The Formation of a Blast Wave by a Very Intense Explosion. I.

Theoretical Discussion
Author(s): Geoffrey Taylor
Reviewed work(s):
Source: Proceedings of the Royal Society of London. Series A, Mathematical and Physical
Sciences, Vol. 201, No. 1065 (Mar. 22, 1950), pp. 159-174
Published by: The Royal Society
Stable URL: http://www.jstor.org/stable/98395 .
Accessed: 07/11/2012 05:43

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact [email protected].

The Royal Society is collaborating with JSTOR to digitize, preserve and extend access to Proceedings of the
Royal Society of London. Series A, Mathematical and Physical Sciences.

http://www.jstor.org
 The formation of a blast wave by a very intense explosion
I. Theoretical discussion
BY SIR GEOFFREY TAYLOR, F.R.S.

(Received 6 October 1949)

SUMMARY AND INTRODUCTION

This paper was written early in 1941 and circulated to the Civil Defence Research
Committee of the Ministry of Home Security in June of that year. The present writer
had been told that it might be possible to produce a bomb in which a very large
amount of energy would be released by nuclear fission-the name atomic bomb had
not then been used-and the work here described represents his first attempt to form
an idea of what mechanical effects might be expected if such an explosion could occur.
In the then common explosive bomb mechanical effects were produced by the sudden
generation of a large amount of gas at a high temperature in a confined space. The
practical question which required an answer was: Would similar effects be produced
if energy could be released in a highly concentrated form unaccompanied by the
generation of gas? This paper has now been declassified, and though it has been
superseded by more complete calculations, it seems appropriate to publish it as it was
first written, without alteration, except for the omission of a few lines, the addition of
this summary, and a comparison with some more recent experimental work, so that
the writings of later workers in this field may be appreciated.
An ideal problem is here discussed. A finite amount of energy is suddenly released
in an infinitely concentrated form. The motion and pressure of the surrounding air is
calculated. It is found that a spherical shock wave is propagated outwards whose
radius R is related to the time t since the explosion started by the equation

R = S(y) t Eipo *,

where po is the atmospheric density, E is the energy released and S(y) a calculated
function of y, the ratio of the specific heats of air.
The effect of the explosion is to force most of the air within the shock front into a
thin shell just inside that front. As the front expands, the maximum pressure de-
creases till, at about 10 atm., the analysis ceases to be accurate. At 20 atm. 45 % of
the energy has been degraded into heat which is not available for doing work and used
up in expanding against atmospheric pressure. This leads to the prediction that an
atomic bomb would be only half as efficient, as a blast-producer, as a high explosive
releasing the same amount of energy.
In the ideal problem the maximum pressure is proportional to R-3, and comparison
with the measured pressures near high explosives, in the range of radii where the two
might be expected to be comparable, shows that these conclusions are borne out by
experiment.

Vol. 2oI. A. (22 March I950) [ 159 ]

160 Sir Geoffrey Taylor
SIMILARITY ASSUMPTION

The propagation and decay of a blast wave in air has been studied for the case
when the maximum excess over atmospheric pressure does not exceed 2 atm. At
great distances R from the explosion centre the pressure excess decays as in a sound
wave proportionally to R-1. At points nearer to the centre it decays more rapidly
than R-1. When the excess pressure is 05 atm., for instance, a logarithmic plot shows
that it varies as R-19. When the excess pressure is 1 5 atm. the decay is proportional
to R-2 8. It is difficult to analyze blast waves in air at points near the explosion centre
because the initial shock wave raises the entropy of the air it traverses by an amount
which depends on the intensity of the shock wave. The passage of a spherical shock
wave, therefore, leaves the air in a state in which the entropy decreases radially so
that after its passage, when the air has returned to atmospheric pressure, the air
temperature decreases with increasing distance from the site of the explosion. For
this reason the density is not a single-valued function of the pressure in a blast wave.
After the passage of the blast wave, the relationship between pressure and density for
any given particle of air is simply the adiabatic one corresponding with the entropy
with which that particle was endowed by the shock wave during its passage past it.
For this reason it is in general necessary to use a form of analysis in which the initial
position of each particle is retained as one of the variables. This introduces great
complexity and, in general, solutions can only be derived by using step-by-step
numerical integration. On the other hand, the great simplicity which has been intro-
duced into two analogous problems, namely, the spherical detonation wave (Taylor
I950) and the air wave surrounding a uniformly expanding sphere (Taylor 1946), by
assuming that the disturbance is similar at all times, merely increasing its linear
dimensions with increasing time from initiation, gives encouragement to an attempt
to apply similar principles to the blast wave produced by a very intense explosion in
a very small volume.
It is clear that the type of similarity which proved to be possible in the two above-
mentioned problems cannot apply to a blast wave because in the latter case the
intensity must decrease with increasing distance while the total energy remains
constant. In the former the energy associated with the motion increased proportion-
ally to the cube of the radius while the pressure and velocity at corresponding points
was independent of time.
The appropriate similarity assumptions for an expanding blast wave of constant
total energy are
pressure, p/po = y = R-3fi, (1)
density, p/po = /f, (2)
radial velocity, u = R-01 (3)

