lnt, J. Rock Mech. Min. Sci. & Geomech. Abstr. Vol. 30, No. 4, pp. 413--429,1993 0148-9062/93$6.00+ 0.

00
Printedin GreatBritain PergamonPressLtd

Fragmentation Mechanism in Crater Blasting

W. L. F O U R N E Y t
R. D. D I C K t
X. J. W A N G t
Y. WEIr
This paper presents the results of a series of model tests that were conducted
to investigate the mechanism of fragmentation in a cratering situation. In
particular, the size of the crushed zone as a function of charge size was
investigated using both P M M A and Homalite models. The increase in stress
level traveling out into the polymeric models as charge sizes were increased was
also investigated. These were of interest since one of the possible mechanisms
being investigated was spall and the magnitude of the reflected wave depends
both upon the energy consumed in creating the crushed zone and the magnitude
of the stress wave leaving the charge site. These tests were conducted in 2-D
models and the charge size was varied from lOOmg of PETN to 600rag. The
results obtained showed that over this range of charge size the size of the
crushed zone quickly reached a size where the ratio of the crushed zone radius
to the borehole radius was about seven and then remained constant for further
increases in charge. The stress level in the outgoing P-wave continued to
increase with charge size. This increase, however, was less than would be
expected. For a four-fold increase in charge, the maximum stress in the
outgoing wave only increased by about 40%.
Additional tests were conducted in Homalite lO0 to investigate the fragmen-
tation mechanism. A multiple spark gap camera in conjunction with a specially
designed smoke diversion device was used to view the fracture formation after
the charge was detonated. A number of different tests were conducted and all
showed that strong multiple spalling occurred such that the material within the
spall zone was fractured to the point where very little residual strength
remained. The material in the near vicinity of the borehole was also greatly
weakened by the system of radial crack created as the P-wave propagated
away from the explosive source and the system of circumferential cracks
created as the PP-wave traveled back across this system of radial cracks.
The mechanism of fragmentation indicated from these tests was one where
the material between the borehole and the free surface is greatly weakened by
the stress waves over the first 50 or so psec after detonation. It is proposed that
this greatly weakened area is then acted upon by the residual pressure in the
borehole to create the final crater.
The feasibility of this mechanism was checked by using a finite element code
to determine the displacement of several points in the crater region and to
compare these displacements with displacements measured in 3-D tests
conducted in models made from a very high strength-low porosity cement. The
agreement with the measure displacements was good, indicating that the
mechanism is reasonable.

INTRODUCTION explosive is placed is subjected to extremely high

pressures--in the 100 kbar range. This reaction results in
When an explosive is detonated a chemical reaction
two types of loading being applied to the material--a
occurs very rapidly and a relatively small quantity of
shock wave loading that quickly travels out into the
explosive is converted into gas at very high pressure. As
material surrounding the borehole and a longer duration
a result of this reaction the material in which the
gas pressurization loading. Over the years there has been
much controversy over the relative contribution of these
tMechanical Engineering Department, University of Maryland, two very different loads on the fraction and fragmen-
College Park, MD 20742, U.S.A. tation of rock. It is the purpose of this paper to describe
413
414 FOL!RNEY eta/.: CRATER BLASTING

a series of experiments which were conducted to When this S-wave strikes the boundary ~t also creatcs
contribute to a better understanding of the mechanism two wave types at points of non-normal incidence a
of crater formation due to the detonation of an explosive shear wave (SS-wave) and a P-wave (SP-wave). These
charge in the vicinity of a free surface. reflected waves travel back over the load site in much the
Figure 1 depicts the propagation sequence of a dila- same fashion as the P-wave does.
tational wave which is produced from a highly dynamic Typical high-speed photographs taken of the dynamic
load being applied at a point in the interior of a 2-D event by a multiple spark gap camera which has been
body. The location of the point source is shown a configured as a dynamic polariscope are shown in Fig. 2.
distance "h" from the nearest free face this being the In this figure the model has been photographed at
geometry of interest, i.e. a cratering geometry. The approx. 200,000 frames/sec while being viewed under
leading edge of the P-wave travels outward with a polarized lighting conditions. The fringes shown are lines
velocity of: of constant shear stress (difference in the principal
stress). From the fringe patterns obtained and the theory
Vp = x/ {E/[p(I - v2)]}, (1)
of elasticity it is possible to determine the state of stress
where v is Poisson's ratio, E is Young's Modulus and p (or strain) in the outgoing P-wave. The strata in the
is the material density. tangential direction ~(, is obtained from the equation:
At time t =h/Vp the wavefront reaches the free a

