Fragblast 1385-514x/02/0601-104$16.

00
2002, Vol. 6, No. 1, pp. 104±136 # Swets & Zeitlinger

The Use of Lagrange Diagrams in Precise Initiation

Blasting. Part I: Two Interacting Blastholes

H.P. ROSSMANITH

ABSTRACT

Using the concept of Lagrange diagrams this contribution details the calculation of the delay time between
blastholes in a row and rows of blastholes with respect to precise initiation timing within the new advanced
blasting technology which is based on the use of electronic detonators.
After introducing the representations of stress waves and cracks, this contribution focuses on the role of
stress wave interaction in optimal fragmentation in surface blasting and bench blasting. Part I of the paper
considers two interacting blast-holes, Part II will be devoted to three or more out of plane interacting
blastholes, whereas Part III will treat the interaction with a free face such as encountered in bench blasting.
A few simplifying assumptions have been made in this paper with respect to the rock mass as well as the
mechanical treatment. The essential assumptions include that the rock mass is treated as a continuum with
®nite tensile and compressive strength and the effects of structural geology are not taken into account. In
addition, the analysis in Part I is simpli®ed by two `educational' assumption, that all waves are plane (i.e.,
one-dimensional) waves and three-dimensional effects of ®nite size blastholes and charges are not taken
into account.
This contribution will also show that knowledge in wave propagation and fracture mechanics is essential
for the successful application of the new blasting technique in industry. In particular, the delay time, the
wave speeds in the rock mass, the shape of the wave pulse and the acoustic impedance mismatch (not
considered in this paper) have become decisive parameters in advanced blasting.
Utilizing the wave speed and wave shapes of detonations, large scale tests in various countries (Australia,
Chile, etc.) have shown that optimal delay timing requires shorter delay times in conjunction with allowing
for a wider drilling pattern and the use of a grossly reduced amount of explosives, i.e., a lower powder
factor. This seemingly contradictory arrangement is fully justi®ed by using scienti®c principles in blasting,
and converting blasting from an art to a scienti®c discipline.

1. INTRODUCTION

The current state of affairs in the mining industry is characterized by an accelerating

decline of the mining industry in many industrial countries such as England,
Germany, Japan and others. Any new development will be driven by several constraint
factors among which cost reduction and improved safety will be most important [1].
Excessive increase of the cost of labor and increasing technological dif®culties have
Address correspondence to: H.P. Rossmanith, Vienna University of Technology, Institute of Mechanics,
Wiedner Hauptstr. 8-10/325, A-1040 Vienna, Austria. E-mail: hans.peter.rossmanith@tuwien.ac.at
 THE USE OF LAGRANGE DIAGRAMS IN DELAY TIME DETERMINATION 105

forced many countries to shut down most or, in some cases, all of their mines which, at
one time, were valued national treasures, as they constituted the lifelines of their
industries, success and national pride.
Around the world, even and controlled fragmentation ± a requirement for the
optimization of the blasting process ± has been a strong motivation for science and
industry to develop new technologies. Controlled fragmentation involves two basic
requirements: knowledge of the geomechanical parameters of the rock and rock mass,
and availability of a precise initiation timing system.
This worldwide scenario led to the sponsoring, during the 70s and 80s, of extensive
lab model blasting testing, performed at the Photo-Mechanics Laboratory of the
Department of Mechanical Engineering of the University of Maryland (Fourney et al.,
1973±1987). This research project was sponsored by the National Science Foundation
and Martin-Marietta and was aimed at establishing a set of rules for blasting to
achieve optimal fragmentation. Similar tests were also performed later at the Fracture
and Photo-Mechanics Laboratory at the Institute of Mechanics of the Vienna Univer-
sity of Technology [2, 3].
At the time, dynamic fracture mechanics [4] had just been developed to a degree
where it was of use to engineers and the application of stress wave propagation and
the newly established discipline of fracture mechanics naturally led to the question:
using the mechanical principles of fracture and wave propagation, what is the proper
timing for optimal fragmentation?
Ultimately, the research work was successful and failed at the same time. It was
successful, because many of the fundamental mechanisms which operate during
blasting and fragmentation could be identi®ed using high speed cameras in
conjunction with dynamic photo-elasticity. The program failed from an applications
point of view, because the lab tests, based on micro-second delay timing, could not be
translated into the milli-second delay time world, because the electric detonators,
available at that time, showed an inherent scatter too large for any of the precise
timing experiments required. Optimal fragmentation based on the application of
scienti®c disciplines of fracture mechanics and wave propagation was not possible at
that time, and therefore, the subject lay dormant for a while.
However, when reliable electronic detonators became available in the mid-90s
interest again arose in precise timing and achieving improved fragmentation. The
world-wide use of explosives with low detonation velocity, together with detonating
systems of timing capability characterized by large scatters, has yielded consistently
poorer blasting results. Using emulsions with higher velocity of detonation, some of
the pertinent problems could be overcome, but non-uniform fragmentation was still a
common problem and severely in¯uences the downstream costs in quarry and mining
operations.
One fact was obvious ever since the laboratory scaled model tests were performed
in the 70s and 80s: a strong dependence of the delay timing on the geophysical rock
106 H.P. ROSSMANITH