where R is the radius of the shock wave forming the outer edge of the disturbance,
po and po are the pressure and density of the undisturbed atmosphere. If r is the radial
co-ordinate, y = rlR and fl, 01 and 2/ are functions of y. It is found that these
assumptions are consistent with the equations of motion and continuity and with the
equation of state of a perfect gas.
 Formation of a blast wave by a very intense explosion. I 161
The equation of motion is au au -Poay
t r p ar (4
Substituting from (1), (2) and (3) in (4) and writingfl,
f Sjfor a f1, a-

-G01+ lqol)
d-dRt + -(1fl+ po 0.= (5)

This can be satisfied if = AR-, (6)

where A is a constant, and

-A(1?0y+,qo)?+q1 +p L-= 0. (7)

The equation of continuity is

ap rp (8)

Substituting from (1), (2), (3) and (6), (8) becomes

- kq71r'+ #'o + 7f 0 + 2 01)= 0. (9)

The equation of state for a perfect gas is

t + U ar (PP-Y) 0 (10)

where y is the ratio of specific heats.

Substituting from (1), (2), (3) and (6), (10) becomes

A(3f1+Yqfl)+rfl f1(-Ay+?1)-q5lfl= 0. (11)

The equations (7), (9) and (11) may be reduced to a non-dimensional form by
substituting f=fja2/A2' (12)

0 = 01/A, (13)
where a is the velocity of sound in air so that a2 = ypo/po. The resulting equations
which contain only one parameter, namely, y, are

95) lf-30, (7a)

Vk' qS+?2q5/y
_ (9a)

+
3f+yf '+ f(-+q5)-5f'=0O. (O. a)

Eliminating 3b' from (lla) by means of (7a) and (9a) the equation for calculatingf'
whenf, 4, ' nd y are given is
-
f '{(q 0)2 _f1/fl= f{-3yq+0(3+1 )-2y22/y}e. (14)
I I-2
162 Sir Geoffrey Taylor
Whenf' has been found from (14), q' can be calculated from (7a) and hence 3f' from
(9 a). Thus if for any value of y, f, q and Vrare known their values can be computed
step-by-step for other values of .

SHOCK-WAVE CONDITIONS

The conditions at the shock wave y = 1 are given by the Rankine-Hugoniot

relations which may be reduced to the form

Pi _ Y-l+(y+l)Y (15
Po Y+1+(y-l)Yl' )

U21
-a-= y{Y-l?(y?l)Y-}, (16)

Uy-1 (y?-1)y 17
where Pi, Ui and Yi represent the values of p, u and y immediately behind the shock
wave and U = dR/dT is the radial velocity of the shock wave.
These conditions cannot be satisfied consistently with the similarity assumptions
represented by (1), (2) and (3). On the other hand, when Yi is large so that the pressure
is high compared with atmospheric pressure, (15), (16) and (17) assume the approxi-
mate asymptotic forms Pi 1y+
Po -'
(16a)
U2 2y
a2=; +lZl (16a)

u _ 2 (17a)
U y+l,
These approximate boundary conditions are consistent with (1), (2), (3) and (6); in
fact (15 a) yields, for the conditions at y = 1,

7 1'+
- (15b)

(16a) yields f = 1/+, (16b)

and (17 a) yields = 21 (17 b)

ENERGY

The total energy E of the disturbance may be regarded as consisting of two parts,
the kinetic energy rR
K.E. = 4J pu2r2 dr,

and the heat energy H.E. = 474 p2 ldr.

