Theme 3

Rock blasting and excavation

Blast vibration analysis, response and control; drilling technology and control; fragmentation;
and road header machines; large diameter shaft drilling

08:30

LE GRAND

Chairman:
Session

SALON

Tan Tjong-Kie
Secretary:

Speaker:

C.K.

- September

Mackenzie

(Canada)

(Australia)

Blasting in Hard Rock:

Techniques
for Diagnosis
Modelling
for Damage and Fragmentation.

and

ABSTRACT:
Models for prediction of rock responses
to blasting have been developed to evaluate
alternative
blast designs for specific applications
and rock types.
Design optimisation,
however,
further requires monitoring
and measurement
to
ensure control over detonation
sequencing,
and to
identify the more important influences controlling
rock response.

1.

100

XNnlODUCTXON

A model of rock fragmentation

by blasting has been
developed by the Julius Kruttschnitt
Mineral
Research Centre in Australia, based on the results
of measurements
and monitoring
at more than 25 mine
sites, totalling more than 200 weeks of site work
over the past three years. The model has permitted
the definition
of rock damage from blasting, and is
being used to optimise blast design for the
achievement
of specific blasting objectives.
The model has the primary feature of directly
utilising fracture data from fracture mapping and
stereoplots,
and identifies the existing fracturing
as the primary influence on both fragmentation
and
damage.
Perhaps of equal importance is the
development
of monitoring
and measurement
techniques
used for the collection of data required
by the model. The use of these monitoring
techniques
has highlighted
some of the basic faults
in blasting,
including lack of control over charge
initiation,
variability
in explosive and initiator
performance,
lack of detailed design ass~ssment
procedures,
and an inadequate understanding
of the
mechanisms
of damage and the factors controlling
damage.
In most applications
to current
Operations,
modelling
is considered to be of
secondary importance to basic control - over
charging procedures,
design, and product
specifications.
The results of some basic blast monitoring
are
presented to emphasise the application
of
monitoring
to the optimisation
of blast
performance.
The improvements
to productivity
reSUlting from the monitoring
have been
Substantial,
and beyond the current scope of
mOdelling.
Mines have reported savings in excess
of $2 million per year in explosive and drilling
costs alone, with powder factors being reduced by
up to 50%, while improving fragmentation,
reducing
damage, and controlling
ground vibration.
2

of full face boring

on the following observations

1. fracturing has the largest single influence on
fragmentation
and damage ;
2. the degree of breakage of a rock is dependent on
the size of the particle and the energy imparted to
it ;
3. during blasting, breakage or fracturing
is
primarily produced by shock energy or brisance,
and
rock displacement
is produced by gas energy.
The basic assumption
of this meChanistic
model of
breakage can therefore be stated : if the size of
particle and the input energy to it are known, then
the degree of breakage of the particle can be
predicted.
This follows from observations
of the
breakage response of thousands of rock particles
in
laboratory breakage tests.
Perhaps the primary
feature of the model is that it is based on
measurements
rather than theoretical
rock or
fracture properties,
and that all predictions
of
the model can be verified through measurement.

2, 1987

(China)

W. Comeau

performance

I'RAGMD!TATXON

The JKMRC

63

2.1

and damage

model

Fracture

8000

(mm)

Figure 1.
Size distribution
before an4
blasting, showing extent of breakage.

after

"'as~t

The intensity of fracturing is measured by line

scanning techniques,
whereby the points of
intersections
of all natural fractures with a scan
line are measured, together with the dip and dip
directions
of the fractures.
These fractures are
then represented
on an equal-area
stereo plot and a
Monte Carlo simulation used to calculate insitu
block size distributions
for three different
joint
spacings distributions.
The rock is therefore
described as a system of irregularly
shaped blocks,
with a size distribution
of blocks being defined by
the natural macro-fracturing.
The fracture measurement
technique immediately
indicates the maximum size of particle which could
be observed in the muckpile, and the percentage
of
particles greater than any particular
size.
If the
desired muckpile fragment size distribution
can be
defined, a comparison
of the two distributions
can
be made as shown in Figure 1. This comparison
provides an immediate indication of the degree of

MODELLING

fragmentation

!lOO
SIZE

is based

1425

by increasing both burden and spacing, reducing the

powder factor by 20% using the standard hole
diameter of l65mm. Figure 2 presents the agreement
between the observed and predicted fragmentation
using the parameters determined from previous
blasting measurements.

breakage required by the explosive, from which a

"breakage index" could be defined.
2.2

Di.~xibu~ion o~ Shock &nervY

Shock energy is measured as ground vibration using

velocity gauges (geophones). An array of gauges is
located around the blast, and the information
recorded either digitally or on analogue tape
recorders. Full vibration waveforms are used to
calculate energy at a point from knowledge of the
rock density, the p-wave velocity of the rock, and
the particle velocity time history at the point of
interest. The measurements of particle velocity
are made close to the blast (from 5m to 25m) to
avoid excessive errors in the calculation of
vibration energy due to extrapolation.
From multiple recordings of particle velocities,
the vibration propagation equations can be defined
for the particular rock/explosive combination in
the blast. The equations used at thElJKHRC take
both frictional and geometric attenuation
mechanisms into account for spherical waves, and
permit detailed extrapolations concerning the shape
and amplitude of ground vibrations at any location.
Through measurement and calculation it is therefore
possible to determine the shock or vibration energy
at any location in or around a blast, and to
establish contours of ground vibration around a
blasthole pattern. It is ths vibration energy
which is considered responsible for the
introduction of new fractures or breakage of the
rock mass.

100.0
00.0
10.0

~
~

70.0

;;

00.0

~
~
5
~

00':>

"'.0
"'.0
20.0
'0.0
00

Fxagmen~.~ion

260

DOD
SIZE

2000

4CIOO

e",ml

riguxe 2. Predicted and observed size

distributions from fragmentation model applied to
actual blast designs.

DalIIageKeasu~~

and De~ini~ion

The mechanistic modelling approach can clearly give

accurate predictions of fragmentation, and can
easily be extended to predict damage. Damage can
now be defined as the ohange in the insitu block
size distribution due to adjacent blasting, and can
be related to the "breakage index" concept
presented earlier, ie. the shift in the pre- and
post-blast block size distributions. Depending on
the application of the measurements of damage, this
definition mayor may not be adequate, highlighting
the need for a more explicit definition of damage.
The definition used above is adequate to compare
the effects of different blast designs on backcracking, or overbreak, but gives little indication
of the reSUlting loss of stability which may
ultimately lead to failure, ore dilution and lost
recovery.

Ke uxe n~

Any program to predict blast fragmentation must be

verified by a measurement of fragmentation.
One
technique being used at the JKHRC is a photographic
process where many photographs of a muckpile are
taken at various' stages of excavation, and all
discernable particles in the photograph digitised
manually to obtain basic dimensions such as area,
perimeter, maximum and minimum projected chord
lengths.
Extensive calibrations have developed a
relationship between the digitised measurements and
the particle size, permitting a size distribution
to be calculated. The technique has been shown to
accurately
describe
observed changes in size
distribution, but is extremely labour intensive,
requiring up to 1 hour to digitise several hundred
particles in a single photograph.
The measurement of fragmentation is becoming more
and more necessary in modern blasting wliere
specific fragmentation objectives are being sought
to permit the application of more cost effective_
mining technology such as continuous mucking,
mobile crushing units, and conveyor transport of
materials after blasting. Although it is possible
to define the required fragmentation for these
applications, it appears that it is not practically
feasible to measure fragmentation so that mining
engineers can further optimise the process of
blasting and primary size reduction. This is
clearly one area where computer aided analysis
could be used, if the problema of boundary
discrimination can be overcome.

2..

'2!

ICJtI:IlNIN.

2.5
2.3

,.

TRB

L%K%TAT%ONS

OF

KODII:LL%NG

Predictive modelling, in contrast to back analysis,

must make assumptions concerning the behaviour of
explosive and the accuracy of blast design
specifications such as blast-hole diameter, hole
accuracy, pattern size, etc. Without some form of
monitoring programme, the real performance of a
blast can not be known,'and the real causes of the
observed performance can not be determined.
Current modelling of explosive rock fragmentation
assumes conformability between actual and designed
charge performance. This will be shown to be a
major limitation of modelling, and a major
impediment to blast design optimisation programmes.
The advances made in modelling and computing are
lost if blast initiation control can not be
strictly maintained. Simple monitoring of charge
performance, and the ability to detect initiation
malfunctions, can lead to far greater gains in cost
effective blasting than predictive modelling.
Whereas mathematically it is possible to vary
blasthole diameters, burdens, spacings, explosive
strengths, and stemming lengths independently over
wide ranges, in practice this can easily lead to
sympathetic charge initiation, charge
desensitisation, or charge dislocation, which can
not be detected without careful monitoring. Delay
intervals can also be theoretically modified over a
wide range, but modelling does not consider the
effects of an inappropriate interval on sequence
reversal, vibration enhancement, or dislocation of
adjacent charges.

Model P~c~ions

Using all of the above techniques, it therefore

becomes possible to model the process of explosive
breakage, using field observations and,measurement
to determine the model parameters. Several normal
blasts were monitored, in terms of vibration
response, muckpile fragmentation, and pre-blast
fracture distribution and the results used to
determine values for critical parameters. Blast
design was then modified and the model used to
predict the changes in muckpile fragmentation from
two different design configurations. In the first,
blast hole diameter was decreased from 165mm to
l14mm, with powder factor being held constant by
reducing both burden and spacing proportionately.
In the second experiment, powder factor was reduced

1426

4.

4.1

BLAST MONITORING TECHNIQtlBS

Instruments located near to a blast can provide

clear records of the detonation of explosive
charges. Figure 3 shows the response to a single
charge, recorded from a geophone located
approximately 20 metres from the charge and
permanently grouted to the rock in a borehole. The
precise initiation time, initiation error, and
vibration amplitude can easily be measured from
this record. Initiation timing provides
information concerning sequencing, misfires and
delay scatter, while vibration amplitude provides
information concerning performance of explosive and
energy available for breakage.
Figure 4 presents the response to a multi-delayed
blast, showing also the blast design. In this
example, the charges are decked, ie. several
charges within the one blasthole, and the nominal
and actual initiation times for-each deck are
shown. Clearly shown is the scatter in the
initiation time for the two 300 ms charges.
Despite the delay errors, these charges are
performing as well as can be expected - observed
delay scatter is within expected limits and all
charges have initiated with strong vibration
response.

1.0

-1.0

YJIg

(KI)

rigure 3. Vibration response to a single

explosive charge, recorded within 20 m of a
blasthole.

