and the drag pressure pi (t) + CD q(t))
expect that this diffraction will result in the propagation of when clearing is completed at t = ta , +tc , decaying linearly
rarefaction waves across the face of the target, leading to thereafter to zero at time t = ta , +td .

123
Prediction of clearing effects in far-field blast loading of finite targets 113

As recently discussed by Rickman and Murrell [6], there tions developed to predict clearing effects. This follows the
are some inconsistencies between these approaches of pre- same principle as the work reported by Rose and Smith [5]
vious authorities in the determination of tc and none of the where the ConWep reflected impulse for an infinite target
methods appears to have a strong physical basis. Further- was reduced by “clearing factors” derived from numerical
more, since the approaches aim to find the average cleared analyses.
impulse, they cannot be used to determine the spatial var- Interestingly, an alternative theoretical method was set out
iation of specific impulse on the target. For example, the in the 1950s, in an apparently little-known publication by
assumption that the effect of clearing begins immediately on Hudson [14]. The rest of this article gives a brief explanation
the shock arrival is valid if one considers the total load on the of the method, sets out the results and provides experimental
target face, since clearing will begin instantaneously at the validation.
free edges of the target. But this assumption cannot be valid
if one considers the local pressure at a point remote from a
free edge; some finite transit time is required before clearing
can commence. 2 Hudson’s acoustic pulse method
ConWep includes a subroutine, the “Loads on Structures”
(LoS) module, which purports to allow for the effect of clear- In a Sandia Corporation Technical Memorandum written by
ing on plane targets normal to a radius from the point of deto- Hudson in 1955 [14], work was presented which attempted
nation of a free air or groundburst explosion and to predict the to provide a robust method to predict the clearing effects seen
spatial distribution of specific impulse. The methods used in when blast waves from nuclear explosion trials impinged on
ConWep to calculate the clearing time and the magnitude of measurement blocks instrumented with blast pressure
the clearing effect are not clear but a brief analysis of typical gauges. The report was not de-classified until 1998, and per-
ConWep predictions is sufficient to cast doubt on their effi- haps for this reason, appears to have been overlooked by the
cacy. For example, consider the case of the blast wave from a general research community; it appears to have been cited
2,000 kg TNT hemispherical groundburst explosion imping- by only one other publication, itself another Sandia techni-
ing normally (i.e. as an approximately plane wave front) on cal memorandum in the 1950s.
a vertical target, 14 m wide and 14 m high at a stand-off Hudson’s report considers the case of a plane, weak
of 200 m. Initially, the entire front face of the target will be Friedlander blast wave impinging normally onto a flat tar-
loaded by a pressure pulse which will begin to decay as shown get of finite dimensions, and uses the Sommerfeld diffraction
in Fig. 1. Simultaneous with the impingement of the blast theory [15] to develop spatial and temporal predictions of the
wave, a rarefaction wave will begin to propagate inwards magnitude of the rarefaction waves as they propagate along
from the free edges at the local sonic velocity in the stagnant the face of the target. Hudson makes four assumptions which
shocked air. As a lower bound, we assume this propagation facilitate his analysis:
velocity to be that of sound in ambient air (say 340 m/s).
This indicates that the clearing waves from the two vertical
• The radius of curvature of the shock front is sufficient
edges of the target should have reached the vertical centreline
large for the assumption that the shock front is plane to
of the target about 20 ms after the arrival of the blast wave.
be reasonable.
Since the positive duration of the blast wave for this event
is ∼70 ms, clearing ought to lead to a reduction from the
full reflected impulse at every point of the target. However,
ConWep LoS predicts that a central rectangular area approxi- 0.2
mately 4 m wide and extending from ground level to a height
Hudson Rarefaction Function/p i,max