 Formatton of a blast wave by a very intense explosion. I 163
In terms of the variables f, , r and I

E
4=
E4TA2 pI
Po 2dy+1
i/rq52y fj2d
J l\a2Q(yP I0' ay,jj}
or since Po = a2po/y, E = BpOA2, where B is a function of y only whose value is

B = 27 fo2y2&d? + -4 jfy2 dd. (18)

Since the two integrals in (18) are both functions of y only it seems that for a given
value of y, A2 is simply proportional to E!po.

NUMERICAL SOLUTION FORy =1-4

When y = 1 4 the boundary values off, q5and Vrat y = 1 are from (15a), (16a).
(1 7a), Aand 6. Values off, 0 and ~f were calculated from ?y 1 0 to ql = 0 5, using
7,

intervals of 0 02 in . Starting each step with values off ', O5',?/r',f, 0 and ?r found in
previous steps, values of f', S' and 3b' at the end of the interval were predicted by
assuming that the previous two values form a geometrical progression with the pre-
dicted one; thus the (s + 1)th term, fS1 in a series of values of f' was taken as
fs-+ = (fS)2/f_1. With this assumed value the mean value of f' in the sth interval was
taken as 1(fgs - +fs) and the increment in f was taken as (0 02) (@)(fs +i +fs). The
values of fs+1, qS'+ and /s+jwere then calculated from formulae (14), (7a) and (9 a).
If they differed appreciably from the predicted values a second approximation was
worked out, replacing the estimated values offs'? by this new calculated value. In
the early stages of the calculation neary = 1 two or three approximations were made,
but in the later stages the estimated value was so close to the calculated one that the
value off' calculated in this first approximation was used directly in the next stage.
The results are given in table 1 and are shown in the curves of figure 1. These
curves and also table 1 show three striking features: (a) the 0 curve rapidly settles
down to a curve which is very nearly a straight line through the origin, (b) the density
curve i/ rapidly approaches the axis Vf= 0, in fact at y = 05 the density is only 0*007
of the density of the undisturbed atmosphere, (c) the pressure becomes practically
constant and equal to 0O436/1P167= 0 37 of the maximum pressure. These facts
suggest that the solution tends to a limiting form as y decreases in which 0 = cy,
= c = constant, f = 0 436, f', Vfand Vf' become small. Substituting for - from
yif
(7a), (14) becomes

(q - 0) 2= yo'( - 0) + -3-y-3,1+(3+1 2 )0- #'(9 (19)

f

Dividing by y- 4 (19) becomes

' 3+270 (20)
(y-0)=y5o,

If the left-hand side which containsf'/f be neglected the approximate solution of (20)
for which 0 vanishes at y =0 is / (21)
164 Sir Geoffrey Taylor
12 [166+

I
6-0

0 404 .

- w// 1- ~~~~20

0 04 0)8
, - r/R

lf (step-by-step calculation); -- + -, curve f (approximate

FIGuRE 1. - -, curves f ancd
formulae). In the other curves the small dots represent the steps of the calculations, the
larger symnbols represent approximate formulae for: A, curve 0; 0, curve 0 -,I/^
o, curve ?If
TABLE 1. STEP-BY-STEP CALCULATION FOR y= 14

1/ f 0
1.00 1-167 0-833 6.000
0*98 0*949 0*798 4*000
0*96 0*808 0*767 2-808
0*94 0*711 0*737 2-052
0*92 0*643 0*711 1P534
0 90 0*593 0*687 1-177
0*88 0*556 0*665 0.919
0*86 0*528 0-644 0-727
0*84 0*507 0-625 0-578
0*82 0*491 0*607 0-462
0-80 0*478 0.590 0 370
0-78 0*468 0*573 0-297
0-76 0*461 0*557 0-239
0-74 0-455 0-542 0d191
0-72 0-450 0*527 0-152
0*70 0*447 0*513 0*120
0*68 0-444 0-498 0-095
0-66 0-442 0-484 0-074
0-64 0-440 0-470 0-058
0-62 0-439 0*456 0-044
0-60 0-438 0-443 0-034
0-58 0*438 0*428 0-026
0*56 0-437 0-415 0-019
0-54 0-437 0-402 0-014
0-52 0-437 0*389 0-010
0A0 0-436 0 375 0 007
 Formation of a blast wave by a very intense explosion. I 165
The line 0- = /y is shown in figure 1. It will be seen that the points calculated by the
step-by-step method nearly run into this line. The difference appears to be due to the
accumulation of errors in calculation.