boundary and begins to reflect back into the media. At ~0 = -- [(1 + v)/E][(./~,)j/h] I(N/rl dr. (3)
/
points where there is non-normal incidence of the
outgoing P-wave, both a dilatational (P) wave and a where ( f , ) j is the dynamic stress fringe value. N is the
distortional (shear or S) wave occurs. The creation of fringe order and r is the radial position. Once (~, is known
both types of waves is due to the stress-free boundary the radial strain ~[ can be found from:
condition that must be met at the free face. Figure 1
{r = {# - - [(1 -{- v)/E](N/h)(/~),t. (4)
shows the development of the P-wave up to the time that
the reflected P-wave (PP-wave) reaches the point source From these strains both stresses and displacements
(borehole). Due to fracturing that occurs at the can be obtained from the stress displacement relations
borehole, a distortional (shear) wave is also created at and the stress-strain relations. Figure 3 shows for a
the explosive source and this travels at a slower velocity given test the fringe order as a function of position
given by: obtained at various times throughout the dynamic event.
Basically, each of the curves represents information
V~ = v/{E/[2P(I + v)]}. (2)
obtained in one photograph from the multiple spark gap
camera.
Figure 3a shows the results obtained for the tangential
t=20.8~sec R=50. h=50.8mm stress as a function of position from the borehole (and
time after detonation). Figure 3b gives the results
obtained for the radial stress as a function of position
from the borehole.
Notice from Fig. 3 that the maximum l¥inge order
decreases as the distance from the borehole (or time)
t=27.7~sec [
increases. This decrease is due both to material damping
and geometric dispersion. As the stress wave travels from
the source some amplitude is lost due to hysteresis losses
in the material. This is the material damping. As the
wavefront travels from the source the front must cover
more and more area and therefore the amplitude
t=34.7~sec / 1"-',M, ~ ..j_ decreases due to geometric dispersion. Most of the
~ ~ P I /"7ram
~ ~eR = 8 4 decrease in amplitude seen in Fig. 3 is due to geometric
dispersion. In a 3-D medium the geometric dispersion
would be even greater than the amount seen here in two
dimensions.
A close examination of the curves showing the
~ ~ ~ / h=50.8ram
tangential and radial stresses after detonation of the
t=41.6~sec explosive in Fig. 3 shows that in the leading edge of the
P-wave a biaxial compressive stress state exists. That is,
both the radial and the tangential stresses are compres-
sive. The radial stress is about three times larger than the
tangential stress. In the trailing part of the outgoing
P-wave both stress components are tensile.
If the compressive stresses in the leading edge of the
Fig. l. Stress wavefronts in a half plane with an internal point source. outgoing wave are sufficiently large (larger than the
 FOURNEY et al.: CRATER BLASTING 415

Fig. 2. Typical fringe patterns from a dynamic point source. Frames 1 4 of 16 frames.

compressive strength of the material) then crushing of and not increase further. At the same time the size of the
the material will occur in both the tangential and radial crushed zone around the borehole should continue to
directions. Early theories for fragmentation mechanisms increase as charge size increases.
assigned all of the fragmentation to the reflection of the
compressive (outgoing wave) into a tensile inward
EXPERIMENTAL SERIES
traveling wave. To maximize fragmentation it would
then be desirable to have the outgoing compressive pulse A series of tests were conducted to investigate what
be as large as possible. If, however, crushing occurs happens to the crush zone size and magnitude of stress
whenever the compressive pulse reaches the compressive in the outward traveling P-wave as charge size increases.
strength of the material it appears that as charge size is Six tests were conducted in P M M A (Plexiglas) with
increased that stress (or fringe order) will reach a peak charge sizes increasing from 220 to 600 mg. In this series
416 F O U R N E Y et al.: CRATER BLASTING

Tangential Stress v~ Distance

Isochrornotic Fringe Test ,~" 2 the fringe patterns for all tests and the size of the crushed
20,000 zone was also measured after each test.
t = 22.3/~Nc Figure 4 shows the decay of the fringe order number
- - - * - - - t =32.0#gec
- . - ~ - . - t = 36.7/~se¢ as a function of distance from the borehole for the
.... 0 ..... t =48.1 /~sec
PMMA material. All of the data were taken during the
t --•-- t =61 .St~sec
lo,0o0 outward propagation of the P-wave. The data for all six
tests fit well a straight line when the log of fringe order
is plotted against log of propagation distance with a
negative slope of 0.805. The scatter in the data shown is
not bad considering the variation in charge size and
borehole diameter for the six tests. The amplitude decay
curve for the fringe order is therefore given by:
N = 111.72/r °'s°~. (5)
-io,ooc |,
•~ 0 loo 200 As indicated earlier this decay is the result of both
material damping and geometric dispersion. Figure 5a
Radial Stress vs. Distance shows the fringe order as a function of charge size for
Isochromatic Fringe Test # 2 various different distances from the borehole. Also
10,000
shown in Fig. 5b is a graph showing log fringe order as
a function of charge size. The lines shown in Fig. 5b are
a least-square linear fit to the date from the tests. For a
four-fold increase in charge size the average fringe order
° increase is about 40%. For the charge sizes investigated