300
"'10

"'7

175

BLAST DESIGN ASSESSMENT

Blast performance monitoring has amply demonstrated

the importance of maintaining tight control over
the distribution and concentration of explosive.
However, even a low average powder factor can
produce localised initiation faults indicative of
overblasting. Simple computer algorithms can be
developed to give blast design engineers a better
assessment of a blasthole and charging pattern than
is currently obtained.
Figure
8(a), for example,
shows how the
concentration of explosive in one section of a
blast has been increased to be more than double the
average. In this case, the concentration was caused
by an increase in blasthole diameter from 70mm to
l15mm in order to achieve the increased hole length
without loss of drilling accuracy. In this section
of the blast, sympathetic detonation of charges is
more likely to occur, resulting in overbreak and
increased damage.
Figure 8 (b) shows how this
charge concentration can be reduced by using a
lower strength explosive in the larger diameter
holes.
Similar applications to fan,patterns in
underground blast designs have resulted in pOWder
factor reductions in excess of 10%, by adjusting
the lengths of explosive columns in blastholes.
Simple design assessment techniques such as these

ts

SCALE(m)

#8

200

rigure 4.
blast.

Malfunctions

Figure 5 presents three of the most common forms of

charge malfunction which have been observed to
occur commonly in blasts throughout Australia,
Canada, USA and Chile. The extent of charge
miSfires is strongly influenced by the complexity
of blast design, number of delayed charges,
duration of blast, and geological structure.
Overcharging is the most common cause of initiation
problems, poor fragmentation, and excessive
overbreak. A common response to poor fragmentation
is to increase powder factor - but this may
aggravate the situation by promoting misfires and
increasing overbreak which tends to be blocky as
defined by pre-existing fractures.
The extent to which misfires and other initiation
problems can be present is shown in Figure 6, where
only 9 out of a total of 20 charge have initiated
as designed. Of the total number of charges, 20%
failed to initiate, and 35% initiated
sympathetically with delays 5 and 8, producing 2
tight bunchings of vibration pulses.
Once identified, charge malfunctions can be
eliminated, restoring initiation control and
invariably using less explosive. Figure 7 presents
a well controlled blast which resulted from
modification of the blast design used in the record
of Figure 6. Misfires have been eliminated, the
number of small deck charges has been reduced,
powder factor has been reduced by approximately
50%, and vibration levels, damage and overbreak
have been reduced.
A fourth common charge malfunction is the reduced
performance of charges, and relates partiCUlarly to
the application of emulsion based explosives.
These explosives have frequently been observed to
produce variable vibration levels, variable
fragmentation, and variable muckpile displacement,
and have also been commonly observed to misfire.
These features suggest that the explosives are not
always detonating at maximum or designed rates, and
that the energy available for breakage is therefore
significantly reduced. For the emulsion-based
explosives to maintain their cost effectiveness,
they must consistently achieve rated performance
levels.
In the light of the evidence concerning the
extent to which initiation problems are occurring
in production, and development or tunnel blasting,
instrumentation to quickly identify the problems
would appear to have greater application to blast
design optimisation than modelling. Blast
performance monitoring has frequently reduced
explosive consumption by around 50%, and averages
approximately 25% with improved control over
fragmentation, vibration and damage.
S.

~~o

Initiation

Vibration response to a multi-delayed

1427

II 4080%

also offer considerable scope for blast design

optimisation beyond the scope of current models.
Future modelling should be able to determine the
point at which sympat-hetic detonation of charges
and charge disruption are likely to occur.

SYMPATHETIC

Desigrd

sequence

-,

80

120%

120 . 160%

160

200%

> 200%

Nominal cell powder factor.

Nom = 0.20kg/t

riqur. 8(a). Concentration of explosive during

loading caused by increased hole diameter.

300

200

100
TIME

(ms)

INSTANTANEOUS

r-

Designed sequence

11

U1
Tf

--,

I.
It

If

40 - 80%

80 - 120%

120 - 160%

Nominal cell powder factor

Nom = 0.20kg/t

400

200

600

1000

800

riqur.8(b).
Even distribution of explosive
obtained by adjusting explosive strength in large
diameter holes.

TIME(ms)

6.

11.
U
o

100

200

300

400

SOO

600

700

TIME (ms)

riqure 5. Common forms of explosive initiation

malfunctions - actual records.

1<

TIME (-)

"

riqure 6. Extent to which misfires can be present

- actual underground blast record.
----!

DELAYS

ton

.>11

NOMI'lAL

.. ,
t
r

4g()

INITIATION TIMES

570 650

725 800

6.1

.
~

~5
TIME

In~luenc. o~ Delay Scatt.r on Blast Design

From knowledge of delay scatter, the probabilities

of delay sequence reversals can be evaluated. This
probability is an additional assessment of the
level of control over the correct initiation of a
blast pattern.
Delays are used in blasting to achieve one of two
specific objectives :
1 to ensure that charges do not initiate together;
2. to promote the instantaneous
initiation of
charges.
Using data describing delay scatter, the
probabilities of achieving these objectives for any
situation can be assessed. Consider the example of
a tunnel or development round shown in Figure 10
which presents the hole pattern and delay
sequencing used in a typical underground round. In
particular, the analysis will consider the

Delay scatter is a well accepted feature of

pyrotechnic delays, though blast design programs
and models rarely appear to account for the
phenomenon. Several thousand non-electric delays
have been studied to determine their precise
scatter patterns, each timed electronically, and
many supported by high speed photography. The
studies were conducted on products from all
Australian suppliers, covering most delay ranges
from 25ms to approximately lOs.
The results indicate that although scatter within
an individual batch is very low (approximately 3 to
4%), the scatter between batches is much higher.
This means that when only one batch and one delay
number is used, the delay scatter may be considered
to be minimal and around 3 or 4%, but if more than
one delay number or batch is used, then.the delay
scatter will increase to a much higher level, up to
13%, depending on the product. Figure 9 shows the
delay scatter for two batches of 500ms delays. The
total scatter when these two batches are mixed is
at least twice the scatter of each individual
batch. From testing up to 40 different batches,
scatter due to batch mixing can increase by a
factor of up to 3, although this is dependent on
the product. For the purposes of this paper, a
delay scatter of 13% will be used to describe the
distribution of delay error, ie, 95% of delays will
detonate within 13% of the nominal or quoted delay
time. Experimentally, this has been demonstrated to
be a realistic definition of delay scatter.

ll:26

DELAY SCA'l'TD

,~
lS)

riqur. 7. Controlled initiation of charges, free

from misfires - actual underground blast record.

1428

rock types, and being much greater for low modulus

plastic rock types.
The period can be related to '
t~e.induced
vibration response of the rock.
The
m1n~um
vibration interval is the time after
detonation before the rock returns to a near-zero
vibration st~te.a~ shown in Figure 3. Vibrations
~roduced by 1nd1v1dual charge detonations
are not
1nfluenced by any other charges, and vibration
levels are well controlled.
Vibration level prediction using the above
concepts can the~ consider the probability
of
enhance~nt
of v1brations from different charges by
evaluat1ng the probability
that delay scatter will
~ause charges to initiate within any particular
1nterval, determined from vibration monitoring
using instrumentation
capable of recording the full
time history of ground vibration at any location.
This approach, for example, has been used to
simplify blast design where large numbers of
delayed deck charges are involved.
Probabilistic
analysis has been used to compare designs
containing large numbers of deck charges with
11
delay intervals, against a reduced number of d:~s
(larger charge weights) with large delay intervals.
In many cases, the larger charge weights and longer
delay i~tervals can be demonstrated
to produce
lower v1brations,
less variability
in ground
vibration
(improved vibration control), and
improved ease and speed of charging.

following probabilities
:
1. that the perimeter holes sharing a common delay
will interact to promote smooth blasting and smooth
walls to the drive ;
2. that the perimeter holes will initiate after the
inner holes have fired and cleared, minimising
damage to the walls and back.
For adjacent blast holes to interact to promote
smooth blasting, they must initiate within a
critical interval, determined by the crack
propagational
velocity of the rock. Assuming a pwave velocity of approximately
6000m/s, a crack
propagational
velocity of one-third of the p-wave
velocity, and a blast-hole
separation of no more
than 1 metre, then adjacent holes must initiate
within 0.5 ma of each other to achieve any
interaction
between charges and to achieve a smooth
blasting effect.
Figure 11 shows the probability
of interaction
as a function of delay time, for a
delay scatter of approximately
13%.
For the high
order detonators
around the tunnel perimeter, the
probability
that adjacent holes will detonate
within 0.5ma is effectively
zero, and the
probability
that they will detonate within 5 ma is
less than 2%. There will therefore be no smooth
blasting effect in tunnel blasting using high
order, long period delays.
A similar analysis of smooth blasting along stope
walls using millisecond
delays is shown in Figure
12, where the easer hole represents an intermediate
hole between rows, on the stope periphery.
Ideally
this hole should detonate at the same time as the
wall hole in the row behind to promote a smooth
stope wall.
The probability
of achieving this
effect is shown to be approximately
3% for 600ms
delays, and increases to only 10% for lOOms delays.
The probability
of sequence reversals and
perimeter holes initiating before the adjacent
inner holes has also been investigated.
Figure 13
Shows the overlap probability
function for the
delays used in the blast, with a delay scatter of
approximately
13%.
The probability
refers to the
overlap potential between successive delay numbers.
A maximum overlap probability
of approximately
14%
is shown for delay periods around 6 seconds. There
is therefore a reasonably high probability that a
perimeter hole will initiate before an inner hole,
promoting wall damage, irregular drives, and
increased ground support costs.
By ensuring that
perimeter holes are separated from inner holes by
at least two intervals, the probability
of overlap
decreases to effectively
zero.
Placing charges on separate delays does not
necessarily
ensure that they will initiate
separately or in the desired sequence, particularly
When standard intervals have been reduced by
Combining delays.
Models of blasting must first
be able to account for the influence of delay
intervals on fragmentation,
and secondly, account
for the influence of out-of-sequence
initiations on
fragmentation
and damage.

6.2

Zttect

ot Delay

Scatter

on Vibration

8
7

Error

o
-1
-2
-3

0.'
-0.'
1

E~o, .1..
-2

2.'
-3
-3.5

-.

Figure g.
detonators.

Control

One of the main reasons for using delay~ in modern

blasting is to control and minimise.vibrations
and
Vibration-induced
damage.
For this reason, many
modern long hole stoping operations are using
mUltiple charges within each blasthole, each charge
separated by inert stemming and separately delayed.
This leads to large numbers of delays required to
achieve desired levels of production tonnage per
blast.
Since standard delay series are limited in
size to no more than approximately
40 delays, it is
Common practice to extend the available series by
COmbining delays, using a collar delay at the top
of the hole to initiate the downline to a second
delay at the bottom of the blasthole.
This
increases the number of available delays but
decreases the interval between delays and therefore
increases the probability
of overlap between
successive delays.
Currently available vibration prediction
equations relate to the maximum charge weight per
delay, and place little emphasis on defining the
minimum period of the delay.
In practice, this
period is controlled totally by the rock type,
being relatively short for high modulus, elastic

Delay

scatter

for different

batches

of
7

17

15

I,
14

15

8
12

10

II

\I

13
18

17

10

16

16

16

15

,,~

0.0
0 0

.4

9.