data such as the speed of the propagation of the elastic stress waves, the shape of the
stress waves and the acoustic impedance of the rock mass. This could now be proven
in the ®eld through initial ®eld testing. Classical blasting did not require the ®rst two
parameters and the third one was only occasionally accounted for.
The theory behind the new blasting technology, developed through dynamic
laboratory scale testing at the University of Maryland/USA starting in the early 1970s
[5] and at the Vienna University of Technology/Austria [2] during the last 25 years,
could now be put to work through ®eld testing at various sites in South Africa,
Australia, the USA and most recently in South America. However, blasting engineers
form a tight group of people who often display excessive skepticism and exhibit a did-
it-always-this-way-and-worked habit. Starting in the late 90s initial trial tests and a
few large scale production tests with emulsion mixtures and advanced electronic
initiation systems including electronic detonators in several continents and performed
by several companies have proven highly successful [6, 7]. Not only have these blasts
yielded excellent uniform fragmentation results but they also have revealed a strong
tendency towards substantial reduction of primary cost of drilling and blasting. In
addition, vibration control can be achieved from an improved viewpoint [8].
However, when the electronic detonators became commercially available, their
real bene®t was not immediately recognized by the industry. Mine operators and
contractors were hesitant to follow the advice of the detonator producers to utilize
electronic detonators instead of the conventional electric ones. And with a price ratio
of more than 10:1 in favor of the electric ones, the move from electric to electronic
was not a light-hearted one. In fact, mere replacement of electric by electronic
detonators did only reduce the scatter in the timing, in itself already a bene®t which
led to certain improvements and savings. However, there is more to it: precise
initation timing opens the door for a new way to blast using shorter delay times and
resulting in better fragmentation and leading to additional considerable savings.
The recent development of reliable electronic detonators and initiation systems,
however, does not only constitute one more turning point [9, 10]. It will fundamentally
change the scene of blasting. Only a radical shift from conventional blasting to advan-
ced technology blasting using precise initiation timing will display the real power of
the electronic detonators and the opportunities offered by the advanced blasting tech-
nology. And this shift includes the use of the basic principles of fracture mechanics
and wave propagation. Blasting experts and technicians will be subjected to a change
of paradigm, i.e., they must learn how to cope with scienti®c issues, as the ef®cient
application of the new technology requires an understanding of the physical processes
which are behind blasting for fragmentation. Hence, science has taken a ®rm hold in
blasting, as exploiting the bene®ts of the new technology will only become possible, if
the mechanisms acting during blasting and fragmentation will be well understood.
The new blasting technology does not require the blaster to enroll at a graduate
program at a university. What is required is a basic understanding of the fundamentals
 THE USE OF LAGRANGE DIAGRAMS IN DELAY TIME DETERMINATION 107

of wave propagation and fracture mechanics, in order to appreciate the bene®ts of the
new technology [11].
New parameters have emerged in the blasting arena: speed and shape of stress
waves and the acoustical impedance have suddenly become important technical items.
Classical blasting does neither care about the speed nor the shape of a blast wave,
because none of the standard equations and ®st-formulae in textbooks and handbooks
on blasting refer to stress waves and their interaction with cracks. Laboratory scale
tests have shown that optimal fragmentation can be achieved, if the blast waves and
cracks interact in a particular way.
When a blasthole detonates a blast wave is radiated into the surrounding material.
For a linear elastic material, theory and lab model tests show that this blasting-
induced stress wave consists of a leading compressive pulse and a trailing tensile
pulse. The length of this wave depends on several factors: the type of explosive
(brisant or non-brisant explosive) and the type of rock. The rock mass takes into
account the structural geology which is one of the controlling factors in blasting and
fragmentation. The length of such a wave varies between a few meters and can be as
long as 150 m in granite bedrock. The character and magnitude of the tensile tail
strongly depends on the structure of the rock mass and, in a highly jointed rock mass,
it may completely be absent. The wave speed (not the particle velocity!) varies
between 2000 m/s (soft sandstone) and 6,500 m/s (basalt and granite).
Considering two adjacent blastholes [12], maximum fragmentation is achieved in
those sections between the blastholes, where the two tensile trailing sections of
the blast waves meet. For simultaneously detonating charges this happens at the
midsection of the spacing of these blastholes; for a delayed charge it occurs off-
centerline. Time-wise, this normally occurs within the range of a few milliseconds,
hence, the inter-hole delay time must be chosen appropriately. In order to make full
use of the adjacent rows, a considerably shorter delay time is chosen than in con-
ventional blasting in order to exploit the superposition effect of the stresswaves.
However, care must still be taken, as the rock material in the leading row must move
before the stress waves of the adjacent row arrive in order to ®nd a free face. Again,
the delay times can considerably be shortened as compared to conventional blast-
ing. Optimal fragmentation has been achieved in all applications with electronic
detonators, particularly in Chuquicamata in Chile, in various open-pit mines in the
Hunter Valley in Australia and elsewhere, with an inter-row delay of between 16 and
25 ms.
The presence of one or several joint sets and faults is still a problem and wave
propagation theory has to be merged with structural geology. The blast design must
take into account the presence of the structural geological features and delay times
must be altered according to the quality of the joints and faults and the thickness,
orientation and quality of the rock layers in the rock mass. Unfortunately, in most
cases the characterisation of joints and faults, i.e., the identi®cation of their position
108 H.P. ROSSMANITH

and the properties, remains highly incomplete and this situation renders the appro-
priate execution very dif®cult if not impossible at present.
A third important quantity is the acoustic impedance, which is the product of wave
propagation velocity and density of the rock. Again, previously, this parameter was
not an ultimate necessity for achieving good fragmentation. Theory shows, that not
only the acoustic impedance of the explosive, which is characterized by the product of
velocity of detonation times the density of the explosive, is important, but even more
the ratio between the impedances between explosives and rock and between two types
of rock in strati®ed rock.
After introducing the representations of stress waves and cracks, this contribution
focuses on the role of stress wave interaction in optimal fragmentation in surface
blasting and bench blasting. Part I of the paper considers two interacting blast-holes,
Part II will be devoted to three or more out of plane interacting blastholes, whereas
Part III will treat the interaction with a free face such as encountered in bench
blasting.
All of the following could basically be treated by using computational mechanics
methods, i.e., by using more or less sophisticated computer software packages which
feature dynamic codes based on either ®nite elements or some other numerical
method. This would allow the user to highly extend the method to include non-linear
material behaviour and introduce into the analysis the effects of geological structure
as well as other additional complexities.
The goal of this paper is to `gently' introduce the reader into the method of Lag-
range diagrams and their use in visualizing and understanding the process of inter-
action of cracks and waves which is one of the basic physical events in fragmentation
by blasting of rock. Hence, in order to achieve this purpose, the elements crack, wave,
etc., are introduced and their interactions are shown.
A few simplifying assumptions have been made in this paper with respect to the
rock mass as well as the mechanical treatment. The essential assumptions are:
a) The rock mass is treated as a continuum with ®nite tensile and compressive
strength and
b) the effects of structural geology are not taken into account.
In addition, the analysis in Part I is simpli®ed by two `educational' assumption,
that
a) all waves are plane (i.e., one-dimensional) waves and
b) three-dimensional effects of ®nite size blastholes and charges are not taken into
account.
This contribution will also show that knowledge in wave propagation and fracture
mechanics is essential for the successful application of the new blasting technique in
industry.
 THE USE OF LAGRANGE DIAGRAMS IN DELAY TIME DETERMINATION 109