APPROXIMATE FORMULAE

The fact that the qScurve seems to leave the straight line 0 q1y rather rapidly
after remaining close to it over the range y = 0 to q=0 5 suggests that an approxi-
mate set of formulae might be found assuming
y/y+ojy(22)
5=-9117

where n is a positive number which may be expected to be more than, say, 3 or 4. If

this formula applies aty = 1,
1I 2 y-l
a 2(
+1 o +l); (23)

inserting 0S= y/y +oyn, q' = n/y+nc'yn-l in (20), the value of f'/f at y 1 is
f '/f= y(n+ 2) (y + 1)/(y-1 ). From (14) and (15b), (l 6 b), (I 7b) the true value of
7,y -3
f'/f at =I1 is 2y2+ . Equating these two forms,
Y- 1 ~utn
7y -1
n - l (24)
The values of oaand n have now been determined to give the correct values of f'/f;
0 and qS'at y = 1, i/' is determined by (9a) so that all the six correct values off, qS, /
f', O', /' are consistent with (22) at =1. Substituting for qSfrom (22) in (20),
f (n + 2)oy2yq-2
f f y 1n1
-yaqocn-1 *( ~~~~~~~~~~(25
The integral of (25) which gives the correct value of f at y = 1 is

logf = log 1'l oY2+7Y 3iog (Y+lYfln-1(26)

At q = 0 5 this gives f = 0457 when y = 1P4. The value calculated by the step-by-
step integration is 0-436, a difference of 5 %.
The approximate form for #/ might be found by inserting the approximate forms
for qSand qS'in (9 a). Thus
log ~f = log Y+ Ij3 + (n + 2) oayVn1 d.(27)
'Y 1 Jr -7 1) Iq yqn (7

Integrating this and substituting for acfrom (23),

log # = log'Yl+ 31logy-2 (;'+)log()' lY). (28)

When y is small this formula gives

f = Dy3l(y-1), (29)
166 Sir Geoffrey Taylor

where log D = log + 2 (y + 5) log (+) (30)

When y = 1P4(30) gives D = 1P76so that

r = 176y75. (29a)
At yi 0 5 this gives r = 0 0097; the step-by-step calculation gives i/ = 0 0073.
At y = 08 formula (28) gives #f = 0-387, while table 1 gives #f = 0 370. At y- 0 9
formula (28) gives Vf= 1 24, while the step-by-step solution gives 1P18. Some points
calculated by the approximate formulae are shown in figure 1.
In the central region of the disturbance the density decreases proportionally to
r3/(y-1); the fact that the pressure is nearly constant there means that the temperature
increases proportionally to r-3/(-1). At first sight it might be supposed that these
very high temperatures involve a high concentration of energy near the centre. This
is not the case, however, for the energy per unit volume of a gas is simply p/(y - 1) so
that the distribution of energy is uniform.
Values off, qSand Vrfor y - calculated by the approximate formulae are given in
table 2.
TABLE 2. APPROXIMATE CALCULATION Y= 1*666
y f 0 fl
1l00 1*250 0*750 4l00
0*95 0*892 0*680 2*30
0*90 0*694 0*620 1*14
0*80 0*519 0*519 0*63
0*70 0*425 0*445 0*29
0*50 0-379 0*300 0-05
0*00 0*344 0*000 0l00

BLAST WAVE EXPRESSED IN TERMS OF THE ENERGY OF THE EXPLOSION

It has been seen in (18) that E/poA2 is a function of y only. Evaluating the in-
tegrals in (18) for the case for y = 1P4,and using the step-by-step calculations, it is
found that
fY fd =-1085 and fy2fdy = 04187.
202