II

12

~
~o

l
8

10

Figure 10.
Development
blast pattern and
sequencing design for 52 mm hole diameter.

1429

18

100

7.

<D

0'5ms NTERACTION PERIOD

~ 5- Om. INTERACTION PERIOD

XX>

6000

Observations of blasting have indicated that high

pressure explosive gas products penetrate along
fractures in all directions around a blasthole.
The forward penetrating gases promote movement and
swell of the muckpile, but the backward penetrating
gases cause swelling behind the blast and the
subsequent loss of friction along fracture
surfaces, decreasing stability. The influence of
the gas is strongly dependent on the state of
fracture of the rock mass.
Simple experiments have been conducted to
demonstrate the flow of gas behind blasts, yielding
information concerning the velocity of propagation,
the timing of burden movement, and its dependence
on blast design variables. In these experiments,
boreholes are drilled behind the blast pattern,
pressure gauges are inserted, and the top of the
borehole is sealed with a concrete plug. Adjacent
to, or even in the borehole, is a vibration gauge
which is used to determine the precise initiation
time of the blasthole near the pressure gauge.
Figure 14 shows the pressure response in a
borehole located 6 metres from the end of a presplit line of blastholes. The gauge, which has
been vibration compensated, still shows a small
vibration response as the nearest blasthole
detonates, and then shows a sudden increase in
pressure followed by a gradual release. In this
experiment, the high pressure gas path has clearly
been along the pre-split line which has extended
beyond the last blasthole, rather than through
natural fractures.

1000O

TIMElms)
5

12

10

14

15

17

DELAY NUMBER

riguze 11. Probability function describing

potential for constructive interaction between
adjacent peripheral holes in a development blast.

GH4t

1Il!S'"Jll!S.

.CI!I!L __

Is"' 1

PREllIUTLNE

PlAN FOR PRE-SPUT

10000

100

CD
~

0'5 INTERACTION
50 INTERACTION

PERIOO
PER100

00000

TIME (SEC!
PRESSURE

100

200

300

400
T1MEIIllll

eoo

riguze 14. Gas pressure response in uncharged

borehole located 6 m from the end of a pre-split
line of blastholes.

700

riguze 12. Probability function describing

potential for constructive interaction between
adjacent peripheral holes in an open stope blast.

In a second experiment, gas pressure and

vibration monitoring were conducted behind a normal
overburden blast in a strip coal mine, producing
the responses shown in Figure 15. The blast was
fired in rows parallel to the free face, in five
rows, with the vibration response from each row
clearly shown on the vibration record from the
radial guage in the 8M 11 hole. These vibration
peaks can also be seen on the gas pressure records
of the other two signals, indicating tht the piezoelectric gauges are not fUlly vibration
compensated. Also clearly visible, however, is the
gas pressure response in the GM 7 hole, showing
first an increase in pressure followed by a
decrease resulting in a negative borehole pressure.
Final return to ambient pressure takes
approximately 1 second after gas first started
flowing into the borehole. Gas propagation
velocities have been calculated from many
experiments to be approximately 8 to 12
metres/second.
The negative pressure indicates that the
blasthole volume is increasing, due probably to a
release-of-load relaxation as the burden in front
of the last blasthole moves away from the bench
face. Measurements of burden movement times based
on this assumption indicate that the rock takes

14

12

~
a::

w 10
>
0
u,
0

j
a::
a.

40

IllXR) R)R PRE-SFL'T

8
6
4
2

2000
4000
6000
8000
NOMINAL INITIATION TIME (ms)
riguze 13. Probability function describing
potential for overlap between successive delay
numbers, likely to produce out-of-sequence
initiation of wall charges in tunnel blasting.

1430

PRESSURE& VI BRATIONRECORDSFOR
THROW BLAST 35-4a

FREEFACE
0

100<>

PRESS

0<Pal

-50-0
0'0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
0

GM7 "GH8
..t
0

8M"
PlAN OF lHROW BLAST 3S-<Ia

20
GH8

8M Il RADIAL

lOCO

PRESS
(KPa)
00

0
0

0
0

1000

GH7

PPV

!mm}sO

/0
TIME (SEC)

10
TIME (SEC.)

rigu~
15. Gas pressure response in uncharged
production blast in heavily fractured rock.

borehole located behind a

approximately 90 ms before'any burden movement

Occurs after detonation of the explosive charges.
Gas penetration has been measured at distances of
8 metres behind a blast, and is proposed as a major
cause of damage and lost stability. The
penetration of gas can be minimised through :
1. minimisation of the time over which the pressure
is acting (maximis~'burden movement velocity) :
2. minimisation of'peak borehole pressure;
3. providing an alternative gas vent such as a
presplit line.
Of these techniques, reduction in the time over
Which the gas pressure is acting is expected to be
the most effective means of controlling gas
penetration. Burden movement velocities can be
maximised, without increasing peak borehole
pressure, by minimising burden dimension. Blast
designs to minimise blast induced damage should
therefore have the following features :
1. few blast rows, promoting unrestricted burden
movement;
2. small burden dimension;
3. pre-split along critical boundaries.
Modelling of blasting must ultimately be able to
predict damage as well as fragmentation. Both
factors are critically dependent on the existing
state of fracturing. Gas pressure plays a critical
role in determining the final muckpile shape, and
also in the extent of gas related damage.
MOdelling must therefore be able to predict the
flow of gas as vell as the volume and pressure of
gas produced in blasting.

ensure that any blast design has the best chance of

producing the desired, or predicted, blast product.

9.20

B. BONAPACE, Austria "Advance and Limits of

TBM-Drivina durina the Excavation of a 22 km Lona
Pressure Tunnel".
J. BRYCH, NGOI NSENGA and XIAO SHAN, Belaium, Zaire,
China "Destructibility of Rocks with Rotary Drill
Bits".
I.W. FARKBR and P. GARRITY, U.K. "Prediction of
Roadheader Cuttina Performance from Fracture
Touahness Consideretions".
GINN HUH, KYUNG WON LBB and HAN UK LIM, Korea
"Determination of Vibration Equation by Empirical
Methods".
FOYER,

Coffee Break.

Y. GOTO, A. KIKUCHI and T. NISHIOKA, Japan "Dynamic

Behaviour of Tunnel Linina due to Adjecent
Blastina"
J.R. GRANT, A.T. SPATHIS and D. BLAIR, Australia "An
Inve.tiaation of the Influence of Charae Lenath upon
Bla.t Vibrations".

8.

Mathematical models have contributed significantly

to our understanding of some of the mechanisms
involved during blasting. However, unlike most
other applications of mathematical models, blasting
models have developed without monitoring and
measurement procedures required to validate
predictions or even confirm the major mechanisms.
Explosives and initiators are assumed to function
exactly as designed or specified, and little
importance is placed on evaluating factors which
may detract from their optimum performance.
Fragmentation and damage, primary factors in
mOdelling exercises, can not be accurately measured
on a large scale. The lack of basic measurement
techniques and instrumentation, which can be used
by the mining industry, is a major limitation to
the application and further development of blasting
models. Instrumentation to monitor the performance
of explosive charges is available and is the bare
minimum of monitorinq instrumentation required to

LB GRAND SALON - Short Presentations

H. HBRAUD and A. RBBBYROTTB, France "Pre.plittina

Test. and Mea.uremants of the Induced Back-Break. in
Granitic Rocks of the French Ma if Central".
T. INAZAKI and Y. TAKAHASHI, Japan "Evaluation of
Rock Ma Quality Utilizina Seismic Tomoaraphy".
N.H. MAERZ, J.A. FRANKLIN, L. ROTHBNBURG and
D.L. COURSEN, Canada, U.S.A. "Measurement of Rock
Fraamentation by Diaital Photoanaly.i.".
A.M. MIRANDA and F.M. HBNDBS, Portuaal "Rock
Weatherina a. a Drillability Parameter".
S. HITANI, T. IWAI and H. ISAHAI, Japan "Relations
between Conditions of Rock Ha and TBH'.
Feasibility"

1431

R.D. SINGH, VlRENDRA SINGH and B.P. KHARE, India "A

Study for Optimization of Fragmentation by Blasting
in a Highly Cleated Thick Coal Seam".
W.H. WILSON and D.C. HOLLOWAY, U.S.A.
"Fragmentation Studies in Instrumented Concrete
Models"
12:00

MARQUETTB,
S ion

Lunch Break and Po.ter

13:30

LB GRAND SALON,
Di.cu ion

Pan.l and Floor

Moderator:

P.A. Lindqvi.t

Panelists:

R.F. Favreau (Canada). O.T. Blindheim

(Norway); C.D. da Gama (Brazil).

Panel contribution:

(Sweden)

C. Dini. da Gama (Brazil)

ABSTRACT:
Six paper. regarding the effect. of rock
bla.ting vibration. and four paper. on the .ubject
of fragmentation con.titut. the kernel of the
author'. contribution for the development of hi.
ta.k a. paneli.t for Theme III of this Congre.s.
After a brief dcription of each paper'. main
a.pect., .pecific comment. about their merit. are
pointed out.
G.n.ral idea. on both topic. are then pre.ented to
complement the author'. view point

'!be article describes the rrethods used in the Soviet

Union to quantify the propagation laws of
blasting
vibrations, and the camonly utilized criteria
to
protect stnlctures fran damagecaused
by
nearby
detonations. For the first aspect, it is
reported
that particle velocity (V) relates to charge
mass
(0) and distance (r) by the fonrulae:
V

r
where K, and K., are numerical coefficients
that
depend On the ~
carxlitions of stnlcture
fOUllda
tions, and on the numberof free faces in the blast-;
respectively.
Values of Kl and K., are given in the paper,
the
first one with aver~e and maxinummagnitudes,
and
the second only with average values. Furthentore, Kl
is indicated for just three ground types, and K2 for
1 to 4 free faces of the :rock mass subjected
to
blasting. '!he Kl and K2 coefficients
are
prOOably
assuming values \41ich eatply with the
dirrensional
oci1erenceof the above iIrlicated expressial, in order
to provide V in ants, after using r in rreters and

in kg.