/V,
t
..

" '
."
:/
/
J it therefore appears that it is possible to continue to
increase the level of the stress wave going from the
explosive source, but not as much as might be expected.
~A
- I0,000 5" ~ - I/ Since the energy is related to the square of the stress a
four-fold increase in charge size could be expected to
result in a stress increase as high as a factor of two.
V ,, Figure 6 shows the results of the crushed zone
-20,0OO I measurements made on the models after the tests were
O IOO 200 concluded. Figure 6a shows the crushed zone volume as
Radial D i s t a n c e [ram) a function of charge size for the PMMA models. The
Fig. 3. Tangential and radial stresses from Test 2 having a charge of upper three points marked " L " show the results for
380 m g of PETN. models which used boreholes of dia. 7.62 mm and the
three lower points marked "S" are for the models with
of tests the borehole was located in the center of the a 5.07 mm dia. borehole. As can be seen from the figure
square model. The borehole diameter used ranged from the crushed zone size is very nearly constant with charge
5 mm for the smallest charge to 7.6 mm for the largest size. The fact that the smaller diameter boreholes
charge. All charges were fully coupled to the borehole resulted in a smaller crushed zone is contrary to what
wall. The multiple spark gap camera was used to record would be expected. Figure 6b shows results for similar

Attenuation Curve for Plexiglass

8-

7-

6-

5-
0 +
4- X
'2
"1=
U. 3-
A
$ ÷
2-

1-

O~
0
P m ~ ~u'loe mm

Fig. 4. Fringe order as a function o f distance from the borehole for various charge sizes.
 FOURNEY et al.: CRATER BLASTING 417

Fringe Order vs. Charge Size

5

7t 351
(a)

+
U.
3-
2.5-
2- m

0
•

+
m

0
+•

1.5-
1-
0.5-
0
0 16o ~ 360 ,=60 s6o e6o 76o 8OO
Chargeb Mm~r~n=

Log Fringe Order vs. Charge Size

1

~ 0.7
0.6
". .
;,= _ . I ± ~ ...... 4-

/
°;11 , ~ . . ~ . 30o
2oo
. . ~ ,.~ .4oo 46o ,~ "r~
/ ~
Charge Size Milligrwxm

I • At75mm + AtlOOmrn .~ At125mm o AtlSOm I

Fig. 5. Fringe order as a function of charge size for PMMA models: (a) fringe order vs charge size; and (b) log fringe order
vs charge size--lines are linear least-square fit.

tests conducted with Homalite models (points with M ation of the crushed zone could be increasing but our
are for intermediate size boreholes). Once again the tests were not sufficiently detailed to determine that.
smaller boreholes resulted in a smaller diameter for the Energy is apparently lost to airblast and pressure leaking
crushed zone. Notice that one test was conducted in through the sides of the model instead o f going into
Homalite with a 100mg charge which resulted in a stress energy and crushing the material.
crushed zone with nearly zero volume. In the range of Looking again at Fig. 3 in which 380 mg of P E T N was
charges between 200 and 600 mg there appears to be no used in a 5 mm diameter borehole it is evident that a
effect on crushed zone volume. The size o f the volume significant level o f loading was reached. As is evident
is different for the different sized boreholes but above from the figure, this loading resulted in a radial
200mg the volume does not appear to increase. compressive stress o f nearly 140 MPa and a tangential
Figure 7 shows the ratio o f the crushed zone diameter to compressive stress of about 34.5 MPa in the leading edge
the borehole diameter for all the models tested and (22.3/~sec after detonation). Also evident from the figure
this ratio appears to be fairly constant with respect to is the lack o f a significant tensile tail. The tensile tail in
charge size. There is some scatter in the data but the the outgoing P-wave pulse is consumed by the initiation
trend is quite clear--as the charge size is increased by a o f radial cracks and the tangential tensile tail is absorbed
factor o f four from 150 to 600 mg there is no apparent in the crushed zone.
increase in the ratio of the crushed zone diameter to
borehole diameter for either the P M M A or the Homalite
FRAGMENTATION MECHANISM
material.
It appears therefore that as the charge size increases A second series o f tests was conducted to investigate
there is a less than expected increase in the stress level the wave reflection from the free surface as well as the
in the outgoing stress wave and that there is no increase fracture formation in the vicinity o f the borehole. These
in the size o f the crushed zone. The extent o f pulveriz- tests were also 2-D and used Homolite 100.
418 F O U R N E Y et al.: C R A T E R BLASTING