In particular,
a) the delay time,
b) the acoustic impedance mismatch,
c) the wave speeds in the rock mass, and
d) the shape of the wave pulse
have become decisive parameters in advanced blasting.
Utilizing the wave speed and wave shapes of detonations, optimal delay timing
requires shorter delay times in conjunction with wider drilling pattern and grossly
reduced amount of explosives, i.e., a lower powder factor. This seemingly
contradictory arrangement is fully justi®ed by using scienti®c principles in blasting,
and converting blasting from an art to a scienti®c discipline.

2. THE LAGRANGE DIAGRAM

A very convenient and illustrative method to present the dynamic behavior associated
with wave and crack propagation is to use the so-called Lagrange diagram. It is a very
simple but extremely powerful method well known to anyone working in mechanics
(see e.g., [13]). In its simplest form where one has a time axis (ordinate) and a position
axis (abscissa) this serves to describe one-dimensional problems. In this paper this
one-dimensional characterisation will be used to show the dynamic development of
stress-wave interaction along the direct connection between two blastholes.
It is known that in a solid various types of waves can propagate. Here, only the bulk
waves and surface waves will be of importance. Within the body of the material two
types of waves can propagate:
P-wave: longitudinal or primary wave/pulse
S-wave: shear or secondary wave/pulse.
These two waves/pulses are of importance in blasting and a full understanding is
required for short time delay calculations as well as optimisation of fragmentation.
Consider, in Figure 1, the ordinate as the time-axis and the abscissa as the spatial
coordinate. Any line through the origin O corresponds to a dynamic event. In Figure 1
the longitudinal P-wave, the S-wave, and the crack are marked with ``P,'' ``S'' and
``C,'' respectively. The tangents of the associated lines are the inverses of the speeds of
the P-wave, S-wave, and the crack, respectively.

2.1. Characterization of a Stress Wave

A stress wave is the propagation of a disturbance in space. Hence, one can describe
the stress wave in two ways, either in space or in time. Figure 2 shows an arbitrary
110 H.P. ROSSMANITH

Fig. 1. Lagrange diagram representing the one-dimensional propagation of longitudinal (P)-wave, shear
(S)-wave and a crack (C).

stress wave of pulse type with ®nite pulse length (Fig. 2a) and ®nite duration (Fig. 2b)
which consists of a leading part and a trailing part. The leading parts are always
characterized by the index ``‡'' and the trailing parts by the index ``ÿ''. The ‡=ÿ
sign has no relationship to either a compressive or tensile pulse; it only indicates the
leading and trailing parts of the pulse. In doing so, it can be applied to any single or
multiple re¯ected pulse with no consideration of the character of the wave/pulse.
One of the important parameters in the new Advanced Blasting Technology is the
pulse length or pulse duration.

Fig. 2. Representation of a one-dimensional stress wave/pulse in the a) space domain and b) time domain.
 THE USE OF LAGRANGE DIAGRAMS IN DELAY TIME DETERMINATION 111

The space description of the pulse is shown in Figure 2a. The pulse length is w ,
the leading part is of length ‡ ÿ
w and the trailing part is of length w . Obviously,
 w ˆ ‡ ÿ
w ‡ w : …1†
The time description of the pulse is shown in Figure 2b. The pulse duration is w , the
leading part is of length w‡ and the trailing part is of length wÿ . Obviously,
w ˆ w‡ ‡ wÿ : …2†
In the space (time) description the wave/pulse is given as a function of position x
(time t). As the different types of waves/pulses travel with different velocities, they
show up in the Lagrangian diagrams as lines of different slope.
The correlation between the space and time description occurs via the wave
velocity: within a time interval dt the elastic wave/pulse travels a distance dx. The
corresponding speeds are cP for a P-wave/pulse (w  P) and cS for a S-wave/pulse
(w  S). These wave speeds are constant if the material is linearly elastic and the
amplitudes are small.
Hence, the following relationships must hold throughout the waves/pulses:
P ˆ cP P …3a†
S ˆ cS s …3b†
for the P- and S-waves/pulses, respectively.
It is convenient to introduce an additional parameter, w , which characterises a sub-
front behind the wave-front. This is shown in Figure 3. The sub-front w is at a
distance  , or the time-interval  behind the wave/pulse front.
Obviously, the following relationships hold:
P ˆ  =P ˆ  =P …4a†
S ˆ  =S ˆ  =S …4b†

Fig. 3. De®nition of sub-fronts in the a) space domain and b) time domain with 0  w  1 and w is either P
or S.
112 H.P. ROSSMANITH