The kinetic energy of the disturbance is therefore

K.E. = 27T(04185)pOA2
= 1164poA2, (31)

while the heat energy is

? H.E. =( -4)(O4) (0 187)pOA2= 4196poA2; (32)

the total energy is t herefore

E = 5*36poA2. (33)
 Formation of a blast wave by a very intense explosion. I 167
PRESSURE
The pressure p at any point is

pR-3fA2 = R-3fPOA -= 0133R-3Ef. (34)

The maximum pressure at any distance corresponds withf = 1P166at r = R. This is

therefore Pmax. = 04155R-3 E. (35)

VELOCITY OF AIR AND SHOCK WAVE

The velocity u of the gas at any point is

u = R-iAq! = R-AEi(Bpo)-14. (36)
The velocity of radial expansion of the disturbance is, from (6),
dR = AR-i = (37)
dt R-Ei(Bpo)-,

so that, if t is the time since the beginning of the explosion,

t = 25R(Bpo)1E-i = 0926RIpLE-+, (38)
when y = 1F4.
The formulae (34) to (38) show some interesting features. Though the pressure
wave is conveyed outwards entirely by the air the magnitude of the pressure depends
only on ER-3 and not on the atmospheric density po. The time scale, however, is
proportional to pi. It is of interest to calculate the pressure-time relationship for
a fixed point, i.e. the pressure to which a fixed object would be subjected as the blast
wave passed over it. If to is the time since initiation taken for the wave to reach
radius Ro the pressure at time t at radius Ro is given by
p (Ro)3 f
Pi B (39)1,=1
where Pi is the pressure in the shock wave as it passed over radius Roat the time to,
R is the radius of the shock wave at time t and y1= RO/R. [f],=1 is the maximum
value of f, namely, 1P166when y = 1P4. y is related to t/to through (38) so that
1 to\W
y = (to/t)f and = [ 1(t [f]qa1i* (40)

Values of p/P1 calculated by (40) for y = 1P4are shown in figure 2.

[0 l

Q05 _

0 20 40 60
t/to
FIGURE 2. Pressure-time curve at a fixed point.
168 Sir Geoffrey Taylor
TEMPERAT-URE

The temperature T at any point is related to the pressure and density by the
relationship T ppo 1- 33ER-3
when y =b4. (41)
To~~i1~PoP
PoP Po~
Po Wf
Sincef tends to a uniform value 0*436 in the central region (r < -R) and */ tends to the
value 3f = 1.76V7f5, T tends to the value
T ER-3 (0.133) (0.436) 5_ (42)
TO P0 1-76 .3~E~-~
Po
Thus the temperature near the centre is very high; for instance, when the wave has
expanded to such a distance that the pressure in the central region is reduced to
atmospheric pressure, po = (0.133) (0.436) ER-3, then (42) gives T/To = 75/1-76
and at q = 0 5, r75 = 181 so that T/I0 = 103. If To = 2730, T = 27,0000. The
temperature left behind by the blast wave is therefore very high, but the energy
density is not high because the density of the gas is correspondingly low.

HEAT ENERGY LEFT IN THE AIR AFTER IT HAS RETURNED TO ATMOSPHERIC PRESSURE

The energy available for doing mechanical work is less than the total heat energy of
the air. The heated air left behind by the shock wave can in fact only do mechanical
work by expanding down to atmospheric pressure, whereas to convert the whole of
the heat energy into mechanical work by adiabatic expansion the air would have to be
expanded to an infinite extent till the pressure was zero. After the blast wave has
been propagated away and the air has returned to atmospheric pressure it is left at
a temperature T1,which is greater than To,the atmospheric temperature. The energy
required to raise the temperature of air from Toto T1is therefore left in the atmosphere
in a form in which it is not available for doing mechanical work directly on the
surrounding atmosphere. This energy, the total amount of which will be denoted by
E1, is wasted as a blast-wave producer.
The energy so wasted at any stage of the disturbance can be calculated by finding
the temperature T1 to which each element of the blast wave would be reduced if it
were expanded adiabatically to atmospheric pressure. If T is the temperature of an
element of the blast wave
T i p V-V)lr
=A2 fR-3 (Y-1)/'Y