'!he IrOStcurious characteristic

of that
fonrulae
is the indication that the particle velocity
is
supposable decreasing with the first power of the
distance, in all cases.
As the explosive charge 0 is considered per delay,
the authors proposed another etpirical expression to
detennine the time interval T between
consecutive
charges, as given by:
T

'lhere is no doobt that the subject of klck Blasting

and Excavatia'l deserves an inportant share of m:ldern
klck Mechanics, due to the manytcpics
that
need
further :research, both theoretical and awlied,
in
order to better cootrol field operations
and
to.
lessen incorrect or excessive explosive ocnsunptia'1S,
thus reducing excavation costs.
'rtle carplex mechanismsof :rock disruption
under
explosive action are still a challenge to manyklck
Mechanics experts, who feel the ne=essity to better
understand the time sequence of events that ocntribu
te to reach a certain fragmentation, \41ich
should
be preViously set, aocording to
eccn:mic optimiza
tion criteria.
aJ.t not only engineering design
is
affected by uncarplete or erroneous Jcnowlec1ge on
this subject; the so-called seocrldary effects
fran
:rock blasting (vibrations, fly:rock, airblast,
dust
and gases) also require scientific
cootributions
a1JIedto reduce their detr:!llental COflSe:llJE!llCs.
In the panelist's
task to tackle this subject,
a
oouple of blasting topics were distributed
by
the
General Reporter: :rock vibrations and fragmentation.
Both are full of Ihysical and mechanical interrelated
phe!IClrella,with very short time intervals
of
oocu
rrE!I'O!and with manythings in camon:
in effect-;
\41enblasting design is adequate (perfect balance of
explosive quantities and :rock volumes)
the
0.0
aspects are the best measures of blasting efficiency.
But \41endesign is mistaken or field
operations
go wrong, either the excessive annmt of explosive
causes vibration prd:llems and/or overfragmentation ,
or subcritical explosive weights produce many :rock
bculders, thus revealing the need for corrections of
several types.
'lherefore, rock vibrations and fragmentation
are
key factors for the sucx:ess of any blasting
oper~
tion, either in mining or civil engineering
"lOrits,
as oonf1nned by the review of ten excellent
papers
on these topics.
2

2 W (1.)1/2
q

\41ere Wstands as the shortest distance (in meters)

fran a charge to a free surface, y is the
specific
weight of the blasted :rock (in t/m') and q is
the
pc:lWder
factor, expressed in kg of explosive per m'
of blasted rock. Uponthe utilization
of this
crite
rion, ~les
are given in the paper
for
blasts
ranging fran 320 kg to 11,000 kg of an
undisclosed
type of explosive.
As far as damagecriteria are cancerned,
the
authors suggest (in Table 4 of their paper) a series
of values for limiting vibration
velocities
in
several kinds of structures, fran hospitals
to mine
openings. It is interesting
to note that
those
limits are reduced to half or 1/3 of theft
single
blast values, when they result fran repeated blasts.
'!he article refers only 5 russian publications
on
the subject, the !lOst recent one being of 1976.
2.2

A detennination of vibration equation by eI'lpiri

cal rrethods, by G. Huh, K.W.lee and
H.U. Liiii
(Korea)

'!he need to protect different types of

structures
located in the vicinity of 0.0 sul::May
lines in seoul,
\41oseexcavation was accarplished by rreans of explo
sives, led the authors of this paper to
deterrn:Lne
the propagation laws of blasting vibrations.
Using field data of nore than 100 blasts, they can
sidered the nutual interference
of the
follow1n9
variables:
- distance between the detonation and registration
points;
- weight of detonated explosive per delay;
- type of explosive (3 types);
- blasting pattern (4 kinds);
- :rock types and oorresponding weathering degrees.
'rtle general expression of vibration
proposed as:

PAPERS <J'.l BLASTING VIBRATICH>

2.1

= ~ ~ 01/3

seismic ocntrol of mine and quarry blasting

in
the USSR,by A.B. Fadeev, L.M. Glosman,
M.I.
Kartuzov, and L.V. safon:JV

W3"

K (-0-)

propagation is

\41ere V (an/s) is the particle velocity, W (kg)

is
the weight of explosive detonated per delay,
0 (m)

1432

the distance, am K am n two parameters \oIDichdepeOO

both 00 the explosive and the rock under
considera
tion. Table 6 of the article .indicates the values 01
K for a range of explosives am blasting patterns
,
am for its calculatioo it is proposed the
expre~
sion:

res were detooated, having lengths fran 0.09 m

'!he purpose of this research was to detect the

effect of charge length on the characteristics
of
vibrations recorded 20 m away, especially the
wave
frequencies am anplitudes.
Results of the experiments have shownthat.as
the
charge length increases,. with
a oc:nstant
charge
diameter, the daninant energy transmited
by
the
stress waves shifts to lower frequencies.
'!he ciani
nant frequency decreased fran 2 kHz (for the 0.09 iii
loog charge) to 0.2 kHz (for the 8 m long
charge) ,
but the rorrespooding peak anplitudes
did
not
increase in prqlOrtion to the length, or the weight,
of the explosive. '!his last aspect is prOOably the
result of the fixed distance for monitoring
detala
tion effects (20 meters) which involves ooly
near=-field oc:nsequencesof the blasts. '!he fact is that
detalation pressure is independent of charge weight,
Mtile the total released energy
increases
with
charge weights, showing its effects
at
larger
distances. '!he autlx>rs E!!!PJasizethe aspect that to
reduce vibrations, in order to protect nearby struc
tures, dElcreasing charge weights to be detooated bY
delay results favorable to their stability,
mostly
as a oc:nsequenceof frequency increases (away fran
structures ressooant frequencies).
lbWeVer, it is not clear if this
behavior
is
maintained for other distance ranges, namely in the
far-field effects of blast vibrations.
Another
interesting cx:nclusion of this research regards the
identification
of the P-wavepeaks
as
the
most
:1rlp:lrtant ones in the near-field range of blasting.

K = Ei (Ri Sc + 0i)
WhereE is a correctioo ratio dependent
en
the
exploSi~S, Ri a constant of each rock type, Sc . the
a:rrpressive strength of rock am Oi
a
correccion
factor according the kind of blasting pattern.
In
the field \oJOrk
K varied bebieen 48 am 138, \oIDile n
(the exponent of distance) ranged fran 1. 5 to 1. 7.
All the errpirical expressions resulted fran exper.!.
mental field data, but no rorrelation
ooefficients
anong the variables are reported. Regarding damage
criteria,
the autb:>rs have used the ooe prcposed by
the GentIanStandard DIN4150, on the basis of peak
particle velocities of vibratioo. '!hey cx:nc1ude the
paper by suggesting the executien of IlOre blasts
in
order to better understam the essential
mechanisms
of blasting vibrations under the prescribed
c::and.!.
tions.
2.3

Dynamicbehavior of tunnel lining due. to

adj~
cent blasting, by Y. Goto, A. Kikuchi, T.Nishi~
ka (Japan)

'!he excavatioo am lining of two parallel

tunnels
owenecI in weathered diorite rock in Ogitsu (Japan)
was perfonnec1with oc:ntinuous llOlitoring of blasting
effects, using acceleraneters am strain meters.
Reciprcx:a1effects were measured in each tunnel, as
the other one was excavated, including the perf0nllCl:!!
ce of oc:ncrete linings that were used
to
S1JR'Ort
rock walls.
Due to the short distance bebieen
both
tunnels
(fran 1.5 to 1.8 meters) special attentioo was given
to the maxinumannmt of explosive charge that could
be detonated safely, withalt damageto
the nearby
cavities am their linings.
As for damagecriteria,
the authors utilized three
methods:
a) oorrelation

of particle

velocities

am

to

8.0 m.

Q1 the other hand, IlOdelling techniques haveproved

adequate to s:iJmJlateexplosive behavior, am further
efforts will be done by the authors in order
to
apply dynamic finite element codes to determine the
differences between near am far-field effects,
and
also to establish the ITUtualinfluences
of
site
parameters am blast patterns on
the
transmitted
vibrations to the surrooOO.ingmedia.

2.5

dist<l!!

Essais de ~e
et measures des arriere-bris dans les roches granitiques
du
Massif
central Francciis, by H. Heraud am A.Rebeyrotte
(France)

oes r

b) defonnations of tunnel linings 1

c) strains induced by the stress waves en the rock
walls ana additiooally caused by the
tunnel
linning defonnations.

'!he paper concerns a oarpariscn

of two explosive
substances as pre splitting agents in a granitic rock
excavation. '!he explosive:;o.were: a detonating
cord
(with a density of 70 g/m) am a water gel
sausage
type (cisalite),
with different physical and dlemical
properties. In terms of detonation velocities,
Mtile
the first one is 6,500 mis, the
seoood is
just
2,850 m/s, according to Table I of the article.
As for loading densities, the detonating rord
is
70 g/m and cisalite is a1IlOst six times greater (400
g/m). Another :1rlp:lrtant aspect of presplitting,
the
deooupling ratio, is also very different: 0.12
for
the detonanting oord and 0.24 for cisalite,
as taken
fran the data given in Table 1. COrresponding hole
spacings were 0.8 m am 1 m, respectively.
'!he excavated rock is described as a
fractured
granite am the presplitting
tests
were carXluctec1
in oc:njunction with a 3 m x 3 m grid of 0.115
m
diameter blast holes, with powder factors
of
the
order of 400 to 410 g/m'.
To measure the quality am efficiency
of prespli
tting in the field experiments, the authors
usee
three methods:

Several graPts of the distributions

of
strain,
defonnation am particle velocity are presented
in
the paper, ,for different explosive weights detalated
in the adjacent tunnel, and a oarpariscn
between
measured am calculated strains is aooooplished.
tmfortunately, no oorrelations
between
tb:>se
variables am damagesare prcp:>sed, neither a measure
of explosive charge weight limits is suggested. rack
of quantificatien
is noted as far as relating dist<l!!
ces charges and vibration variables (either defonna
ti~,
strain or particle velocity). HoweVer, F ig 4"
i
of the paper gives a general idea of that rorre 1at on,
\oIDichcould be inproved by means of nultiple variable
regression analysis. In addition to this,
it h ~
suggested that the authors proceed their researc
order to develop reliable criteria for the stability
of nearby excavations.
2.4

An investigation of the influence

of
charge
length upa1 blast vibrations, by J.R. Grant, A.
T. Spathis am D.P. Blair (Austrcilia)

a) definition of quality .indices1

b) pseudo-frequency measurements1
c) microseismic m:nitoring.

'!he paper describes a series of full scale experiments

for the measurementof stress waves in a oen~ 1
hole, surrcmlded by 8 boreholes located at
equid~~
tant points of a 20 m radius cirCle.
In each
of
these points, explosive charges of a PEm-'lNl' ~

Twoquality iIxlices were proposed to

llOl1i
tor tlle
state of the rock face after the blasts: P (profile
:i.niex) as the ratio of the real to the ideal profile
lengths, and C (resulting borehole :i.niex) as
the
ratio of visible borehole lengths to the theoretical

1433

averaged three times greater, Wicating a superposf,

tia1 of effects fran adjacent charges;
b)
with
respect to frequencies, it was dem:nstrated that in
general as the distances increased there
was
a
reduetioo of daninant frequency,
although
that
tendency changed fran site to site;
c)
Rayleigh
waves, generated aloog the grourrl-air
interface ,
have been identified as the main ocrrpcnent of the
low frequency wave-trains, and a canbinatioo
of
Rayleigh and Love waves coold be traced in several
cases; d) results of interpretatioo of single charge
blasts W1cated a propagatioo pattern of systematic
attenuation with distance, acoarpanied by a
pro
gress1ve reductien of daninant frequencies,
and
increase in pulse duratioo. In oc:nclusion,
it
is
suggested that a canbinatien of geological,
lcx:al
and technical factors have contriblted
to
the
peculiar situation uOOeranalysis,
and
although
additiOOal studies are proposed by
the autb:>r,
particularly on the subsurface oooditions
of
the
region, it is suggested to the mine operator that
addit1Cl1al care nust by taken in order to avoid the
supel:POSitiooeffects of nultiple
hole
delayed
blasts, that cootribute to the generatioo
of
low
frequency vibration waves.