for the P- and S-wave/pulse, respectively. Hence, any position behind a wave front can
be considered a sub-front. Per de®nitionem the wave/pulse is always con®ned
between the front and the end of the wave/pulse.
This one-dimensional method of characterisation can immediately be applied to a
number of dynamic problems which are important in blasting in a rock mass with
widely spaced healed joints.
Consider an isotropic and homogeneous linear elastic material.
a) Detonation of a spherical charge: The velocity of detonation is in®nite and as the
problem is rotationally symmetrical, the dynamics in all directions is the same.
Hence, the problem is a one-dimensional one and the above mentioned description
of the dynamics is appropriate.
b) Instantaneous detonation of an (in®nitely long) column charge: The velocity of
detonation is in®nite and in each cross-sectional plane all directions are equivalent.
Again, a one-dimensional description is appropriate for the waves.
It is important to note, that the term one-dimensional does not imply that there is
only one stress, e.g., r . In spherical blasting problems there is also a non-zero
tangential stress.
Although the appearance of the cracks ruins the overall symmetry, consideration is
here restricted to the situation where the borehole crack runs in the plane formed by
the two blast-holes. The case of crack-wave interaction where a stress wave hits a
crack in an oblique fashion will be treated in Part II of this paper.
Blasting a ®nite length column charge constitutes a fully three-dimensional
problem because of the highly instationary character of the blast event. In fact, the
detonation of a ®nite length column is characterised by three distinctive dynamic
elemental events: the instationary initiation, quasi-stationary or steady-state blast, and
instationary termination of the blast. In real blasting of short columns the intermediate
steady-state phase may be completely absent or hidden in the dynamics of the
instationary parts. In addition, the stress and strain ®eld associated with a ®nite
velocity detonating charge is very complex and becomes dependent on the ratios
between the velocity of detonation and the wave speeds in the rock mass [14; Uenishi
& Rossmanith, 1998]. Nevertheless, for the steady-state phase of constant velocity
detonation the one-dimensional representation is appropriate.
If a plane stress wave propagates in a linear elastic material the shape of the pulse
is invariant and the length of the pulse does not change. For spherical and cylindrical
as well as ®nite detonation speed wave propagation the shape of the pulse is not
invariant but changes shape with distance traveled. This effect is called dispersion. In
the following treatment dispersion effects are neglected.
From the preceding we can conclude that the characterisation of blast-problems
using a one-dimensional Lagrange-diagram representation is useful and will render
important insight into wave-crack interaction.
 THE USE OF LAGRANGE DIAGRAMS IN DELAY TIME DETERMINATION 113

2.2. Stress Waves Emerging From a Detonating Blasthole

In the following it is assumed that the blasthole is in®nitely long and, therefore,
conditions of steady-state detonation will prevail. The results of this consideration
will be utlilized in the discussion of real blasting problems.
Figure 4 gives a graphical illustration of the differences between a long and a short
wave/pulse. Consider a position x from the origin of the coordinate system which can
be identi®ed with the blasthole. Then, in Figure 4a which depicts a short wave/pulse,
the two stress waves/pulses, i.e., the P-wave and the S-wave, have already completely
separated and, if a recording instrument (e.g., a geophone, accelerometer, etc.) was
placed at position x , one would register two individual and completely separated

Fig. 4. (a) De®nition of fronts and ends of emerging P-wave (PF , PE ) and S-wave (SF , SE ) for a short pulse.
(b) De®nition of fronts and ends of emerging P-wave …PF ; PE † and S-wave …SF ; SE † for a long pulse.
114 H.P. ROSSMANITH

Fig. 5. Overlap …t ˆ tph1 † and separation …t ˆ tph2 † of a P-wave and an S-wave radiating from a blasthole
where
P and
S are the lengths of the P-wave and S-wave, respectively.

signals. For the long wave/pulse shown in Figure 4b and the same position x , the P-
and S-waves would not be completely separated but still partially overlap. The latter
situation makes the interpretation of the signal recorded more dif®cult if not
impossible (unless additional information is available).
The same situation can also be studied in connection with taking photographs as
shown in Figure 5. Suppose that two photographs of the dynamic action are taken at
times tph1 and tph2 after detonation of the blasthole.
One can immediately see, that at the earlier time tph1 the two stress waves/pulses
have not completely separated, whereas at a later time tph2 they are clearly separated.
However, one also notices that the wave lengths for the P- and S-waves are different.
This is due to the different wave velocities and the fact, that as long as there is stress
acting at the source, i.e., gas acting at the blasthole wall (at the origin of the coordinate
system in Fig. 5) P- and S-wave creation will continue. As the P-wave and S-wave
information spreads out with different velocities, the wave/pulse lengths will lengthen
due to dispersion. If dispersion effects are neglected in a ®rst order analysis, and it is
assumed the lengths of the pulses does not change with distance traveled, the P- and
S-pulse lengths are different but the P- and S-pulse have the same duration. Under
these assumptions it holds:
P ˆ S ˆ : …5†
This is graphically seen in Figures 4a and 4b.
Employing Equations (3a, b) and (5) one obtains
P =cP ˆ S =cS …6†
 THE USE OF LAGRANGE DIAGRAMS IN DELAY TIME DETERMINATION 115

Fig. 6. Geometrical con®guration of two adjacent blastholes.

which represents a relationship between the wave/pulse velocities and the associated
wave/pulse lengths. For real detonation waves (which are dispersive in nature) the
(here parallel) lines for the fronts and the ends of the P- and S-pulse would (slightly)
diverge when the wave travels from left to right.

2.3. Stress Wave Interaction From Detonation of Two Blastholes

Consider two adjacent blastholes, shown in Figure 6, which are separated by the
spacing s. In the following the wave/pulse interaction between the blastholes will be
studied in the light of fragmentation improvement.

2.3.1. Interaction Between Two Arbitrary Elastic Waves

In a general blast in dif®cult rock which is characterized by complex structural
geology the blastholes in the blast pattern may be charged with different amounts of
explosives. In addition, this amount of explosives may also be arranged in different
ways taking into account possible weak zones, faults, etc., which intersect the blast-
hole or are located nearby. Hence, the charge density distribution along the blasthole
may vary from hole to hole, i.e., the explosives topography and therefore the energy
density distribution in the rock mass may vary considerably.
Nevertheless, in many practical applications (and also from an economic point of
view) most blastholes are charged alike. If, in the simplest case, the explosives charges
in two similar holes are detonated simultaneously, the stress waves will meet ®rst in
the plane of symmetry [15].
Figure 7a shows the interaction of two short wave/pulses, whereas Figure 7b shows
the same interaction for two longer waves/pulses. Generally, several zones of
interaction can be identi®ed:

1) P‡ ‡
1 P2 Interaction of the leading compressive parts of the P-waves (retrograde
hatched)
2) Pÿ ÿ
1 P2 Interaction of the trailing tensile parts of the P-waves (prograde hatched)
3) S1 S2 Interaction of the S-waves (horizontally hatched) and
4) A range of mixed wave interactions, e.g., P1 S2 of the P1 wave of blasthole #1 with
the S2 shear-wave from blasthole.
116 H.P. ROSSMANITH

One recognizes immediately that, in Figures 7a and 7b, the interaction of short
waves/pulses results in small and separated interaction regimes whereas the inter-
action of longer waves/pulses yields spread-out and largely super-imposed interaction
regimes.