PMMA Crushed Zone Volume were formed as a result of the tensile stresses which occur
1600- in the trailing portion of the outgoing P-wave. As these
(a)
~- 1400-
radial cracks are formed, the amplitude of the tensile
E stress is decreased greatly. At the time shown in Fig. 9a
._u 1200-
J~ the P-wave has reached the free boundary and has
~o
¢D
1000- •L reflected back into the model a distance of about 35 mm.
•L The fracture patterns shown in Fig. 9 were typical for the
:~ coo.
results obtained in the six different tests conducted in this
6oo- series. At 33.8~sec (Fig. 9a) about 24 radial cracks
4oo- mS initiated and are traveling away from the borehole very
mS mS
rapidly. There is a tremendous amount of energy driving
0 200-
these cracks and they are attempting to branch to
0
6OO
effectively use up the excessive energy that is available
0 100 200 300 400 500
Charge Size Milligrams since they cannot travel any faster than they are already
moving [3]. Upon reaching the free face the biaxial
compresion in the leading edge of the wave is converted
Homolite Crushed Zone Volume into biaxial tension and this tensile wave travels back
600-
into the model towards the borehole. The conversion of
(b)
the compressive stress to the tensile stress occurs such
500-
E
.o_
that very near the free surface there is a shadow zone [4]
.0
~,mo- •M •M •L in which the value of the tensile stress is less than
the tensile strength of the material--so no fracturing
occurs. Outside of this shadow zone, however, the tensile
BS stress in the inward traveling wave exceeds the tensile
strength of the material and fracturing occurs. The
•S •S
initiation of this spall fracturing is very evident in the
2 lOO-
o
photograph shown in Fig. 9a. It is especially clear in the

BR
°o 16o ~o ~o 46o ,~o coo Crush Zone Diameter Ratio for PMMA
Charge Size Milligrams 20-
Fig. 6. Crush zone volume produced by various size PETN charges in 18- (a)
P M M A and homolite.
16-
14-
In these tests the charge was placed 50 mm from the ~
12-
free surface and a special "T-shaped" tube was used on .~_
10-
both sides of the 2-D model to conduct the smoke and o"6
O 8- II •
debris created by the explosive away from the fracture • II

site. In this way the high-speed camera could be used to 6-

maximum advantage to view the fracture formation 4-

process. Figure 8 shows the model geometry used in the 2-
test--including the smoke detection device. Figure 9 0
15o a:~ a~o aro 35o 460 4so 56o s~o 600
presents four photographs of the 16 taken with the Charge Size Milligrams
high-speed camera during one of the several tests
conducted. These were taken between 33.8 and 50.7 #sec
after detonation of the explosive and represent frames Crush Zone Diameter Ratio for Homolite
20-
5-8. The size of charge used ranged from 100 to 250 mg
18- (b)
with the smaller charge being used in a single test
which utilized a model o f 3.2 mm thickness whereas all 16-
o f the other tests used models which were 6.4 mm in 14-
thickness. 12-
The photographs presented in Fig. 9 are intended to 10
show how the fracture and fragmentation occurs in a "~
8
cratering situation. The speed of the outgoing P-wave in 6
I "
• •
this material (Homolite 100) is 2460 m/sec. As seen from
4
the first frame the P-wave has traveled 85 mm from the
explosive site in 33.8 #sec. At this time many radial 2

fractures have formed and have propagated to a distance 0

of about 25 mm from the borehole. This propagation Charge Size Milligrams
occurred at an average velocity o f about 737 m/sec or Fig. 7. Crush zone diameter ratio for P M M A and homolite for various
about 30% o f the P-wave speed. These radial cracks charge sizes.
 FOURNEY et al.: CRATERBLASTING 419

photographs presented in the next three frames. In Fig. 9 sufficiently high, enough energy remains after the first
the horizontal grid markings shown are 12.7 mm apart fracture initiations to initiate other fractures as the wave
above the borehole and 25.4mm apart below the continues to propagate inward. This is called multiple
borehole. Note that there are multiple initiation sites all spalling. Multiple spalling is evident in Figs 9b--d where
along the reflected wavefront----cspecially evident in many lines of spall are visible. A close examination of
Fig. 9b. As the tensile stresses in the forward edge of the Fig. 9b for example shows four lines of initiation sites by
inward traveling P-wave initiate fractures the amplitudes multiple spalling.
of the tensile stresses decrease since energy is used for the Also shown in Figs 9b-d are the creation of circum-
initiation process. If the amplitude upon reflection is ferential cracks as the front of the PP-wave travels over