Imgth of all boreholes . '!be value of 100% is

the
qltirrun for both iJXiices, altlnIgh their
variatia1
is ~.ite:
bad presplitting
leads
higher than 1
values of P, and smaller than 1 values of c.
<:arparisal of these iJXiices for the two fore-men
tialed explosives indicate advantages for cisalite.With respect to black-eak detenninations, techni
ques of neasuring pseWofrequences and microseismic
events were utilized. '!be first ale is based on the
correlation .between pseJdofrequence and duratia1 of
wave propagation to the fracturing status
of
the
rock mass, so that it is as SOOl'i (or harder)
when
the frequency is higher and sign duratia1 smaller. By
plotting these two variables in a (x, y) gratil, dis
tinct point clusters identify the results
of
eacl1
blast and a oarparisal of explosives is aooall'lished
as Fig. 4 in the paper shows.
Microseismic ncnitoring, .by means of contin\DUSly
logqing the remaining rock, in tenns of ,propagation
velocities,
also prClll'1des
a method to neasure
the
darnagecaused in the remaining rock. 'Ihis
logqing
was perfonned in water filled boreholes,
.drilled
pexpendicularly to the presplitted rock face,
and
also concluded for the superiority of cisalite.
'lhe enly \D'1favorableoarparisoo for cisalite
was
the CIteregarding costs, revealing a slight iJx:rease
with respect to the 70 g/m detalating cord. fiJWever,
other types of charging densities (for exanple
40
g/m) were not subjecte:i to the same type of analysis.
JesuIts of this research, which is obviously valid
for the specific oooditials of the rock uOOerstudy,
as well as the particular explosives and blasting
patterns, madepossible a valid oarparison of .their
actions.
'lhe interesting cari:>1nation
of
presplitting
perfonnance indices and gEqi'lys1cal techniques
for
the assessrrent of black-eak, are the
ootstaIxling
aspects of the paper.

to

2.6

an

PAPERS CN FRACMNrATICN

3.1

IDW frequency vibrations fran

surface
mine
blasting over abandaled underground mines,
by
D.E. Siskind (U.S.A.)

'lhe author describes field work and

interpretatioo
of results of blasting vibratials in a
particular
si tuatien where low frequency waves were a
reasoo
for ooncem.
'lhe detalatials
occurre3 in a surface coal
mine,
and their effects were measured
at
residential
structures located at distances rangihg fran 300 to
3 JOOO meters. In between the mine and
the
houses
there is an abandaled underground coal mine,
Iiotlose
cavities are partially filled with
~
gravely
drift. Because of the al:normd low frequency
waves
reaching the houses and causing damage, an investiga
tion was c:alducted by the U.S. Bureau of Mines
the participation of the following factors: a) blast
characteristics,
namely charge weight per delay and
nultideck fired, as well as two single-charge detala
tials that were specially c:alducted in order
identify the generatien and wave propagation mechan
isms; b) the influence of ground type (a thick zone
of glacial till) en the . reduced
frequency
of
transmitted stress waves; c) subsurface c:x:JlIlOSitien
and extent of abandcnedmine workings over
the
propagatien distances; dJ residential
structures
perfonnance and strength \.In3erdynamic stresses.
As IlUChas 235 detalatials
were subjected to the
analyses, by means of recording their
anplitudes,
frequencies and durations, at 7 houses in a
period
of 11 mooths. In additiCll,
seven
blasts
were
specially prepared for research purposes, and their
effects rronitored 17 m aMa':f, so that wave character
istics
were measured before roodif1cations
1np:>seO
by the mediumoocurred,
Main oc:nclusions of the study, as pointed out
by
the author, are: a) single charge blasts
showed
nonrel vibratioo anplitudes (with respect toprev1ous
~
experience) Iiotlilevibratials fran
productien
blasts, for identical charge weights
per
delay,

on
to

1434

Meas\lI'eIle1tof rock fragmentatien

by digital
FhOtoanalysis, by N.H. Maerz, J.A. Franklin, L.
RothE!nburg(Canada) and D.L. coursen (U.S.A.)

'lhe measurementof fragmentatioo,

or
the
size
distribltioo
of blocks resulting fran rock blasting,
is one of the most challen;ri.r:1
questials in contellpQ
ray DynamicRockM:Jchanics.As the authors
p:>inted
out, there have been attstpts to solve this pr<i:>lems
by several techniques, ranging fran sieving in small
scale blasts, or by means of eJ!Pirical fontU1.ae and
cx:rrp1tersinulations, as well as ~atilic
methods.
~
to difficulties
of these techniques, particularly
the last ones, the authors propose a new ccncept of
interPreting photos taken 00 the fragments
fran
blasting. '!be technique is based 00 digital ~
lysis and aims to neasure sizes
of all. fragments
(both overlapping and IlCIl-<lverlappingones)
in
order to build their size distributioo.
In steads of
taking PJotos vertically over the nuc:kpile (with a
camera located within a ballooo)
or through
a
vertical cut of the pile, they selecte:i to ~atil
the fragments inside haulage trucks,
at
regular
intervals. 'lhe blocks were loaded
along
the
centerline of the nuc:kpile and were taken perperdi
cularly to the bench subjected to detalation.
'lhe subsequent phases of this process are:
a) manual ~t
of oversize blocks,
not
included in the photos;
b) digitalization,
through maJUlalmethods, of the
~,
in order to transfer a two d:imensia\al
image to the cx:rrp1ter1tleIlDry;
c) calculation of block areas fran the
digitized
image, and evaluatioo of their "sizes",
by
means of the equivalent circles
(equal area
ones). oren.-classhistograms are then constructed
for the presentatioo of frequency distributials,
in tenns of nUll'berof blocks a each size,
or
equivalent diameter, per square meter
of the
nu::k pile surface;
.
d) detenninatien of real block size distribltioo,
using two alten'lative methods: analytical
and
eJ!Pirical. '!he first
CIte uses
\D'1fo~
functions to coovert fran two to three
d:1nen
sions, Iiotlich
are based 00 geanetric
prd:labilI
ties and stereology theory, Iiotlile the seoc:ni
methodutilized small scale particles
of
the
same rock that were sieved, as well as
p,oto
graphed and digitized, so that a
correlatioii'
between 2-D and 3-D diameters was established.
Jelevant problems such as fragment overlawing,

missing of fines, and anisot.rq:>ic stacking were

delt with, in order to develop a
reallistic
awroach to the size distribltiem
of fragments
fran blasting.
Although the authors state that this is an
early
state of research on the topic, further developtents
em the Iilotoanalysis technique (namely,
autana?-c
digitalizatiem
and better unfolding expressio~ W1.11
make ita IOOrereliable methcdology.
3.2

An investigatioo into the effect

geanetry on rock fragmentatiem, by
and V.R. sastry (Il'Xiia)

of
blast
D.P. Singh

&nall scale blasts in sandstone blocks ....erecorrlucted

by the authars to detennine the nutual ccnsequerx::es
for fragmentation that can be at~iblted
to three
geanetric d1mensicmsof rock blasting: Wtden; spacing
and bench height. Twotypes
of experiments
....ere
perfoIllled: single hole and IILl1tiplehole siIILl1tanec:us
detonatioos, the first being dedicated to stuiy the
relative effects of Wrden and height,
while the
secon::lones were aimed to analyse the nutual interf~
rences of the three d.inensioos. A total of 83 tests
were executed and in each
one
the
resulting
fragments wer~ sieved in order to deteJ:mine
their
size grading, using detonating fuse as explosive in
all cases.
In teDllS of fragmentation, the autix:>rshavedefined
several interesting ooncepts:
a) optiJm.mlfragmentation Wrden, as the
distarK=e
that maximizes mass surface area
after
the
single hole blasts, and far each bench height.
Mass surface area is defined as the prcduct of
total fragment mass and the retl surface
area
created. by the blast;
b) subcpt.1JmJm
and over opt.1JmJm
burdens resulted in
carparison With opt1mJmvalues, when nultiple
hole siIILl1taneousdetonations were aocatplished,
and after Varying spacing to burden ratios;
c) opt.1JmJm
breakage burden,
as
the distance
correspcnllng to max:1mJm
mass of
fragments,
far single hole tests;
d) In addition, measurements of sc:me ccmninution
parameters were also
established:
average
fragment size, new surface area,
mass.su:tace
area, fine and coarse fragmentation
indices,
and apertures or sizes at 25%, 50% and
75%
throu:j:lwuch the total mass
of
debris had
passed.
Upc:nopt:ainin} size distribltion
histograms
all tests, the authors discussed the results
presented several interesting
oonclusions,
follows:

after
and

a) while opt.1JmJm
fragmentation Wrden was
51%of bench height, opt.1JmJm
breakage
~
aboot 72%;

aroun:i
burden

as

b) increases of bench height produced

increases
of opt.1JmJm
fragmentation Wrdens, all
other
factors constant;
c) increases of burden caused
fragmentation, particularly

a reduction
on
far short benches;

d) ratios of 3 to 4 in spacing to
optiJllumfragmentation results;

burden

gave

e) when these ratios were greater than 5,

each
hole behaved independently, with the formation
of hl.1llpsbetween them;
f) far subopt:iJ'mJm
as well as far optimJrn bw:dens
11'
the mass surface area increased proper ona yti
to bench height dimensions.
These general
conclusions are of qualitative nature, and c::an
be utilized to 1nprove ar control fragmentation
em real scale rock blasting.