f
Also
Also zT=ppo -- A2 R-3,
~~~~~~To
pop ~fa2
T,1 flly ZA2R-3 1/y
hence = fb (A2 (43)

The total heat energy per unit mass of air at temperature T1 is

T
x (gas constant) = ( 1)po0
 Formation of a blast wave by a very intense explosion. 1 169
The increase in heat energy per unit mass over that which the air contained before the
passage of the disturbance is therefore P (7 I). The increase per unit

volume of gas within the disturbed sphere is therefore Po-1 (, _1) . Hence from (43)
,y To
the total energy wasted when -the sphere has expanded to radius B is

E = 4i7TR3PoJ{f (A21i?3)117 } 2dy. (44)

This expression may conveniently be reduced to non-dimensional form by dividing

by the total energy E of the explosion which is related to A by the formula (18).
After inserting a2/y for po/po, this gives

El 47B ( A2A-")f(3 y)l'ff/2dq f ; 2d ] (45)

437poR3 is the total mass of air in the sphere of radius R. This is also 47TR3p4 Vf2 dy,
so that 1
Jfbq2dy =-. (46)

The quantity A2R-3/a2 is related to the maximum pressure Pi at the shock wave by
the equation Pi A21-3
- =-
Y1-Yi p0 2a2
t2 [f In==

where Yi is the pressure in the shock wave expressed in atmospheres. (46) therefore
reduces to El BLyi/7 [ ]ff 2d
dy j(7 -

For y = 1P4numerical integration gives

B = 5-36 (see (18)), ft=1 = P166, ffl1h12dq = 0219.
(47) reduces therefore to E I
-1- = l-[O-958yI1,,4-1P63]. (48)

Some values.of E1/E are given in the second line of table 3.

It is clear that E1/E must continually increase as R increases, and Yi decreases
because the contribution toEl due to the air enclosed in the shock-wave surface when
its radius is R2, say, remains unchanged when this air subsequently expands. A
further positive contribution to E1 is made by each subsequent layer of air included

TABLE 3

y, (atm. at shock wave) 10,000 1,000 100 50 20 10* 5*

E1/E (proportion of 0.069 0-132 0.240 0.281 0.325 0.333* 0.28*
energy wasted)
(El+E2)/E 0*096 0*189 0 337 0 393 0 455 -

* Formulae inaccurate when Yi < 10.

170 Sir Geoffrey Taylor

within the disturbance. The fact that formula (48) gives a value of E1/E which in-
creases till yi is reduced to 10 and then subsequently decreases is due to the in-
accuracy of the approximate boundary conditions (I 5a), (IG6a)and (17 a), which are
used to replace the true boundary conditions (15), (16) and (17).
When y = I -4 and y1 = 10 the true value of pl/p0 is 3-8 instead of 6-0 as is assumed,
the true value of U2/a2 iS887 instead of 8-6 and the true value of n,/ U is 0 74 instead of
0-83.
When yi = 5 the errors are much larger, niamely, pl/po is 2-8 inistead of 6*O,U2/a2 iS
4*4 instead of 4-3, and ut1/U is 0-64 instead of 0.83. The proportion of the energy
wasted, namely, E1/E, is shown as a function of yi in figure 3, yi being plotted on
a logarithmic scale.
o.5

0-4-

t0-3

0 10O 2-0 3-0 4-0

log10 Yi
FiGURE 3. Heavy line, (E + E2)/E; thin line EIIE. (E + E2)/E is the proportion of the
initial energy which is no longer available for doing work in propagation. E2/E is the
work done by heated air expanding against atmospheric pressure (see note added October
1949 (p. 172)).

It will be seen that the limiting value of E1/E is certainly greater than 0-32, its
value for yi = 20. It is not possible to find out how much greater without tracing the
development of the blast wave using laborious step-by-step methods fdr values of y,
less than, say, 10 or 20.

COMPARISON WIETH HIGH EXPLOSIVES

The range within which any comparison between the foregoing theory and the
blast waves close to actual high explosives can be made is severely limited. In the
first place the condition that the initial disturbance is so concentrated that the mass
of the material in which the energy is originally concentrated is small compared with
the mass of the air involved in the disturbance at any time limits the comparable
condition during a real explosion to one in which the whole mass of air involved is
several times that of the explosive. In the second place the modified form of the
 Formation of a blacstwave by a very intense explosion. I 171
shock-wave condition used in the analysis is only nearly correct when the rise in
pressure at the shock-wave front is several-say at least 5 or 10-atmospheres. In a
real explosive this limits the range of radii of shock wave over which comparison could
be made to narrow limits. Thus with 10 lb. of C.E.* the radius R at which the weight
of explosive is equal to that of the air in the blast wave is 3 ft., while at 3-8 ft. the air is
only double the weight of the explosive. The pressure in the blast wave at a radius of
6 ft. was found to be 9 atm., while at 8 ft. it was about 5 atm. It seems, therefore, that
in this case the range in which approximate agreement with the present theory could
be expected only extends from 3 8 to 6ft. from the 1Olb. charge.
Taking the energy released on exploding C.E. to be 0-95kcal./g. the energy
released when 1Olb. is exploded is 1-8 x 1014ergs. If this energy had been released
instantaneously at a point as in the foregoing calculations the maximum pressure at
distance R given by (35) is

Pmax.R3 (0 155) (1 80 x 1014) = 2-79 x 1013 ergs.

= (49)
Expressed in terms of atmospheres Pmax. is identical with Y1* If R is expressed in
feet, (49) becomes 2-79 x 103
YIR3 = (30.45)3 x 106 ( )

The line representing this relationship on a logarithmic scale is shown in figure 4.

I. 0 _
+8
08 +6 R=7 ft.
_~+5 range for
comparison
YIR'=990 4 3-8 ft.
~~~~~~~~R=
0