(a) SMOKE GUIDING TUBE DEVICE

Imake tube

4 5 0 ntnl

A-section

/
Free Surface

(b)
?Omm
Borehole 305 mm

3 0 5 mm
Fig. 8. (a) Drawing of smoke deflection device used to make fracture patterns visible in the borehole vicinity. (b) Model
geometry used in test series.
 Z

r~

!
 FOURNEY et al.: CRATER BLASTING 421

Fig. 9. Fracture patterns caused by charge in vicinity of free surface: (a) frame 5--33.8 psec after detonation; (b) frame 6 at
40.4psec; (c) flame 7 taken 46.1 #sec after detonation; and (d) flame 8 at 50.7 psec.
422 |:()URNEY eta/.: CRATER BLASTING

the outward propagating radial cracks. The stress state borehole. When the final fracture surface is viewed after
in the PP-wave has the effect of turning the radial cracks the blast is concluded the features on the surface are the
into a circumferential direction and causing multiple outlines of the radial cracks which form the outline of
branching to occur. This was noted in an earlier series the crater and these do grow from the borehole outward.
of tests [5] and we called it "barrier branching" since the The mechanism of the formation of the crater is,
PP-wavefront acts momentarily as a barrier to the radial however, heavily dependent on the spall, multiple spatl
crack propagation. After the PP-wave passes by, the and the radial and circumferential crack network which
cracks once again turn and propagate in the radial is formed as the PP-wave passes back over the extending
direction. radial cracks. Once these mechanisms occur and the
Figure 9d gives a good picture of what the fracture material between the borehole and the free surface is
pattern looks like right after most of the initial wave weakened, the gas pressure then fills the fractures that
action has passed. Note that at the time in question are connected to the borehole and the pressure then lifts
namely about 50~sec after detonation--the area very and throws the material out of the crater.
near the free surface is fairly well fractured from the Figure 10 presents two frames taken relatively late in
multiple spalling. The material there has very little another dynamic test. The two photographs shown in
strength. Although the original direction of initiation that figure were taken at 230 and 237 gsec after deton-
was parallel to the PP-wavefront, the state of stress in the ation. In these figures movement of the fragments near
front is biaxial tension and fractures have formed and the free face is evident. The free lace is no longer straight
grown both parallel and perpendicular to the wavefront. due to movement of the pieces above the spall fracture
On the side o f the borehole towards the free face, radial described earlier. Close examination of the figure also
cracks have propagated about half the distance to the shows radial cracks just outside the crater that are still
free face. On this side of the borehole the radial cracks propagating towards the free face. These are the radial
have all been connected together with the circumferential cracks that are felt to define the final crater.
cracking that was formed as a result of the barrier
branching. As can be seen from Fig. 9d the material
COMPUTATIONAL INVESTIGATION
contained in the 90 ° sector just between the free surface
and the borehole is well fragmented even at this very In an effort to verify the mechanism of crater frag-
short time after detonation. (The black grid in Fig. 9 mentation being proposed and to compare experimental
divides the models into four equal segments of 90 ': each.) results obtained from 3-D testing it was necessary to
In the area outside of the 90 ~' segment only radial perform finite-element calculations.
fractures exist, not a combination of radial, circumferen- The proposed mechanism was based primarily on
tial and spall fractures. results obtained from the dynamic photoelastic tests.
One of the biggest puzzles to some of the authors These were conducted in 2-D models. We desired to
about the mechanism of crater blasting has been post compare these results to the results of tests conducted in
mortem experimental evidence that has shown that the 3-D models in materials that were more rock-like in
crater travels from the borehole outward towards the behavior.
free surface as opposed to traveling from the free surface A 2-D static finite element computation was
inwards [6]. If the main mechanism was that of spalling conducted to predict the displacement (magnitude and
or of multiple spalling the fracture surface should show direction) of selected points within a model which could
a propagation that would be from the free surface be compared to displacement results obtained from
inward. In previous cratering experiments, evidence is instrumented concrete models. The best results for
quite clear that the final fracture surface for the crater comparison were obtained in cylindrical models made
is created by fractures that start at the borehole and from high pozzolan concrete. Those models possessed
propagate outward towards the free surface. Results high strength and low porosity and reacted to the
from this current test series clear up the previous mis- explosive loading in a fashion that simulated as close as
understanding in this regard. possible an actual rock structure. The fact that it was a
The process of crater formation is as follows. The cement, however, permitted the insertion of various
multiple spalling occurs as indicated in the high-speed forms of instrumentation into the models.
frames presented in Fig. 9. The radial cracking and the Of particular interest are the results obtained from
circumferential cracking which occurs between the two tests. These were constructed with velocity gage
borehole and free face further weakens the material in loops designed to measure vertical velocities in one test
the 90 ° sector immediately above the borehote. This and radial velocities in the second test. The two tests
sector is very weak when compared to the material in the were conducted with spherical charges of I g of PETN.
other three sectors. As time passes the residual gas With the low-porosity, low-compaction, high-strength
pressure acts on the material surrounding the borehole and rock-like fracturing of the rock stimulant the tests
and begins to move the material in the direction of provide a good database for predicting quantitatively
greatest weakness. This direction is in the 9 0 sector rock motion.
identified earlier and the radial cracks which define that Figure ! 1 shows the geometry of the model used for
sector are filled with gas pressure and grow to the free the two cement tests and the placement of the gages. The
surface as the material in the sector is thrown from the figure also depicts how we envision the 2-D photoelastic
 FOURNEY et al.: CRATER BLASTING 423