0.1
of 9 m experiences the same reflected impulse as it would
0
were the target infinite in extent, implying that the clearing 0 0.5 1 1.5 2 2.5 Time (t/t d) 3
wave is assumed to have propagated just 5 m in from each -0.1

free edge by the end of the positive loading phase of the blast -0.2
wave. This raises serious questions as to the accuracy of the -0.3
ConWep LoS approach to blast clearing, and suggests that η = 0.001 η=0.005 η=0.03 η=0.1
-0.4
this approach may significantly under-estimate the effect of
η=0.3 η=0.5 η=0.7 η=1.0
clearing and thereby over-estimate the actual impulse expe- -0.5

rienced by a target. -0.6

This issue has been discussed in detail by Rickman and -0.7
Murrell [6], who present the results of a large number of
well-controlled experiments, together with empirical rela- Fig. 2 Hudson rarefaction functions

123
114 A. Tyas et al.

Fig. 3 Test arrangement (a)

General arrangement of test (b)
(a)
arena (b) Target face

of the reflected pressure pulse for an infinite target and the

relevant Hudson rarefaction wave. Hudson gives some exam-
ples of the use of this approach compared with experimental
traces from nuclear explosions; whilst these show reasonable
qualitative correlation, the experimental records were not of
sufficiently high resolution to make overall quantitative com-
parisons. As a result, a number of tests were conducted at the
University of Sheffield Blast & Impact Laboratory, Buxton,
UK in order to check the accuracy of the Hudson predictions.

Fig. 4 Pressure gauge locations in the target face 3 Experimental work

A number of simple explosive tests were conducted, each

• The dimension of the target structure in the direction of of which used 250 g C4 hemispherical explosive charges,
propagation of the blast wave is sufficiently long for rare- detonated on a 100 mm square by 50 mm thick steel plate
faction waves from the rear face to be unable to influence placed on a level, flat concrete ground slab. The steel plate
the front face loading in the timescale of interest. was used to reduce damage to the concrete slab. Whilst this
• The rarefaction wave propagates into stagnant shocked does not strictly produce the idealized ground-burst condi-
air. tions assumed in the ConWep predictions, it is considered
• The shock is sufficiently weak for sonic velocity behind that the discrepancy is negligible. This is discussed in Sect. 4
the shock front to be approximately equal to the velocity below. The target was a large reinforced concrete block,
of the incident shock. placed on the same ground slab and clad with 20 mm thick
mild steel plate to give mounting points for pressure gauges
Whilst neither a detailed presentation of the resulting equa- and provide flat, square target faces. The side of the target fac-
tions, nor tabulated numerical values of their solutions are ing the explosive charge was 675 mm high by 710 mm wide
presented in [14], Hudson does present graphical results and the length of the block on a line away from the charge
(Fig. 2 of [14]) for the amplitude of the clearing wave pres- was 1,930 mm (thus satisfying one of the assumptions of the
sure as a function of η = x/λ, (the distance between a point Hudson model, namely that the target be sufficiently deep
on the target and a free edge divided by the spatial length that no rarefaction wave from the back face would have time
of the positive phase of the incident Friedlander pulse) and to affect the front face loading). The overall test set-up is
t/td .1 Figure 2 of this paper shows examples of these rarefac- shown in Fig. 3.
tion waves, normalised against pi,max for selected values of η. Two pressure gauges (Kulite HKM 7 bar) were placed in
The rarefaction waves clearly become more rounded, lower in the front plate, one (G1) at the centre of the face, the other
amplitude and longer in duration as η increases (which itself (G2) vertically above it and 25% of the height of the face
implies a greater distance of travel from the free edge(s)). below the top edge. The gauge locations are shown in Fig. 4.
The time history of the cleared pressure at any point on Tests were conducted at four stand-offs, 4, 6, 8 and 10 m from
a given target surface can be found from the superposition the target face. Assuming a TNT equivalence value of 1.20
for C4 explosive, the stand-offs equate to scaled distances
1 NB: Hudson’s time parameter, the abscissa of Fig. 2 in [14] is actu-
of 6.0–14.9 m/kg1/3 , with the scaled target face dimensions
ally presented as a normalised length scale, but can be shown to be being approximately 1 m/kg1/3 . A total of eight tests were
effectively t/td . conducted, two at each stand-off. The explosive charges were