Fig. 7. (a) Lagrange diagram showing interaction of short pulses emerging from two simultaneously
detonated adjacent blastholes and the corresponding interaction wave interaction patterns. (b)
Lagrange diagram showing interaction of long pulses emerging from two simultaneously detonated
adjacent blastholes and the corresponding interaction wave interaction patterns.
 THE USE OF LAGRANGE DIAGRAMS IN DELAY TIME DETERMINATION 117

Fig. 8. Identi®cation of interaction locus x of the sub-fronts w1 and w2 for two waves w1 and w2 radiating
from two delayed detonated blastholes.

Assuming stress pulses of equal magnitude in both situations, the potential for
fragmentation spread over a wider zone is much higher for the long pulses than for the
shorter ones. If the blasts produce very long stress waves then the interaction regimes
will cover the entire ®eld between the blastholes and, if the energy density is high
enough, fragmentation is expected to be excellent throughout.
Blasting practice reveals that a short pulse is of higher amplitude than a long pulse.
Hence, the extension of the respective stress ®elds are different and the potential for
fragmentation for the longer pulse will be diminished.
Consider the interaction of two arbitrary waves, w1 and w2 (e.g., a P-wave from
blasthole #1 and a delayed S-wave from blasthole #2 as shown in Fig. 8). It is
convenient to label the corresponding waves, time intervals, sub-fronts, velocities,
etc., with indices `w1 ' and `w2 '. Then, the wave velocities are cw1 and cw2 .
In the general case one has to allow for initiation time delays
01 and
02 , i.e.,
time delays with respect to the origin of the Lagrange diagram. Employing the relative
delay interval

12 ˆ
02 ÿ
01 ; …7†
one obtains from Figure 8 the following condition

01 ‡ w1 w1 ‡ x =cw1 ˆ t ˆ
02 ‡ w2 w2 ‡ …s ÿ x †=cw2 …8†
from which, upon elimination of the time t , general equations for the position and
time of the interaction are obtained in the form:
x ˆ …1 ‡ †ÿ1 ‰s ‡ cw2 …w2 w2 ÿ w1 w1 ‡
12 †Š …9†
ÿ1
t ˆ …1 ‡ † ‰s=cw1 ‡
01 ‡ P1 P1 ‡ …
02 ‡ P2 P2 †Š: …10†
118 H.P. ROSSMANITH

where the abbreviation  ˆ cw2 =cw1 has been introduced.

Employing Equation (7), a time-shifted relationship is obtained:
t ˆ
01 ‡ w1 w1 ‡ x =cw1
ˆ
01 ‡ w1 w1 ‡ …1 ‡ †ÿ1 ‰s=cw1 ‡ …w2 w2 ÿ w1 w1 ‡
12 †Š …11†
Various special cases can be derived from Equations (9±11).

2.3.2. Interaction of Two P-Waves

In many practical cases one wishes to move the zone of intense wave interaction off
the center towards one side or the other. This becomes necessary, e.g., when, in order
to increase fragmentation in the vicinity of a blasthole, the stress waves from one
blasthole interact with the radial borehole cracks from an adjacent hole. This can be
accomplished by delaying the initiation of the blastholes.
Suppose that for the interaction of two P-waves (P1 ; P2 ) the sub-fronts P1 and P2
are expected to meet at at time t at position x (position x and time t in Fig. 8). The
calculation of the distance and time follows from Figure 8.
Specializing (8) yields

01 ‡ P1 P1 ‡ x =cP ˆ t ˆ
02 ‡ P2 P2 ‡ …s ÿ x †=cP …12†
from which, eliminating the time t , follows for the position
x ˆ 12 ‰s ‡ cP …P2 P2 ÿ P1 P1 ‡
12 †Š …13†
and the time
t ˆ 12 ‰s=cP ‡ …P1 P1 ‡ P2 P2 ‡
01 ‡
02 †Š
ˆ
01 ‡ 12 ‰s=cP ‡ …P1 P1 ‡ P2 P2 ‡
12 †Š …14†
If the time axis is shifted by the amount t ÿ
01 the new origin of the time axis
coincides with the initiation of blasthole #1. Given symmetrical conditions one
obtains x ˆ 1=2 s, i.e., the waves/pulses will meet in the plane of symmetry between
the blastholes.
The spread of the interaction regime and the duration of the interaction can be
determined by applying Equations (9) and (10) to the four points A, B, C and D as
shown in Figure 8. In the general case these points have the positions

xA ˆ …1 ‡ †ÿ1 ‰s ‡ cw2
12 Š …15a†

ÿ1
xB ˆ …1 ‡ † ‰s ‡ cw2 …w2 ‡
12 †Š …15b†
ÿ1
xC ˆ …1 ‡ † ‰s ‡ cw2 …w2 ÿ w1 ‡
12 †Š …15c†
ÿ1
xD ˆ …1 ‡ † ‰s ‡ cw2 …ÿw1 ‡
12 †Š …15d†
 THE USE OF LAGRANGE DIAGRAMS IN DELAY TIME DETERMINATION 119