Fig. 10. Photographs taken very late in the dynamic event showing the radial fractures that define the final crater shape.

models were taken from the 3-D tests. The 3-D models cylindrical charge were used for comparison purposes
were cylindrical in shape with a diameter of 254 m m and for these two gages.
a height of 208 mm. The center of the charge was Typical results obtained from the gages are shown in
57.1 m m from the upper free face. The locations of the Fig. 12. In this figure results are given for radial velocity
six velocity gages used in each test are given in Fig. 11 b for the two gages nearest the charge for both a cylindri-
which shows only the upper-right-hand quarter of the cal charge test (ROC-7) and the spherical charge (ROC-
model. Test ROC-9 was used to measure vertical velocity 10). As shown in the figure, the initial velocity for both
and ROC-10 for radial velocity. All six gages were tests is outward and some rebound occurs at about
successfully recorded in test ROC-10. In ROC-9 gages 1 30/~sec which results in a short period of inward vel-
and 4 were lost due to amplifier malfunction. The results ocity. The figure also points out that there are slightly
from gages similarly located in a model which used a higher radial velocities in the case of a cylindrical charge
424 FOURNEY el al.: CRATER BLASTING

(o) Top CylindricalModel Side

/,/ ~Charge
i/
~-)
\,

//
PlateModel

I
(b)
ToP FREE SURFACE

EMG - 3 EMG - 5 EMG -6

, .

Igm PETN ~ I m
g.NG-Z ~ W
0
0

,:o-,\I ':: J/ >-

rr
0
g
0
m
VEIOCI TY GAGES w
n

CHt~RGE 2B. 4
mm
Igm PETN

Fig. 11. (a) Visualization of how 2-D models were "taken" from 3-D models. (b) Location of velocity gages in super cement
models.

than for a spherical one. As can be observed from the 3-D tests described above in the following fashion. The
figure the results obtained for the cylindrical charge were 2-D plane stress models tested photoelastically were
not greatly different from the results obtained from the assumed to have been taken from the 3-D tests as shown
spherical charges. in Fig. l la. As shown in that figure the 2-D model
The velocity data recorded was integrated over time to represents a section taken from the 3-D model through
produce displacements in the radial and vertical direc- the borehole.
tions and these displacements are shown in Fig. 13. In The fragmentation pattern that existed at 50 #sec in
Fig. 13 the vertical and radial components have been the high-speed photoelastic models was taken as the
plotted together to yield particle trajectories. The trajec- starting geometry for a static finite element analysis. This
tories given for gages 1 and 4 in Fig. 13 are really the fractured and fragmented model was loaded with a static
trajectories for the radial displacement of a spherical pressure of 1600 psi. The magnitude of the pressure was
charge and a vertical displacement for a cylindrical charge. determined from an examination of pressure vs time
In order to determine if the fragmentation mechanism determined in similar 3-D testing configurations which
proposed is valid it was assumed that the 2-D models had no stemming and was free to vent to the
tested using dynamic photoelasticity were related to the atmosphere. One such curve is shown in Fig. 14. The
 F O U R N E Y et al.: C R A T E R BLASTING 425