1435

3.3

A stuiy far optimization of fragmentation

by
blasting in a highly cleated thick coal seam,
by R.D. Singh, V. Singh and B.P. Khare (In:1ia)

An In:1ian coal surface mine

uses bench blasting
techniques to excavate and prcxiuce the are out
of
a 18 III thick seam, that is mined in t'oIObenches. The
coal is banded, With layers ranging fran 0.30 m to
2.74 ill thick, and a system of discootinuities
is
present, With t'oIOprincipal sets of cleats
interse.
ting each other at an angle of 1120.
Calpressive
strength of the coal was fouzx1to be (in laboratory
tests) of the order of 10-15 MPa aloog
bedding
planes and 8-20 MPain the direction
perperxiicular
to these planes.
In tenns of breakage the authors irdicated
that
the coal used to fragment aloog cleats
and becHing
planes, and this fact led to the farmation of many
ll.llp that did rot pass
a 4.5 III x 3 III grizzly.
Furthenoclre; the pre-existing blast patterns
were
rot adequate in tenns of fragmentation, because they
prcxiuced manyJ::oulders.
The patterns were formed by 6" diameter
holes,
With 4 III burden and 2.5 III spacing, and 4 III distarx:e
between successive rows. ANFO
explosive was
used
With Dynak, and deck loading was practiced,
With
siIILlltaneous firing of each reM ()f holes,
and
a
delay between 15 and' 51 ms between rows.
Explosive
conSUlTptionranged for 0.4 kg/ton in the top bench
and 0.2 kg/ton in the bottan one.
In order to irnprCllTe
breakage the authors
chan;ed
the drilling mesh to 3 III Wrden and 4 III spacing and
takin;r into account the preserx:e
of
cleats
and
bedding planes, thayobtained an oversize
fraction
(greater than 0.3 m blocks) of 10 to 15%of
total
blasted coal, in stead of 25-30%with the
previous
patterns.
The changes were based in stOOies ceniucted With
cratering tests, trial blasts and catparisoo
of
results. The cratering experiments gave irdications
aboot theoretical burdens
and
spacings,
rot
IOOdifications of these were ajusted to account
far
the existeoce of ..discontinuities in the coal.
They <bserved also that ooarser fragmentation and
overl:lreak oocurred when the direction
of
blasting
was out of the lesser angle of cleat
intersection.
Consequently, better breakage was achieved
when
blasting directioo was out of the open angle between
cleat sets.
In teDllS of microdelays between adjacent
holes,
the authors reccmnenda delay inteNal of 9 IllS
per
meter of Wrden, while between rows they
suggest
a delay of 2 to 3 times greater than the one existing
between holes of the 'samexo;
A OCIltJOSi
tion of the above mentioned
solutions,
provided the way to <btain
better
fragmentation
degrees in the coal mine under analysis.
3.4

Fragmentatioo stu:iies in instrunented

coocrete
IOOdels,by W.H. Wilson and D.C. Holloway (USA)

The paper describes a series of ten laboratory tests

on cancrete blocks subjected to explosive
action,
in order to stuiy the mechanismsof
fracture
and
fragmentation. These ~
were IOOnitared
by
means of various instrurrents, such as strain
gages,
cracl< detection gages, acceleraneters,
fiber
optic
probes and high speed filming.
In the tests, eight were ceniucted in 0.5 I113
oon
crete blocks for sin;rle-hole blasts, and two others
hlKitwo ~elayed
holes, all of them loaded With
PEINcharges fran 1. 5 to 5 grams in 6.4 IlDI diameter
boreholes.
Charge bw:dens varied between 7.6 nm to 130 1lDI,
and the ratio burden/dialleter ranged fran 11.5
to
20.5, so that several situations were covered, although
corresp::lniing f~tation
results are rot reported.
The experimental setting inclOOed additional cement
blocks to constraint the block W1derstu:1y, working
as m::mentum
traps.

am

Fracture developnent in both the tq>

bench
faces of the CEmentIlDdel are the main aspects
of
this research. Collected data showedthat the
rrost
iIrp:>rtant cracks in tel:IllSof breakage are the ones
foJ:Inedearly at the bench face in front
of
the
charged borehole, as a result of the tensile stresses
iIduced by the reflexion of ?-waves at that surface.
Also the top surface above the drillhole is subjected
initially
to the samemechanism,
later
00
the
gas expansion action develop these fractures, rather
than creating newones.
'!his inportant conclusion" which was confirmed by
measurementsof dynamic strains at the berx:h face,
indicated the iIrp:>rtant role of the stress waves in
the fragmentatioo process. Another eviden::e
was
that lateral areas of the berx:h face, away fran the
borehole position, sustain high CO'I'pressions, star
ting iItmediately after the passage of stress
waves"";
and caused by the gas pressure which ben:1s
plate-like fragments created by the radial
fracturing
process.
Duratioo times of gas expansion were rreasured at 1
to 1.5 IllSafter detooation, and extrapolating
this
infonnation to field blastin; the authors concluded
that the wtx>le~
is finished after 13 to 16
IllSper meter of l:uJ:den. In tel:IllSof
fragmentation,
high speed films showedthat while
the
largest
fragments were foJ:Inedearly in the process as radial
fractures, IOOStof the smallest ones cane fran
the
area around the boreholes. However,the final crater
of the berx:h face is explained by the radial
frac~
res.
The general conclusion of this research indicates
that laboratory blasting tests,
using
rronitoring
techniques can be of great help to m:x3elreal
scale
blasting with the aim of cptirnizing this
inpJrtant
cperatioo.

!UBI!Il!l.
T","

... ""01.11

am

Althc:o:Jha very small portion of the total explosive

energy released in oonnal rock blasting
cperations
is cawerted into seismic m:>tion (typically
aboot
1%), their damagepotential to nearby structures and
a cause for eatplaints fran neighbours, makes this
effect of blasting a very inpJrtant ale.
The theoretical aspects of the problem are very
c:lCq)lexif the real 3- dimensional situation
is
analysed, because of the interferen=e
of
many
variables (explosive properties,
charge
loading
parameters and mechanical characteristics
of rock).
The experimental awroach to the problem is based
upon instrumentation to record vibrations CMay fran
the shot point ani to correlate their
magnitudes
with obsezved effects 00 points of interest, such as
hcuses, structural famdations, slcpes, tunnels,ete.
In general, this awroach is coosidered sufficiently
accurate to develcp relationships
between
the
arocunts of detalated explosives,
the distawes
of
propagation ani the vibration intensities.
Dtpirical
fonnulae are thus utilized to predict
effects
of
future detalations in the same rock mass,
for
a
certain ran;e of distarx:es ani taking into considera
tioo the directions of wave propagation,
inclu:linij
their orientations, because of geanetric variations
al.oD:Jthe rock mass.
As Fig. 1 depicts, for two gee:px>nes located
at
the samedistawe d fran the blast, one can
expect
higher aI\illitu:ies of vibration in the record
taken
i:lehindthe ben::h than in front of it.
To characterize vibration levels as measured by
eD;J1neeringseisrOOgraphsit is ccmnc:n to
record
particle velocities,
though accelerations,
displace
nents and frequencies are also of interest.
The reason to preferably utilize vibration velocity
is mainly due to the fact that it correlates
best
with observe:l damage, as a result
of
extensive
research ccnlucted at the U.s. Bureau of Mines since
1942, as well as Nitro Nobel of Sweden, and other
insti tutions.

STRUCTURI

---l

d.-----~
Fig. 1 - Geanetric constraints
ing in bench blasting.

of vibration

nonttor

It llUSt be pointed oot that the

proposition
of
safety levels of vibration (for exanple, 5 an/s
as
safe peak particle velocity) is independent of the
seismic waves frequency, at least in the interval of
2 to 500 hertz. Manyquestions arise
fran
these
erpirical rules:
a) are the safe particle
types of structures?

velocities

b) at wuch point of the structure

values be considered?

valid for
shoold

all
these

c) 00w do the propagation laws chan;e

with
the
direction (anisotrcpy), rock type (heterogenei
ties), stress-strain
behavior
(anelasticityT
and the presence
of faults or
fractures
(discontinuities)?
d) what is the effect of explosive charge geanetry
(SFherical or cylindrical)., decoupJ.inJ ratio,
and depth of burden,
on ,those expressions
relating charge weight, C1istaxx:l;!
and vibration
velocity?
Manyother doubts exist in the en]ineer's
mind,
especially whenhe is asked
to
establish
the
max1nurn
charge per delay, in order to
protect
a
certain structure located near the blast. A careful
stu:1y has to include the dynanic stren;th prcperties
of that structure, which is a canplex
interaction
of its geanetry, mechanical characteristics
of their
constituent materials and distaxx:e to the shot point.
Therefore, the cx:moon
met:b::xis
of safety
prediction
are not; adequate for application,
mainly
because
structural reliability
is not expressed in tel:IllS of
particle velocity of vibration. An alternative method
described by the author elsewhere (Gama,1978) pro
poses the dynamic tensile stren;Jth of the
weakest
IlSIlbers of the structure to be protected,
as
the
safety criterion under blastin; vibrations.
In order to develcp this approach it is necessary
to obtain the charge - weight relationships
in
tenns of dynamic stresses, ani this dale by the use
of the well koownequation:
a=pcv

where a is the stress associated

with the
wave
travellirq in a mediumof mass-density p,
with
a
propagation velocity. and for a vibration
velocity
y. Alt:hoJgh this expression is originally valid for
uni-directional propagation, it can be awroximately
extended for field cperations if velocity v Ls given
in tenns of experinental data, namely:
-

if

O-C

where Q is the detonated charge weight, a

distarx:e
n fran the point of interest, and a, b and c
are
numerical constants, depending on field
conditions
ani blasting pattern.
Under this approach, the stress wave ~se
in
given by:
a = p c a

if

O-C

thus making possible to establish

1436

safety criteria

in

furx::tion of stresses.
advantages:

This concept allows additional

a) it accounts for the type of foun:1ation rock an:l

material properties fonning the structure to be
protecteci;
b) it agrees with experimental evidence that
a
magnification .factor of the order of 2 to 5
(and sanetillles nore) for stresses
transmited
in rock in opposition to soil.
The ratio
of
characteristic
~es
(the
prcducts
pc)
beb\leen these two media explains properly that
difference of effects;
c) it confiJ:ms failure limits
of rocks.
wx1er
blastin; actions, as described
by
several
authors. For exanple, Brawner (1974) iIrlicates,
that rock blastin; design ran;e
is
to
be
aoccnplished for particle velocities of 50 to
100 in/s, and usin;
typical
.characteristic
~es
of rocks we obtain dynamic tensile
strengths of the order 5 to 50 MPa, Wti.ch is
confirmed by the values given by AttelNell and
Fanrer (1976). Aoother evidence
canes
fran
Langefors (1963) who specified particle velocity
limits for granitic rocks,
shaoTing
the
initiation
of cracks for peak velocities
of
60 em/a, AssuniD:Ja density of 2.65 an:l P-wave
velocity of 6,000 mls for those rocks, we cb~
a dynamic tensile stren;th of 10 MPa,Wti.ch as
CCIlpatible with "in situ" deteJ:minations.
Consequently, the criterion of structural
can be implementedthrough the evaluation of
factors, defined as the ratio:
crt
Fs = a

safety
safety

The peak stress equation

was ci>tained,
after
substi tuting for a cofferdam core material
density
of 2.3 and a seismic velocity of 600 m/s,
thus
resUlting:

o = 0.08 01.98 0-097 (MPa)

and estimating a dynamic tensile strength
of
the
clay core as 62 KPa, the various safety factors were
represented as Fig. 2 shows.
Aoother case study regards the
evaluation
of
minimlmwidth for a basaltic wall constituting
the
interrrediate roa&.ay linking the two margins to the
deviation canal in the Parana river,
durin;
the
construction of Itaip.1 Dam.That basaltic wall
was
62 m high, with vertical sides, and it was designed
to resist ground vibrations fran adjacent explosions
performed to excavate the canal.
"In situ" measurementsrevealed a propagation law
of the type:

0.53 0146 0-219

(m/s)

ard using the properties

of basalt
c = 4,800 m/s) the expression
for
stresses is:
= 7.26 01.46 0-219

(p = 2.8 an:l
dynamic peak

(MPa)

Using a dynamic tensile strength for basalt of the

order of 35 MPa, it was possible to establish
a
safety criterion for that structure,
which
is
represented in the gri3{tlof Fig. 3.