?2

0-6 1.0 14 1[8

log1oYi (atm.)
FIGURE 4. Blast pressures near 10 lb. charge of C.E. compared with calculated blast pressures
due to instantaneous release of energy of 10 lb. C.E. at a point. The numbers against the
points on the curve give distances in feet.

Though no suitable pressure measurements have been made, the maximum

pressure in the blast from 1Olb. of C.E. has been found indirectly by observing the
velocity of expansion of the luminous zone and, at greater radii, the blast-wave front.
These values taken from a curve given in a report on some experiments made by the
Road Research Laboratory are given in table 4. The observed values of U in ft./sec.
given in column 2 of this table and the values of Yi (in atmospheres) found from the
* C.E. is the name by which a certain high explosive used in many experiments by the
Ministry of Home Security was known.
172 Sir Geoffrey Taylor
shock-wave formulae are given in column 3, where they are described as observed
values though they were not observed directly. The 'observed' values are shown in
figure 4. The values of Yi calculated from (50) are given in column 4.
TABLE 4
Yi (atm.) y1 calculated for
observed with concentrated explosion
R (ft.) U (ft./sec.) C.E. by (50)
8 2350 6*2
6 3100 9*3 4.6a
5 3800 14-0 79 crangeof
4 4820 22*6 155J comparison
3 6200 37.5
2 8540 71-8