and associated times

tA ˆ
01 ‡ …1 ‡ †ÿ1 ‰s=cw1 ‡ 
12 Š …16a†
ÿ1
tB ˆ
01 ‡ …1 ‡ † ‰s=cw1 ‡ …w2 ‡
12 †Š …16b†
ÿ1
tC ˆ
01 ‡ w1 ‡ …1 ‡ † ‰s=cw1 ‡ …w2 ÿ w1 ‡
12 †Š …16c†
ÿ1
tD ˆ
01 ‡ w1 ‡ …1 ‡ † ‰s=cw1 ‡ …ÿw1 ‡
12 †Š …16d†
The spatial extension of the interaction zone is given by
xB ÿ xD ˆ …1 ‡ †ÿ1 ‰cw2 …w2 ‡ w1 †Š …17†
and the duration of the interaction is given by
tC ÿ tA ˆ w1 ‡ …1 ‡ †ÿ1 ‰…w2 ÿ w1 †Š ˆ …1 ‡ †ÿ1 …w1 ‡ w2 † …18†
Equations (17) and (18) exhibit that the spatial extension as well as the duration of
the interaction are linearly proportional to the wave lengths (this can also be seen from
Figure 8 where the distances AC and BD are directly proportional to the lengths of
either pulse).
From these equations one can also de®ne an `interaction speed' by
vInt ˆ …xB ÿ xD †=…tC ÿ tA †
ˆ cw2 …w1 ‡ w2 †=…w1 ‡ w2 † ˆ cw2 …1 ‡   †=…1 ‡   † …19†
where
  ˆ w2 =w1 : …20†
Notice, that w1 does not have to be necessarily equal to w2 , because the two
blastholes may be different and/or contain different explosive charges which would
in¯uence the structure of the stress wave.

3. WAVE ± CRACK INTERACTION

The preceding equations can also be used to investigate the interaction between any
type of stress wave that radiates from a blasthole and a crack which emerges from an
adjacent blasthole [4, 16±20].
Denote the crack speed with cc and the crack extension with rc . In the Lagrange
diagram a crack shows up with one line only. However, it will normally take some
small but ®nite time interval to incubate and initiate the crack. In the Lagrange
diagram in Figure 9a this is expressed by a small delay
c. The ensuing crack
movement can be rather complex and is best discussed on the basis of the crack path
drawn in Figures 9a±c.
120 H.P. ROSSMANITH

Fig. 9. Formation of blasthole cracks and their interaction with P-wave and S-wave from blasthole #2 (IP ˆ
point of reinitiation due to the P-wave, IS ˆ point of reinitiation due to S-wave). (b) Interaction of a
running blasthole crack at blasthole #1 with wave w2 from blasthole #2.

Upon detonation a strong shock wave is emitted from the blasthole wall. Due to the
radial expansion this shock wave rapidly decays into an elastic stress wave which is
responsible for the initial ground movement in the form of vibrations. The radial
 THE USE OF LAGRANGE DIAGRAMS IN DELAY TIME DETERMINATION 121

Fig. 9. (c) Interaction of an arrested blasthole crack at blasthole #1 with wave w2 from blasthole #2.

extent of the shock wave regime is in the order of a few blasthole diameters and has
been suppressed in Figures 9a±c.
The detonation is associated with the rapid production of a large volume of
pressurized gas which performs work by expanding the blasthole wall perimeter and
by initiating cracks at the blasthole wall. The expansion of the blasthole volume and
the creation and opening of the blasthole cracks enlarges the volume available to the
gas, and thereby reduces the pressure. During the later stage the gas pressure in the
blasthole and the cracks will be further reduced if the gas can vent to open surfaces.
Although the radial blasthole cracks are initially driven by the stress waves, at least
until they are outdistanced by the waves, their further fate is more likely to be
controlled by the gas pressure scenario. With the stress wave effect gone-by and the
gas pressure dropping the stress intensity factor at the crack tips will become smaller
than a critical value, and the cracks will arrest. This occurs at a radius rc . It should be
remarked that, despite decades of intense and devoted research by various research
groups around the world, the details of three-dimensional blast-induced crack-wave
interaction are not completely understood.
Without loss of generality, Figures 9a±c show the dynamic situation which pertains
to the plane comprising and connecting two adjacent blastholes, where it has tacitly
been assumed that a radial blasthole crack is propagating in this plane, i.e., the crack is
directed from hole to hole. For further reference this plane shall be denoted `split
plane.' If the radially emerging crack expands in a plane which is not coincident with
the split plane and is hit by a stress wave one has oblique interaction between a crack
and a stress wave. This situation will be discussed in detail in Part II of this paper.
Suppose that an axi-radial crack propagates in the split plane (marked by the
solid line emerging from blasthole #1. A (longitudinal) P-wave and a slower
122 H.P. ROSSMANITH

S-wave expand from blasthole #2. Various scenarios of crack-wave interaction are
possible:
a) Running crack interacts with P-wave,
b) Arrested crack interacts with P-wave,
c) Crack re-initiated due to interaction with P-wave interacts with S-wave, and
d) Re-arrested crack interacts with the S-wave.
Cases a), b) and d) are shown in Figures 9b, 9c and 9a, respectively.
When the P-wave interacts with the (possibly already arrested) crack, the crack
may reinitiate and extend its length in the split plane. Under certain conditions the
crack may divide into two or more branches; this has regularly been observed in lab
scaled tests with larger amounts of explosives. Crack branching assists the formation
of an improved fragmentation pattern. The fragmentation pattern caused by the
interaction of a set of radially emerging blasthole cracks with a circularly crested
stress wave expanding from an adjacent blasthole is not rotationally symmetrical. The
shape of the resulting fragmentation pattern (damage pattern) leans towards the
adjacent blasthole. Ruling out any anisotropy of the rock mass, the damage pattern is
symmetrical at least in a statistical sense.
When the P-wave from blasthole #2 has swept over blasthole #1 and outdistanced
the crack motions, these blasthole cracks will again arrest because no energy will be
available for further crack propagation. One can safely assume that this fragmentation
pattern is contained in a circle of radius rP . In fact, the maximum extent of this regime
will occur in the split plane as shown in Figure 9a±c.
When the tips of these cracks are hit by the (usually powerful) S-wave they may
reinitiate a second time and extend the fragmentation pattern, which ®nally will be
contained within a circle of radius rS . Again, this interaction pattern results in a non-
rotationally symmetrical fragmentation pattern.
In a real three-dimensional situation the wave front shapes, depending on the velo-
city or detonation, vary from conical to pseudo-spherical for supersonic to subsonic
detonation, respectively [14, 23]. Again, the resulting interaction patterns are not
axi-symmetrically. Hence, the resulting fracture or fragmentation pattern is highly
dependent on the direction of incidence of the stress waves.