40

where crt is the dynamic tensile strenqth

of
the
Structure to be protected (or its weakestmernber)and
cr is the stress associated with the seismic
wave
originated by the blast.
'l\> additional
aspects
must be kept in miIrl: if particle
velocity
.is
considered as peak values, so o is
the
resulting
peak stress, and the chan;e fran eatp:'essive
to
tensile stress is due to wave reflection
mechanisms
because of the well-krown smaller resistance
of
solids in tension than in CCIIpression.
J:plications of dynamic stress criterion have been
described by Gama,1978, and later on utilized
in
the prediction of vibration effects on
residen~al
buildings of different types, located in the vic~nity
of quarry blasts.
In the case of an earth dam (see Fig. 2) that had
to be protected fran the dem:>lition by explosives of
a large ccncrete wall', the propagation law in terms
of peak particle velocity was detennined as:
v = 0.058 01.98 0-097

':

u
Z

...
Q

,
10

CHAIlG[ WII"'"

'00

'I~

DILAY (.,

Fig. 3 - Safety criterion for Itaip.1 dam intermediate

roadwayof deviation canal, sul::mitedto
vibrations
fran nearby excavation.

These exanples demonstrated the possibility

of
taking appropriate decisions, with respect
to the
cliJlensicns of a struture to be designed, or to the
amoontof explosive to be detonated per delay, within
a new frameworkof analysis for ground
vibrations
resultin; fran rock blastin;. Those decisions
can
take into account the choice of an adequate
safety
factor and its cansequerces in tenns
of
costs,
allowing the possibility of protecting in different
ways, the various structures subjected
to
the

...

..

10

(m/s)

100

10

~.

'0

!!!
Q

I
Q

CH."OE

WEIGHT PEI't

DELA.

Fig. 2 - Safety criterion

for
COfferdam,suI::mi
ted to vibrations
dE!lOlition.

10

(k,)

the
SCbradinb::>
fran
nearby

1437

Furthermore, the use of stresses

in stead
of
particle velocities seems to be a
!lOre familiar
concept to en;ineers, even tOOughthis a simplified
approach, which order of magnitude is
not greater
than the experimental errors camonly recorded
in
vibration !lOnitoring.
.Additional effort.; have to be developed in order
to better quantify reliable criteria
for
damage/
failure/stability
of rock structures subnited
to
grouIrl vibrations, in order to protect thEITI ad~
tely.

attitudes a,rXispacings. MicrClCalplterprograms are

available to establish the size distributions
of
blocks in a rock bench that is subjected to such a
survey.
As far as breakage is conce:rned, ale can develop a
rccdel of the influen:::e of natural fracturing,
as
Fig. 5 represents.
.
For increasing degrees of natural joint density it
can be ooserved the correspon:i1ng size distribution
curves, indicating the
use
of
less
explosive
col18Ul'lPtion
to reach a desired fragmentation. It is
envisaged that for each real blasting situation
it
is possible to develop the two curves depicted
in
Fig. 4, tlU1sleading the design of blastin]
I:'CA.Inds
to hig~ level en;r1neering. The process can be even
IlOre sq:h1sticated if adequate consideration
is
given to the spacial location of bigger blocks within
the bench to be excavated, SO that the placement of
explosive charges is deteJ:rninedon the basis
of
their indiv1dualeffects
in teJ:msof reducing
the
size of those blocks, in order to Wild
a desired
size distr1b.1tion of fragments for the total blast.

An increasing awareness of
the 1rlTportar::e
of
fragmentation in rock blasting can be confiImed by
the organization of international synposia
on the
subject, every foor years (the first was held
in
Lulea, Swedenin 1983 am the second in Keystone, Co
in 1987). Manyresearch grc:upsare IlC70I
involved in
stu:1ies for better urderstaniing the~,
so
that field operations can be adequately designed to
reach a certain fragmentation degree
that
will
m1n1Jn1Ze
total costs of rock excavation
(drilling
am blasting), transportation (lcs1:1.nqam hauling)
am medlan1cal breakage (c:rushi.n;Jam milling)
Early in 1971 the author pmlished a paper proposing
the utilization
of earminution theory to quantify the
size distribution of fragments as a reault
of
a
blast. The governing law, which was shewn to
be
consistent with BoIn' s theory of camdn1tion, indica
ted that the percent ClIllllative undersize weight ot
the fragment fraction size S, is given by:
c
P = a r)J (.2 )
s
B
litlere Wis the total explosive energy espended per
unit weight of blasted rock, B is the Wrden of the
charges am a, b, c am nunerical factors depen:1ent
on explosive type, rock properties
and
blasting
pattern. This empirical relationship was established
on the basis of laboratery scale tests, am it has
been oonfiImed in real acale blasts
(for example,
TogtX)I1,
1916 and Borquez, 1981).
Due to the fact that ordinary blasting
operations
are oc:n:luctedin jointed rock, ajustments of
that
law have been proposed (Gama,1983).to include the
natural state of fracturing in the
rock
IIBSS
subjected to the detalation. The newexpression was:
Ps =

Sc

W- (B)

Fig. 4 - Block size distribution curves prior

after blast, and hatched
area prcpxtional
explosive energy used in fragmentation.

(F50)

~
FSOrepresents the average block size in the
beD::h, tfle other synmls keeping the same
mean1nj
as before, and the coefficient d
is a
positive
l'lUli:ler
.
To ootain this relationship, results
of
seven
quarry blasts in jointed basalt were analysed
in
detail, including a manual survey of block
size
distributions.
Although these cases were characterized
by an average size of blocks before blast (FSO)always
less than the Wrden of the charges (B), indicating
situations ~
the effect of jointing is daninant,
it seems prcbable that the expression
contains the
general treDi of variation for the pherx:Inerm.
Manyauthors stress the ~
of discontimJ!
ties in the consequences of rock blasting,
!VIlely
frae.Jll!lltationand flyrock, leading the research in
this area to ce::n:Jentrateon scph1sticated 1ns~
tation, high speed c1nematography,ete. For exaIlple,
Amerson, Winzer and Ritter (1985) state that
"tile
fracture density, measured on the face prior to the
tests, can be related in a general way
to gross
chan]es in fraqoent size distr1b.1tion". However,the
ned1an1cs of breakage are so oatplex,
with
wave
reflex10ns at external and internal
free
faces,
microdelay influences, time
sequrm::e of
gas
expansion am their II8.ltualinterferences, that IlOre
research in needed to fully explain this PJienanenon.
In a qualitative way, it can be confiImed that it
exists a close relation bebieen the size distr1b.1tions
of blocks before am after detalation,
as Fig. 4
depicts. The hatched area is proportional
to
the
fraction of eJCPlosiveenergy effectively
used
in
fragmentation, or in pratical tenns, to the powder
factor (specific energy consutption), which is the
variable Win the abaI1efOIllll1ae.
The technique to establish the natural block size
grading of a rock bench was desCribed
elsewhere
(Gama,1977) am it is based on voll.llletric distribu
tion of blocks, upcn a survey of
discontinuitY

100%

and
to

----

..~
..~

a:

>

i13

SIZE

Fig. 5 - Typical size distribution curves of

rock
blocks prior to blasting, for I - intact, II - ~
rate jointing am III - highly jointed rock mass.

Pock blasting is still CJr.eof the n-ost challenging

areas of Pock Mechanics, in teJ:ms of
theoretical
research to be developed to
subsidise
reliable
en;r1neering design, leading to sc:;unjfield results.
N1ile ground vibration is a deterious effect
that
re;iuoes naturally as rational design is applied, the
subject of fragmentation is of
great
econ:mic
1np:lrtanCe, thus justifying cont.i.l'unls efforts
to
1JlprOvesuch a design.
Creativity and innOVationhave a great
role
to
participate in this task to makebetter rock blasting
nethods, for the benefit
of civil
am
mining
en;r1neering

1438

am

Anderson, D.A.; WINZER,

S.R.
Ritter, A.P.
1985
- The relationships
bebr.leenrock
structure
delay timing, am fragment size am distrilution
in explosively - Ioeded rock. Fragmentation
by
Blasting (1st Edition). SOCiety of
Experimental
Mechanics, p.41-62.

am

Ash, R.L. 1985 - Flexural rupture as a rock breakage

mechanism in blasting. Fragmentation by Blasting
(1st Edition). Society of Experimental Mechanics,
p.24-29.
Attewell, P.B. and Fanner, I.W. 1976 - Principles
el'XJineering geology. Chapnan and Hall.

Tan Tjong Kie (China)

My contribution/qu tion i. directed to:
Y. Goto, A. Kikuchi and T. Ni.hioka (Japan).
Dynamic behaviour of tunnel lining due to adjacent
blasting

of

j30rquez, G.V. 1981 - Estimating drilling ani blasting

costs - An analysis ani prediction m:rlel. ~
ring and Mining Journal, Jan. p.83,..89.
Brawner, C.O. 1974 - Rock mechanics
in open pit
miniI]. Proc. 3rd. Con;jress of Int.
Soc.
Rock
Mechanics. Vol. I-A. p.755-815, Denver.
Coates, D.F. 1981 - Rock mechanics principles
(3rd.
ed.). Dep. Energy, Mines ani Resoorces, ottawa.
Dick, R.A.; Fletcher, L.R. ani D'Andrea, D.V.
1983
- Explosives ani blasting procedures maIUJal.U.S.
B.1reauof Mines, Infcirmation Circular
8925.
WashiI]ton
Gama, C.D. 1971 - Size distriOOtion general law of
fragments resulting fran rock blasting.
Transa
tions SOCiety of MiniI] EI]ineers of AlME. Vol.
250. p.314-316.
Gama, C.D. 1977 - eatputer ncdel far
block
size
analysis of jointed rock masses. Proc. 15th Al'CCM
~um.
JID.t.1.Brisbane, p.30S-315.
Gama, C.D. 1978 - Dynamicstresses
fran
blasting
vibrations. A better damage criterion.
zroc. Int.
Synp. Rock Mechanics Related to Dam Foordations.
Rio de Janeiro, CIfoIR/AIMS,
p.III.1-III.14.