123
Prediction of clearing effects in far-field blast loading of finite targets 115

25 30
Experiment Gauge 1 Experimental Gauge 1

Hudson Prediction Gauge 1 25 Hudson Prediction Gauge 1

20
ConWep Incident Pressure ConWep Incident Pressure
20
Overpressure (kPa)

Overpressure (kPa)
15 ConWep Reflected Pressure ConWep Reflected Pressure

ConWep Incident + Dynamic 15 ConWep Incident + Dynamic

Pressure Pressure
10
10

5
5

0
0
23 23.5 24 24.5 25 25.5 26 26.5 27
17.5 18 18.5 19 19.5 20 20.5 21
-5 -5
Time after detonation (ms) Time after detonation (ms)
S=10m S=8m
50 100
Experiment Gauge 1 Experiment Gauge 1

40 Hudson Prediction Gauge 1 80 Hudson Prediction Gauge 1

ConWep Incident Pressure

Overpressure (kPa)

Overpressure (kPa)
ConWep Incident Pressure
30 60
ConWep Reflected Pressure
ConWep Reflected Pressure
ConWep Incident + Dynamic
20 ConWep Incident + Dynamic 40 Pressure
Pressure

10 20

0 0
12 12.5 13 13.5 14 14.5 15 15.5 16 7 7.5 8 8.5 9 9.5 10
-10 -20
Time after detonation (ms) Time after detonation (ms)
S=6m S=4m
Fig. 5 Pressure–time signals, Gauge G1

detonated using electrically activated L2A1 detonators, and tion was used to enable Hudson to produce solutions of the
the pressure signals were recorded using a LeCroy Wave- acoustic problem, and whilst his description does not make
Jet324 digital oscilloscope with a sample rate of 100 kHz. it absolutely clear, it appears that he also uses this to give the
The oscilloscope was triggered by the failure of a break-wire simplification
wrapped around the detonator, to synchronize the pressure
records with the time of break-out of the detonator. λ ≈ a 0 td (3)
The Hudson clearing method was used to derive predic-
tions of the cleared-reflected loading, taking ConWep pre- This assumption, with a0 = 340 m/s and td taken from Con-
dictions for the basic reflected pressure on an infinite target. Wep has been used in the current work to determine λ for
In [14], there is some lack of clarity in the way in which each stand-off, and hence the values of η at the gauge loca-
the spatial length of the positive phase of the incident blast tions. The relevant Hudson rarefaction waves are then found
pressure wave, λ, is determined. The length of the wave is: by interpolating between the different η curves in Fig. 2.
In the experimental work described here, three rarefaction
a +td
t waves propagate across the target face: one from each free
λ= U (t) dt (2) vertical edge and one from the free top edge. Whilst Hudson
ta
makes clear the fact that his analysis does not account for
interactions of the rarefaction waves, in the current work the
where, U , the velocity of propagation of the incident wave three rarefaction functions have simply been superimposed.
decreases as the magnitude of the incident pressure decays Figures 5 and 6 show examples of pressure–time histories
along the Friedlander curve. The limitation of his method for the four stand-offs, with experimental data from the two
to weak shocks allows Hudson to assume that U ≈ a ≈ gauges together with ConWep predictions for the pr (t), pi (t)
a0 , where a is the sonic velocity behind the reflected shock and pi (t) + CD q(t) (ignoring clearing and assuming CD = 1
and a0 is the ambient sonic velocity in air. This assump- as indicated in [13] for front wall loading) and the ConWep

123
116 A. Tyas et al.