3.1. The General Interaction Scenario

The general scenario of crack-wave interaction is shown in Figure 9a where the
position and time of interaction with a delayed P- and S-wave can easily be identi®ed.
In Figure 9a the P-wave and the S-wave impinge on the tip of the crack at times when
the crack tip has already arrested and cause re-initiation (P-wave induced re-initiation
and S-wave re-initiation). The case where the crack upon re-initiation chages the
direction of propagation is treated in Part II of this paper.
 THE USE OF LAGRANGE DIAGRAMS IN DELAY TIME DETERMINATION 123

When the blasthole #1 detonates P- and S-waves propagate into the material and
a set of borehole cracks emerge from the blasthole. These cracks are initiated due to
excessive tangential strain along the borehole wall and once the cracks are born, they
draw energy from the stress waves and from the gas pressurisation of the blasthole.
The cracks typically propagate at a speed lower than 1/3 of the Rayleigh-wave speed,
i.e., approximately 1/6 of the P-wave speed [5]. Hence, the stress waves quickly
outdistance the cracks and the cracks draw their energy from the gas energy.
Crack arrest in conjunction with blasting is a highly complex matter which involves
the joint action of stress waves and the presence of the gas pressure acting on the
fracture walls. Recent ®ndings from well-controlled laboratory scale contained
detonation experiments have shown that the delayed gas front rushes into the radially
expanding crack with a speed close to the speed of sound in the gas with higher
temperature [21, 22].

3.2. Interaction With a Propagating Crack

Suppose that the crack is still propagating when the sub-front w2 of the stress wave
impinges on the crack tip as shown in detail in Figure 9b. The position of interaction
of the running crack tips with the sub-front w2 of the wave w2 follows from the
equivalence

c ‡ r  =cc ˆ t ˆ
12 ‡ w2 w2 ‡ …s ÿ r  †=cw2 …21†
and takes the form
r  ˆ cw2 …1 ‡ cw2 †ÿ1 ‰s ‡ cw2 …w2 w2 ‡
12 ÿ
c †Š …22†
where
cw2 ˆ cc =cw2 : …23†

Interaction takes place at the time t
t ˆ …1 ‡ cw2 †ÿ1 ‰s=cw2 ‡ w2 w2 ‡
12 ‡
c cw2 Š …24†
In general, the ®rst wave to interact with the crack is the P-wave from an adjacent
hole. Hence, Equations (22) and (24) take the form

rP2 ˆ cP2 …1 ‡ cP2 †ÿ1 ‰s ‡ cP2 …P2 P2 ‡
12 ÿ
c †Š …25a†
 ÿ1
tP2 ˆ …1 ‡ cP2 † ‰s=cP2 ‡ P2 P2 ‡
12 ‡
c cP2 Š …25b†
rF
From Equations (22) and (24) also follow the position and the time tF of interaction
of the front of the wave w2 and the crack tip by setting w2 ˆ 0
rF ˆ cw2 …1 ‡ cw2 †ÿ1 ‰s ‡ cw2 …
12 ÿ
c †Š …26a†
ÿ1
tF ˆ …1 ‡ cw2 † ‰s=cw2 ‡
12 ‡
c cw2 Š …26b†
Obviously, it holds tF 
< t and rF <r .
124 H.P. ROSSMANITH

3.3. Interaction With an Arrested Crack

In the next scenario it is assumed that the crack has already come to arrest before its
tip is hit by the stress wave w2 and the crack is re-initiated when the sub-front w2
interacts with the arrested crack tips. This situation is depicted in Figure 9c. In this
case, Equation (21) must be modi®ed and replaced by

c ‡ rac =cc ‡
a ˆ tri ˆ
12 ‡ w2w2 ‡ …s ÿ rac †=cw2 …27†
where
a is the arrest time interval, i.e., the time from arrest until re-initiation due to
the wave w2 . Hence, rac ˆ rri , but the associated times are different, i.e., tac < tri .
It follows that the position is given by
rri ˆ cw2 …1 ‡ cw2 †ÿ1 ‰s ‡ cw2 …w2 w2 ‡
12 ÿ …
c ‡
a ††Š …28†
and the time tri for re-initiation is
tri ˆ …1 ‡ cw2 †ÿ1 ‰s=cw2 ‡ w2 w2 ‡
12 ‡ cw2 …
c ‡
a †Š …29†
If it is assumed that, after re-initiation, the crack grows with cc0 (which does not
necessarily have to be the same velocity as cc ) as long as it interacts with the wave w2 .
The ®nal length, rE , is then determined from …w2 ˆ 1:†

c ‡ rac =cc ‡
a ‡ …rE ÿ rac †=cc0 ˆ tE ˆ
12 ‡ w2 ‡ …s ÿ rE †=cw2 …30†
to yield
rE ˆ …1 ‡ 0 =cw2 †fs ‡ cw2 ‰w2 ‡
12 ÿ …
c ‡
a †Š ÿ rac ÿ1 0
cw2 …1 ÿ  †g …31†
where
0 ˆ cc =cc0 : …32†
and rac ˆ rri is taken from Equation (28).
The time tE associated with secondary arrest after the P-wave has passed is given
by
tE ˆ
12 ‡ w2 ‡ …s ÿ rE †=cw2 : …33†
If the wave w2 is a P-wave, the subscript `w2 ' in Equation (33) can be replaced by `P'
to give
tEP ˆ
12 ‡ P2 ‡ …s ÿ rEP †=cP2 …34†
where rE ˆ rEP and tE ˆ tEP for a P-wave.