A F O S R O C K - VELOCITY G A G E DATA
ROC-7 (CYLINDRICAL)VS ROC-10 (SPHERICAL)
50

40
ROC-7._ EMG-1
_

~ ROC-10 EMG-1
E 30

o~ 2o
LLI
>
-J
<
: . .~"" ROC-10 EMG-2
...."".~;..,.~~ "'\; ~
,, I I ...'
.I.I/
: :.f
'
".~,~,
%'~.'~ii,.
,,~

" ' " . . . . ' ~ V'. j.je...".,;/"

%~o~o,
°°°°°°,i
-10 I I I I I I I I I I I
0 10 20 30 40 50 60 70 80 90 100
TIME (microsec)
Fig. 12. Radial velocities from concrete models with spherical and cylindrical charges.

DISPLACEMENT PATH 50X Scale

.J
A,?(ISOF SYMMETRY
"-• t/,,,'-- TOP =
FREE SURFACE

,;--~.~,= - + - ~.~,om
' @ ~.~,=
" /
~ T EMG-~// E~G-~ EMG-6

A 1.0 - 'T
,,
I
CYLINDRICALSIDE

z
w

5
08

06
1O 7 r

PETN CHARGE - 1.0 gm

EMG-~

u)_ 0.4-
o
.-I
AFOSROCK-GAGEHODOGRAPH
0.2-
n-
ROC-10 (RADIAL) + ROC-9 ~ERTICAL)
o I I I I
0 0.2 0.4 0,6 0.8 1.0
Fig. 13. Particle motion t r ~ t o r i e s obtained ~ o m ~ m e n t models with spherical charges.
426 FOURNEY ela/.: CRATER BLASTING

2400"

80o

0 i I i I i I i I I I '
0 30 60 90 120 450 180 210 240
Fig. 14. Pressure profile from model tests. This trace was used to determine pressure to be used in computational effort.

model which was used to determine that pressure-time 254 × 177 x 6.4 m m thick. M A R C 4, a commercial finite
curve was unstemmed and was fragmented during the element program was used to obtain the solution. Figure
test and therefore vented both through the borehole and 15b presents an enlarged view of the borehole area of the
through the resulting fractures. The pressure value model. The curved boundary at the top of Fig. 15a
chosen is therefore felt to be representative of what represents the boundary of the spalled region as observed
would have been found if the pressure in the borehole of in the photoelastic tests. It is assumed therefore than the
the photoelastic tests had been measured. material between the spall front and the original straight
Figure 15 shows the finite element grid used to free surface is incapable of carrying any load. The height
represent the model. Figure 15a shows half of the of this non-load carrying area is 10 m m in the finite
assumed symmetric model. The solution obtained element computation. It is also assumed that 16 radial
is a plane stress solution with a plate size of cracks have propagated to a distance of 17 m m from

Fig. 15a. Finite element grid enlarged showing upper quarter of model.
 FOURNEY et al.: CRATER BLASTING 427

Fig. 15b. Finite element grid in vicinity of borehole.

the borehole center at the time that the pressure is when the strongest wave activity was finished. It was felt
applied. that if good correlation could be shown between these
In addition, for the seven cracks around the top of the finite element computations and the results of the testing
borehole it is assumed that circumferential cracking has in the 3-D cement models it would add credibility to the
occurred. Note that there is no circumferential cracking proposed mechanism of fragmentation.
assumed for the radial cracks located around the bottom A comparison of the results obtained from the finite
o f the borehole. In the finite element computation the element computations and the measurements made in
1600psi pressure was assumed to act on all crack the "super" cement are shown in Fig. 16. The lines
surfaces--radial and circumferential. The material represent the results obtained from integration o f the
properties used in the finite element computation were velocity gage data and show the displacements over the
those of Homolite 100. In particular, a modulus of first 100/~sec after detonation of the charge. The arrow
500,000 psi and a Poisson's ratio of 0.2 were used with represents the vector displacement obtained from the
a borehole diameter of 5 mm to calculate displacements finite element computation for the static pressure loading
at the six locations where the velocity gages were placed of the fragmented model with 16 radial cracks and short
in tests ROC-9 and ROC-10. circumferential cracks at the tips of the radial cracks
T o summarize, the finite element computation that around the top of the borehole. Since the material
was carried out was a static one and acted on a geometry properties used in the finite element computations were
which corresponded to the fragmentation pattern that for Homolite and the tests were conducted in cement,
was seen in the dynamic tests at a time corresponding to some modification had to be made to reflect these
RMMS 30/4~H
 Displacements at gage location 3 Displacements at 9o9e location 5 Displacements at gage location 6
0.8 0.i 0. 0¢