25 30
Experiment Gauge 2 Experimental Gauge 2

Hudson Prediction Gauge 2 25 Hudson Prediction Gauge 2

20
ConWep Incident Pressure ConWep Incident Pressure
Overpressure (kPa)

20

Overpressure (kPa)
15 ConWep Reflected Pressure
ConWep Reflected Pressure
15 ConWep Incident + Dynamic
ConWep Incident + Dynamic
Pressure
10 Pressure
10

5
5

0
0
23 23.5 24 24.5 25 25.5 26 26.5 27
17.5 18 18.5 19 19.5 20 20.5 21
-5 -5
Time after detonation (ms) Time after detonation (ms)
S=10m S=8m
50 100
Experiment Gauge 2 Experiment Gauge 2

40 Hudson Prediction Gauge 2 80 Hudson Prediction Gauge 2

ConWep Incident Pressure

Overpressure (kPa)

Overpressure (kPa)
ConWep Incident Pressure
30 60
ConWep Reflected Pressure ConWep Reflected Pressure

20 ConWep Incident + Dynamic 40 ConWep Incident + Dynamic

Pressure Pressure

10 20

0 0
12 12.5 13 13.5 14 14.5 15 15.5 16 7 7.5 8 8.5 9 9.5 10
-10 -20
Time after detonation (ms) Time after detonation (ms)
S=6m S=4m
Fig. 6 Pressure–time signals, Gauge G2

prediction for pr (t) corrected by the Hudson rarefaction func- and observed experimentally for the arrival of the rarefaction
tions calculated as described above. In each case, the time waves at the gauge points. In the case of gauge G1 (Fig. 5)
datum of the experimental signals was shifted so that the where the gauge is almost equidistant from the two side
arrival time of the shock at G1 corresponded to that pre- edges and the top edge, the three rarefaction waves arrive
dicted by ConWep; in no case was it necessary to time-shift nearly simultaneously and their superposition results in a
the experimental data by more than 1.7% of the arrival time pronounced and rapid drop in pressure around 1 ms after
to achieve this correlation. the arrival of the blast wave. For gauge G2 (Fig. 6) the rar-
Figure 7 shows the examples of the cumulative impulse efaction wave from the top edge arrives much earlier than
against time traces for shots at each of the four stand-offs. those from the sides, and two separate drops in pressure are
apparent, around 0.5 and 1 ms after the arrival of the blast
wave.
4 Discussion After the arrival of all three rarefaction waves, the
pressure–time histories at the two gauge points are very sim-
Inspection of the graphs in Figs. 5, 6 and 7 reveals a number ilar, and the magnitude of overpressure quickly becomes sig-
of key aspects shared by both the experimental traces and the nificantly lower than the pi (t) + CD q(t) value assumed in
Hudson predictions. TM5-1300, and indeed becomes negative well before the end
Before clearing commences, the experimental results are of the incident wave. This early onset of negative pressure
very similar to the ConWep predictions for the reflected pres- is entirely due to the superposition of the various rarefaction
sures. This gives credence to the supposition that the 50 mm waves; it occurs much too early to be related to the well-
steel ground plate did not materially affect the assumption of known negative phase which occurs at the end of the positive
idealized groundburst conditions. In general, there is excel- phase of the blast wave due to over-expansion of the propa-
lent consistency in the times predicted by the Hudson method gating medium.

123
Prediction of clearing effects in far-field blast loading of finite targets 117

Specific Reflected-Cleared Impulse 20 25

18

Specific Reflected-Cleared Impulse

16 20
14
12 15
(kPa.ms)

(kPa.ms)
10
Experiment Gauge 1
Experiment Gauge 1
8 10
Experiment Gauge 2
6 Experiment Gauge 2
Hudson Prediction Gauge 1
4 Hudson Prediction Gauge 1 5
Hudson Prediction Gauge 2
2 Hudson Prediction Gauge 2
0 0
23 23.5 24 24.5 25 25.5 26 26.5 27 17.5 18 18.5 19 19.5 20 20.5 21
Time (ms) Time (ms)
S=10m S=8m
35 70