Though the observed values are higher than those calculated, it will be noticed that
in the range of radii 3-8 to 6 ft., in which comparison can be made, the observed curve
is nearly parallel to the theoretical line y R3 990. In this range, therefore, the
intensity of the shock wave varies nearly as the inverse cube of the distance from the
explosion. The fact that the observed values are about twice as great as those calcu-
lated on the assumption that the energy is emitted instantaneously at a point may
perhaps be due to the fact that the measurements used in table 4 correspond with
conditions on the central plane perpendicular to the axis of symmetry of the
cylindrical charge used. The velocity of propagation of the luminous zone is greater
on this plane and on the axis of symmetry than in other radial directions so that the
pressures deduced in column 3 of table 4 are greater than the mean pressures at the
corresponding radii.
On the other hand, it has been seen that by the time the maximum pressure has
fallen to 20 atm., 32% of the energy has been left behind in the neighbourhood of the
concentrated explosive source, raising the air temperature there to very high values.
The burnt gases of a real high explosive are at a very much lower temperature even
while they are at the high pressure of the detonation wave. Their temperature is still
lower when they have expanded adiabatically to atmospheric pressure, so that little
heat energy is left in them. To this extent, therefore, a real high explosive may be
expected to be more efficient as a blast producer than the theoretical infinitely con-
centrated source here considered.
Note added, October1949. The data on which the comparison was based between
the pressures deduced by theory and those observed near detonating explosives were
obtained in 1940. More recent data obtained at the Road Research Laboratory
using a mixture of the two explosives R.D.X. and T.N.T. have been given by Dr
Marley. These are given in table 5, which shows the values of U observed for various
TABLE 5. PRESSURE Y1Po AT DISTANCE R FROM EXPLOSION OF WEIGHT W OF
T.N.T.-R.D.X. MIXTURE
R/W- (ft./lb.-) 0*5 1.0 1.5 2-0 2*5 3*0 3.5
U (thousands ft./sec.) 143 11.0 8.4 6*6 5-1 40 3.3
y, (atm.) 198 117 68-3 42.0 25*1 15*4 10*5
R/EIx 104(cm./ergsl) 5.39 10 8 16-2 21-5 27-0 32.4 37.8
 Formation of a blast wave by a very intense explosion. I 173
values of R/WI. R the distance from the explosive is expressed in feet and W its
weight in pounds. The third line in table 5 shows the result of deducing Yi from U
using y = 1-4 in (16) and a = 1100ft./sec. in (16).
For comparison with the concentrated point-source explosion, the value of RE-,
expressed in cm. (erg.)-+ is found by multiplying the figures in line 1, table 5, by
(354)V (1200 x42 107) = 1-078x 1o-3. The first factor converts ft. (lb.)-} to

cm. (g.)-i, and the second replaces 1 g. by the equivalent energy released by this
explosive, namely, 1200 cal. The values of RE-i are given in line 4, table 5. In
figure 5 values of log10 Yi are plotted against log10RE-1, and the theoretical values
for a point source of the same energy as the chemical explosive are plotted in the
same diagram. Comparing figures 4 and 5 it seems that the more recent shock-wave
velocity results are qualitatively similar to the older ones in their relation to the
point-source theory. The range of values of Yi for which comparison between theory
and observation might be significant, is marked in figure 5.

_N
2-0 -
'20~~~~~~~~~~~~~~~~~~~~~~~~~'

07
\; grange for
__ _______\_comparison .9
0
10 14

o
-4 -3 -2
log1oRE-,
FIGuR.E 5. Blast pressures near a chemical explosive (R.D.X.+T.N.T.) compared with
theoretical pressure for conceiutrated explosion with same release of energy. Heavy line
(upper part) is taken from shock-wave velocity measurements. Heavy line (lower part) is
from piezo-electric crystals. Thin line, y, =0O 155E7(poR3). The figures against the points
represent the ratio of the mass of the air within the shock wave to the mass of the explosive.

It will be seen that the chemical explosive is a more efficient blast producer than
a point source of the same energy. The ratio of the pressures in the range of com-
parison is about 3 to 1. This is more than might be expected in view of the calculation
of E1/E as a function of yi which is given in table 3. E1 is the heat energy which
is unavailable for doing mechanical work after expanding to pressure po Of the
remaining energy, E - a part E2 is used in doing work against atmospheric
pressure during the expansion of the heated air. The remaining energy, namely,
E-1 -E2, is available for propagating the blast wave.
174 Sir Geoffrey Taylor
To find E2, the work done by unit volume of the gas at radius yR in expanding to
atmospheric pressure is Tp l) Po- From (43)

-1
T, A2 \(1-r)Ir
= - R-3f)
p
and p-f a=f
A2R-3
T- a2 Po
aPo

hence E2 47TR34 {(R-3fA2) 1} 2dq, (51)

A2 _ __ _
but =- , so that
a2]?3 (fX=1
( 1
E2 -4iR3pO[YF f1fy'r2d
dq - q2d5
dq ~~~~~~~~(52)
E E 4=1

The first integral has already been calculated and found to be 0-219 when y = 1-4
(see (47) and (48)). Substituting for Prax. from (35),

E2 = 4__(_ _155)[_ (1 11YY

_ 1 (53)

Values of (E2+ E1)/E have been added as a third line in table 3 and a corresponding
curve to figure 3.

REFERENCES

Taylor, Sir Geoffrey I946 Proc. Roy. Soc. A, 186, 273.

Taylor, Sir Geoffrey I950 Proc. Roy. Soc. A, 201, 175.