0.7 0.; 0.
11vleosumd
0.6 0.i
0.5 Calculated 0.'
0.4 0.4 0.,
Measured
0.3 0..~
Colculoted
0.2 0.; Meosured
0.1 0.'
~___~.._~..~ Colculoted
0 C (
--0.1 -0.1 --~).
I I I I i I i I , I
-0.2 -o.:, ~L
0.2 0.4 0.6 0.8 1.0 1.2 0 0.2 0:4 ' C}.6 ' 0:8 1.0 1.~ 0 ' a2 ' 0:4 ' 0'.6 * 0.8 1.0 t.2
Radial dl I (ram) Radial displacement (ram) Radial displacement (ram)
Displacements at gage location 2 Displacements at gage location 4
0.8 0.E
©
0.7 0.7 C
~o
0.6 O.E Z
Colculate<l m
0.5 0.~ .<
0.4 0.4
a3 0.3
O.2 0.,~ ¢3
0.1 >
0.1
0
-0.1
j -0.1
( ~o
o:
i i i
>
-0.2 -0.~
02 0., 06 ;8 ,o ,~ o 0.~ 0, 0.e ;8 ,o ,~
I~ldlal displacement (ram) Radial displacement (ram)
Oispl~ a t 9090 Ioc:oUon I
0.8
0.7
0.6
0.5
0.4
0.3 Calculated
0.2
0.1
0
-0.1
-0.2 i
0,2 0.4 0.6 0.8 1.0 1.2
Radial disl~tmmont (mml
Fig. 16. Comparison of experimental results with predictions from finite element model coupled to results from 2-D tests.
 FOURNEY et al.: CRATER BLASTING 429

differences in properties. In order to correct for this, the geometry when acted upon by residual long-time gas
magnitude of the displacement computed for gage 3 was pressurization describes very well the fracturing which
set equal to the magnitude of displacement measured in occurs in crater blasting. This mechanism, even though
the cement at 100/~sec. The finite element calculations fracturing does begin at the free face and works its way
represent the conditions in the dynamic tests at 50 # sec. back towards the charge, is defined by radial (conical)
As can be seen from Fig. 16, the agreement is good-- cracks which originate at the borehole and propagate to
especially at early times for gages 2-5. The finite element the free face.
result agrees very well with the displacement measured
at gage 1 at later times but the agreement is not very Acknowledgements--This work was conducted under Contract
good with gage 6--at any time. Looking at the overall AFOSR-88-0280. Dr Spencer Wu at AFOSR was our technical
monitor during the contract period and we appreciate his support and
trend the measured velocity at gage 3 (the gage above the guidance.
charge and nearest the free surface) has a much higher
vertical component of displacement than predicted by Accepted for publication 30 December 1992.
the computation (and the fragmentation mechanism).
For all of the other gage locations the finite element REFERENCES
prediction is quite good except for gage 6 (the gage
farthest from the borehole and closest to the free sur- !. Dally J. W. and Riley W. F. Stress wave propagation in a half
plane due to a transient point load. Dev. Theoret. Appl. Mech. 8
face). For that location the measured displacement also (1966).
had a larger vertical component than the finite element 2. Porter D. D. A role of the borehole pressure in blasting: the
prediction. formation of cracks. Ph.D. Thesis, University of Minnesota,
Minneapolis (1979).
Overall the agreement is taken to be a good colla- 3. Irwin G. W. et al. On the determination of the a - K relationship
boration of the fragmentation mechanisms being for birefringent polymers. Expl Mech. 19, 4 (1979).
proposed. It therefore appears that the stress wave 4. Barker D. B. and Fourney W. L. Photoelastic investigation of
fragmentation mechanisms, Pan ll--Flaw initiated network.
action is quite important in crater blasting. The National Science Foundation Report (1978).
contribution of the very first mechanism proposed in the 5. Barker D. B., Fourney W. L. and Dally J. W. Photoelastic
early 1960s (spalling) is quite important. When coupled investigation of fragmentation mechanisms, Part I--Borehole
crack network. National Science Foundation Report (1978).
with the combination of radial cracking and circumfer- 6. Fourney W. L., Dick R. D. and Simha K. R. Y. Model study of
ential barrier branching the preconditioned rock crater blasting. Rock Mech. Rock Engng 21, 183-2005 (1988).