Specific Reflected-Cleared Impulse

Specific Reflected-Cleared Impulse

30 60

25 50

(kPa.ms)
(kPa.ms)

20 40
Experiment Gauge 1
15 30
Experiment Gauge 2 Experiment Gauge 1

10 Hudson Prediction Gauge 1 20 Experiment Gauge 2

Hudson Prediction Gauge 1

5 Hudson Prediction Gauge 2 10
Hudson Prediction Gauge 2
0 0
12 12.5 13 13.5 14 14.5 15 15.5 16 7 7.5 8 8.5 9 9.5 10
Time (ms) Time (ms)
S=6m S=4m
Fig. 7 Cumulative impulse versus time

At the longer stand-offs, correlation between the experi- est at 8 m. The ConWep predictions are 27.4–56.8% higher
mental results and the predicted traces appears to be excellent than the experimental average at G1, and 22.2–34.8% higher
in the early stages of clearing, although the Hudson method at G2, with the discrepancies at both gauge positions rising
appears to slightly over-predict the later effect of clearing, monotonically with stand-off.
most especially the rate of development of the negative pres- Rose and Smith’s numerical and experimental study of
sure. This is possibly due to the fact that the rarefaction waves clearing [5] concluded that at short scaled distances, and/or
are not propagating into stagnant air by this time. with large scaled target dimensions, clearing was insignifi-
At 4 m stand-off, the correlation is slightly less good, with cant, with the average specific impulse over the target face
the Hudson method over-estimating the clearing effect from tending towards the value for an infinite reflecting surface.
the outset. This is thought to be primarily due to the decreas- For the inverse scenario, they noted that clearing was highly
ing validity of the assumption of a plane incident shock (at important and the impulse tended towards the incident value;
4 m stand-off, the radius from the detonation point to the that is, the target presented no significant obstacle to the free
top of the target is inclined at ∼ 9.6◦ to the horizontal). The passage of the blast wave. This finding is reinforced by the
resulting, slightly oblique shock will produce an air flow par- data shown in Fig. 9, where values of average specific impulse
allel to the target surface, retarding the propagation velocity on the target in this study have been calculated from ConWep
of the rarefaction waves. LoS, and by the Hudson method (by calculating the Hudson
Comparison of the experimental and Hudson prediction cleared specific impulses at each point on a 100 × 100 grid
results with the relevant ConWep LoS predictions for total on the target face and numerically integrating the results to
positive impulse is illuminating (Fig. 8). The Hudson predic- determine the average value). The Hudson predictions clearly
tions are between 1.25–3.10% lower than the average experi- show a migration from values tending towards the ConWep
mental impulses at G1, and 0.86–4.86% lower at G2. In both incident impulse at long stand-off, to values tending towards
cases, the largest discrepancy is at 4 m stand-off and the low- the ConWep reflected impulse for an infinite target at shorter

123
118 A. Tyas et al.

(a) 80
Total Postive Specific Impulse (kPa.ms)
5 Conclusions
70
Hudson Prediction
Analytical predictions of the effect of clearing of blast waves
60
Experiment on a finite sized target developed more than 50 years ago have
50 Conwep LoS been presented and validated by carefully controlled exper-
40 imental results. The experimental data has indicated that,
within the range of scaled distances considered, the predic-
30
tions of the widely used ConWep tool for arrival time and
20
early-stage reflected blast loading are very accurate. How-
10 ever, the ConWep LoS module appears to significantly over-
0 predict the cleared impulse on a finite sized target, whereas
0 2 4 6 8 10 12 the predictions from the Hudson acoustic method have been
Stand-off (m)
shown to demonstrate excellent correlation with carefully
(b) controlled experimental results. It is therefore recommended
80 that the Hudson method is used for predicting both local
Peak Reflected Specific Impulse (kPa.ms)