The authors pr.s.nt sn int.re.ting pap.r on the

.xcavation by bla.ting of twin parall.l tunn.ls
etr.ngth.ned by the rock-bolt-ehotcr.t. m.thod.
What mak the probl.m attractiv. i. that it m.
daring to kp the mutual di.tance betw n the
linings only as small 1.5 to 1.8 m, wh.r.as a
Young's modulus of the rock ma of 1-3 GPa and it.
low propagation velocity of 0.9-1.2 km/.
and
furth.r Vp - 1.5-3.0 km/s
and V. -0.8-1.5 km/.
for rock .ampl.s are clear indications that the rock
ma mu.t be fiur.d and not of good quality. In
order to .tudy the po ibl. risk of damag. to the
s.parating rock wall the authors have p.rform.d many
dynamic m.a.urem.nt. with the h.lp of pi.zoel.ctric
acc.l.romet.rs and strain gauge.; u.ing the
rult. the di.placement. and particl. v.locitie.
have be.n comput.d.
On the basis of table 1 and figur.s 5 to 12, I will
now pr.sent the following analysi.: Th. dynamic
str.ss ad due to a propagation str.s. wave in a
solid can b. giv.n by the following formula:
ad - p.Cp.Vp
wh.r.by the d.n.ity p - y/g - volume w.ight/98l em/
81;

Gama, C.D. 1983 - Use of caml1.nution theory to predict

fragmentation of jointed rock masses subjected to
blasting. Proc. 1st Int. Synp. Rock Fragmentation
by Blasting. Vol. 2, IAllea, p.S6S-S80.
Gama, C.D. 1984 - Microoatp.1ter simulation of rock
blasting to predict fragmentation. Proc. 2Sth U.S.
Synp. on Rock Mechanics,ai. AlME,Urbana.
Hagan, T.N. 1983 - The influen::e
of
controllable
blast parameters on fragmentation
ani
miniI]
costs. Proc. 1st Int. Synp. Rock Fragmentation by
Blasting, Vol. 1, !Allea, p.31-S2.
Kolsky, H. 1963 - Stress waves in solids.
Publications, NewYork.

m.asured level of 900 mm/s

would definitely be far
more dangerous if the distance from blasting to
lining was 50 m instead of 2-5 m. What is the
.xperience of Mr. Goto concerning the frequ.ncy
content of bla.t vibration?

Dover

Lan:.:Jefors,U. am Kihlstran, B. 1963 - The modern

technique of rock blasting. John Wiley &
sons,
NewYork.
Tognon, A.A. 1976 - Analytical study of blasts
in
basalt fran several dam quarries (in portuguese) .
Proc. 1st Brazilian Congress of EI]ineering Geology.
Rio de Janeiro, ARiE, Vol. 2, p.283-296.

Cp - v.locity of wave propagation (1.53.0 km/s)

Vp - particl. v.locity within the rock ma .
Th. author. have comput.d the particl. velociti.s at
the fr urfac., which w. denote by Vp.urf which
are in the ord.r of 10-90 cm/s. Th. int.rnal
particl. v.locity Vp i. known to b.:
Vp - ~ Vp.urf
With the above data I g.t ad - 2 - 35 kg/em";
th str may b. conqu.ntly compr iv. and
t.nsil. due to multiple r.fl.ction.. A. the rock
wall conc.rn.d do not .e.m very .ound, dynamic
t.n.il tr s of this ord.r of m.gnitud. may
g.n.r.t. v.rtical fiur in the .eparating rock
wall and the frh concr.t. lining and furth.r
.palling at the unlin.d cavity .urfac.. Und.r the
influ.nc. of the v.rtical .tr due to the
ov.rburd.n, long t.rm cracking l.ading to po ibl.
in.t.biliti i. not unlik.ly. H.nc., in addition
to the rockbolting-.hotcr.t. t.chniqu., I r.commend
c.m.nt grouting to pr.v.nt further di.int.gration;
furth.rmore rock bolting prior to and not aft.r the
.xcavation will b. pr.f.rabl

Floor Di.cu ion.:

Rock Bla.ting and Excavation - Th.m. III

Tor. L. By (Norw.y)
My contribution/qu tion i. dir.ct.d to:

Mod.rator Summary:

P.A. Lindqvi.t (Sw.d.n)

Y. Goto, Dynamic b.haviour of tunn.l lining,

1 INTRODUCTION
Maximum .cc.ptabl. particl. v.locity on the fre.h
concr.t. lining was di.cu.s.d. Th. critical
importanc. of the bla.t vibration fr.qu.ncy cont.nt
was not treat.d. At the v.ry .hort di.tanc. in this
actual proj.ct, the fr.qu.nci probably w.re v.ry
high, that i. many hundred hertz. At this high
fr.qu.ncy l.vel, damage will .a.ily occur. Th.

Theme III of the Sixth International.Congress of Rock

Mechanics is rock blasting and excavation. In the proceedings, a total of 30 papers on drilling. mechanical rock excavation and blasting were presented. A
summary of technical areas covered in the papers' is
shown in table 1.

1439

Table 1. Technical

Blasting dl!sign
R D Singh. V Singh & Khare

areas of papers in Theme III

Vibration

3
2

Drilling - general
Percussive drilling
Rotary drilling
Oil well drilling

criteria

Fadeev. Glosman. Kartuzov

Huh. Lee & Lim
Goto. Kikuchi & Nishioka
Grant. Spathis & Blair
Siskind

1
_2_
8

Mechanical

No of papers

Drilling

&Safonov

excavation
Presplitting

3
4

TBM

Roadheader
Blind raise boring
Static demolisher

Heraud & Rebeyrotte

1
Measurement

_1_

of fragmentation

Maarz. Franklin.

Blasting

Rothenburg

Rock mass characterization/cost

1
4
3
1
_4_

Rock mass characterization

Blasting - general
Blasting - Open pit
Presplitting
Tunneling - various aspects

Inazaki & Takahashi

PottIer & John
Mikura

& Coursen

model

_3_

13

13

As can be seen there is a good balance between

ling. mechanical excavation and blasting.
It is more interesting perhaps to look at the
lems treated by the different authors. In table
pers on drilling and mechanical excavation have
grouped under four main headings.
Table 2. Selected
excavation

problems

- drilling

In summary the papers of Theme III show that active

and successful research and technical development is
being carried out throughout the world in this area.
Since the last Rock Mechanics Congress a few major developments must be mentioned. Among these are the increased use of medium high pressure water jets to assist roadheader cutting and the wide establishment of
blast vibration criteria based on local assumptions
but often influenced by international experience.
Papers of Theme III also include a few important case
studies that contribute to the long term development
of rock blasting and excavation technoloFY

drilprob2 pabeen

and mechanical
No of papers

Prediction

of penetration

rate/performance/feasability
2 HIGHLIGHTS

Bernaola & Oyanguren

Howarth & Rowlands
Miranda & Mendes
Farmer & Garrity
Bonapace
Mitani. lwai lie Isahai
Design of bits/drill

steel/equipment

Brysch. Nsenga & Shan

Brighenti & Mesini
Nishimatsu. Okubo & Jinno
Xu. Tang & Zou
Borehole

Considering the high quality. it is a difficult task

to select and comment on specific papers. A few are
however worth highlighting.
The paper "Fragmentation studies in instrumented
concrete models" by W H Wilson and D C Holloway uses
a combination of different observation techniques
which makes it possible to draw new and firm conclusions. This is a key paper and gives. in.my opinion.
a major contribution to the understanding of mechanics
leading to fragmentation.
The paper by J R Grant and co-workers entitled "An
investigation of the influence of charge length upon
blasting vibrations" indicates a possibility to detect
detontation cut off. at present a serious problem in
underground mining.
The paper "Measurement of rock fragmentation" by
N H HBerz and co-workers is an instructive example
of how the use of computers opens new ways to solve
difficult engineering problems.
The paper by J Brych and co-workers shows the impressive result of several HSC and PhD theses under
the title "Destructability of rocks with rotation
drilling bits".

stability
1

Kaiser & Maloney

Cutting mechanisms/water

jet

Hood. Geier & Xu

Iihoshi. Nakao. Torii & Ishii
Ip & Fowell
Kutter & Jiitte
Sekula. Krupa. Koci. Krepelka & Olos
Fukuda. Kumasaka. Ohara & Ishijima

_6_
17

3 FUTURE DEVELOPMENTS
In Table 3. finally.
the authors have been
It is notable that no
blasting and empirical
presented.

blasting problems taken up by

classified under six headings.
papers on computer modelling of
modelling of fragmentation were

Table 3. Selected problems

Blasting

- blasting

theory

Singh & Sastry

Wilson & Holloway

When looking into the future there are several aspects

to consider. One of them is the length of time. In the
discussion of future research needs I have taken a
fairly short term perspective. It is arguable whether
scientists should consider the near future. It is my
opinion. however. there are coming or soon will come
changes that will have a major impact on the selection
of research problems.
Befor starting your research you first have to choose the technical area to work in. Then you have to
select the scientific problem and finally you have to
consider the kind of method you are going to use. such

1440

as laboratory testing, field testing and computer modelling. Ideally all should be used on the same problem, but this is certainly not always the case.
In the following presentation a few important technical developments will be identified and the consequences specified. Future research needs will then b~
given.

models and blast design practice with emphasis

tual problems and new explosives. .

Development of fast methods to monitor blast performance with respect to delay time, total charge detonation, vibrations and fragmentation.
Further development of presplitting
ting methods with new explosives.

Future developments

Consequences

New materials

High speed drilling and high

efficiency equipment will
reduce the cost of holes

Medium high pressure

water jets

TBM and Roadheader mine development and mining in hard

rock will be introduced

High capacity
equipment

High availability of equipment will be very important

Increased scale in
underground mining
and excavation

Larger holes (diameter 50100 mm, length 10-50 m) will

be introduced even in small
scale mining

Automation and robotization will increase

Increased accuracy
ling and blasting

Monitoring and hole

logging will increase

in dril-

Dramatic increase in aquisition, analysis storage and

presentation of geodata

Data handling

Precision caps and

pumpable, variable
strength explosives
4 RESEARCH

Dramatic increase in blasting

design possibilites

NEEDS

With identified developments and consequences

the following research needs are proposed.
4.1 Drilling

and mechanical

in mind

excavation

Case studies with emphasis on new experience.

Development

of cutting removal models.

Development
rOck.

of Roadheader

prediction

models in hard

Development of fast methods to measure hole length

and hole deviation, models and techniques to reduce
hole deviation.
Development of new methods by which geodata can be
obtained:
Measurement while drilling (MWD) for hammar drilling including model development.
Sampling of cuttings while drilling and chemical
and physical analysis including model development.
Hammar drilling followed by geophysical hole logging including model development.
Development
models.

of drilling and mechanical

fragmentation

Development of integrated systems for efficient aquisition, analysis, storage and presentation of geodata.
4.2 Blasting
Data collection

and further development

of empirical

1441

on ac-

and smooth blas-

Further development on theoretical (first principal)

blasting models, including influence of rock type,
rock mass and explosives behaviour.