70 specific impulse and total impulse on a target. The Hudson

method is independent of scale, and may, in principle, be
60 Hudson Prediction
Experiment used to predict the clearing-related reduction in both positive
50 Conwep LoS duration and reflected impulse experienced by a target in
40 larger scale events, providing that the assumptions of rela-
30
tively weak shock and normal incidence of the blast wave are
acceptable.
20

10

0
0 2 4 6 8 10 12 References
Stand-off (m)
1. Kingery, C.N., Bulmash, G.: Airblast parameters from TNT spher-
Fig. 8 Positive impulse versus stand-off a Gauge 1, b Gauge 2 ical air burst and hemispherical surface burst, Report ARBL-TR-
02555, U.S. Army BRL, Aberdeen Proving Ground, MD (1984)
2. Hyde, D.W.: Conventional Weapons Program (ConWep). US
Army Waterways Experimental Station, Vicksburg, USA (1991)
1000
Hudson Prediction
3. Hunter, J.K., Keller, J.B.: Weak shock diffraction. Wave Motion
Average Reflected Specific Impulse

6, 79–89 (1984)
Conwep LoS 4. Smith, P.D., Rose, T.A., Saotonglang, E.: Clearing of blast waves
ConWep Reflected (infinite from building facades. Proc. Instn Civ. Eng. Struct. Bldg. 134,
target) 193–199 (1999)
ConWep Incident
(kPa.ms)

5. Rose, T.A., Smith, P.D.: An approach to the problem of blast

100 wave clearing on finite structures using empirical procedures based
on numerical calculations. In: 16th Symposium on the Military
Aspects of Blast and Shock (MABS16), Oxford, UK (2000)
6. Rickman, D.D., Murrell, D.W.: Development of an improved meth-
odology for predicting airblast pressure relief on a directly loaded
wall. ASME J. Press. Vessel Technol. 129, 195–204 (2007)
10 7. Rose, T.A., Smith, P.D., May, J.H.: The interaction of oblique blast
1 10
waves with buildings. Shock Waves 16, 35–44 (2006)
Stand-off (m) 8. Shi, Y., Hao, H., Zhong-Xian, L.: Numerical simulation of blast
wave interaction with structure columns. Shock Waves 17, 113–133
Fig. 9 Average predicted impulse on target face versus stand-off (2007)
9. Trélat, S., Sochet, I., Autrusson, B., Loiseau, O., Cheval, K.: Strong
explosion near a parallelepipedic structure. Shock Waves 16,
349–357 (2007)
10. Ballantyne, G.J., Whittaker, A.S., Dargush, G.F., Aref, A.J.: Air-
stand-offs. These latter predictions should be treated with blast effects on structural shapes of finite width. J. Struct. Eng.
some caution, since at short stand-offs the assumption of 136(2), 152–159 (2010)
weak shock conditions and normal incidence of the shock 11. Kinney, G.F., Graham, K.J.: Explosive Shocks in Air. Springer,
Berlin (1985)
at all points on the target become increasingly inaccurate. 12. Anon.: Dept. of Army—Structures to resist the effects of acciden-
Nevertheless, the general trends in the Hudson predictions tal explosions, Technical Manual TM 5-1300. Department of the
are logical. Army, Washington, DC (1990)

123
Prediction of clearing effects in far-field blast loading of finite targets 119

13. Anon.: Design and analysis of hardened structures to conventional Memorandum SC-TM-191-55-51 (1955). Available from
weapons effects TM 5-855-1. Departments of the Army, Navy, Air https://www.osti.gov/opennet/servlets/purl/16340913-1cPgoX/16
Force, and the Defense Special Weapons Agency, Washington DC, 340913.pdf. Accessed 07 September 2010
(1998) 15. Sommerfeld, A.J.W.: Mathematische theorie der diffraction. Math.
14. Hudson, C.C.: Sound pulse approximations to blast loading Ann. 47, 317–374 (1896)
(with comments on transient drag), Sandia Corporation Technical

123