3.4. Blast Wave Induced Re-Initiation of Cracks

In general, the crack will interact ®rst with the P-wave and then with the S-wave.
Various scenarios can occur: the wave front may hit the crack tips while it is still
moving or the crack could have come to arrest before the next wave arrives to interact.
 THE USE OF LAGRANGE DIAGRAMS IN DELAY TIME DETERMINATION 125

Fig. 10. Interaction of twice arrested blasthole crack at blasthole #1 with P- and S-waves from blasthole #2.

Figure 10 shows a case of higher order re-initiation of an arrested borehole crack

due to the S-wave interaction (In this case it is secondary re-initiation). Again, the
preceding equations need to be generalized to include the arrest time
aw which
signi®es the time interval between crack arrest just after P-wave passage and re-
initiation when the S-wave interacts. Both these waves originate at blasthole #2.
Similarly, it is assumed that the crack initiates a small time interval behind the S-wave
front passage, i.e., the crack re-initiates when the sub-front S2 S2 hits the crack tip.
From Figure 10 one can see that the S-wave induced re-initiation radius, riS is equal
to the previous arrest radius, i.e., rEP ˆ riS and is determined from a time equivalence

c ‡ rac =cc ‡
ac ‡ …riS ÿ rac †=cc0 ‡
aw ˆ ta1 ˆ
ˆ
12 ‡ S2 S2 ‡ …s ÿ riS †=cS2 …35†
where ta1 is the time of ®rst arrest of the crack tip.
Eliminating the time ta1 and employing the ratios
c0 S2 ˆ cc0 =cS2 0 ˆ cc =cc0
…36†
c00S2 ˆ cc00 =cS2 00 ˆ cc0 =cc00
one obtains from Equation (35) the position of the S-wave re-initiation
riS ˆ …1 ‡ c0 S2 †ÿ1 ffs ‡ cS2 ‰S2 S2 ‡
12 ÿ …
c ‡
ac ‡
aw †Šgc0 S2 ÿ
0
ÿ rac …1= ÿ 1†g …37†
126 H.P. ROSSMANITH

and the associated time tiS ,

tiS ˆ
12 ‡ S2 S2 ‡ …s ÿ riS †=cS2 : …38†
It is interesting to note that, using the equivalence rE ˆ riS in conjunction with
Equation (37) allows the calculation of the arrest time interval
aw which, in Equation
(37), is the only unknown.
During crack ± S-wave interaction the crack may extend with velocity cc00 until
the S-wave has passed. This occurs at time taS at radius rES . These quantities follow
from a generalisation of Equation (35). Regarding Figure 10 one obtains the time
equivalence
0

c ‡ rac =cc ‡
ac ‡ …riS ÿ rac †=cc ‡
aw ‡ …raES ÿ riS †=c00c ˆ taS ˆ
ˆ
12 ‡ S2 ‡ …s ÿ rES †=cS2 : …39†
from which follows for the radius rES
rES ˆ c00 S2 …1 ‡ c0 S2 †ÿ1 fs ‡ cS2 ‰S2 ‡
12 ÿ …
c ‡
ac ‡
aw †Š
0 00
ÿ rac …1 ÿ  †=c ‡ riS …1 ÿ  †=c0 S2 Šg: …40†
The time of crack arrest after the S-wave has passed is given by
taS ˆ
12 ‡ S2 ‡ …s ÿ rES †=cS2 : …41†
After the S-wave from blasthole #2 has passed further crack extension is only
possible due to the stress waves from other adjacent blastholes and re¯ected waves or,
more important, due to the gas pressure in the blasthole. This gas pressure may,
however, have been reduced to a value which is smaller than the one needed to
produce a stress intensity factor at the crack tips to further crack re-initiation.
The effect of gas pressure on crack extension is included in the assumptions of the
crack speeds cc , cc0 and cc00 and does not need to be considered separately.

4. OPTIMIZATION OF FRAGMENTATION BETWEEN TWO BLASTHOLES

In the following precise initiation delay timing (also called Advanced Blasting
Technology) will be applied to optimize fragmentation blasting. An optimal frag-
mentation pattern is achieved if the chain of ensuing operations of loading, hauling,
crushing, grinding, etc. yields a better output. In quarrying the requested output is
usually an amount of broken rock with a speci®ed fragment distribution arranged in
the form of an appropriate muckpile characterized by a certain degree of diggability
for easy handling and hauling. In mining the requested output would be maximum
high grade ore content mass ¯ow in concentration plant, etc. In other situations the
requested output may be different, e.g., less ®ne material (®nes) without loosing the
crusher throughput.
 THE USE OF LAGRANGE DIAGRAMS IN DELAY TIME DETERMINATION 127

In this paper optimal fragmentation is understood in the sense of achieving a

uniform fragmentation pattern with a certain speci®ed fragment distribution around
and between the blastholes during blasting. Meeting the required output is a delicate
problem and a step forward towards achieving this goal is taken if the delay time
between the blastholes in a row (inter-hole delay time) and between the rows (inter-
row delay time) is selected appropriately. Initial trial tests, performed in an Australian
gold mine near Kambalda and several collieries in the Hunter Valley in New South
Wales, by the contractor Thiess & Co., and in Chilean CODELCO-owned copper
mines in Chuquicamata and near Santiago de Chile, revealed that the application of
short delay time initiation resulted in a better blasting product characterised by a more
uniform fragment distribution within a narrow band. Non-ideal delay timing resulted
in an increased number of large bolders left for secondary fragmentation.
Optimisation (in the above-mentioned sense) of fragmentation in a general blast
pattern design is not a trivial matter and requires the study of the interaction of more
than just two adjacent blastholes. In fact, each additional blast will spread out stress
waves which are going to interact with the entire volume of the rock involved in the
blast operation, i.e., the volume of rock already blasted and the volume of rock still
to be blasted. The general problem is fairly complicated because of the non-
commutativity of the creation of fragmentation in blasting.
Despite of the complexity of the fragmentation problem, it can be reduced to a
number of basic problems:
a) the (simultaneous or delayed) interaction of two adjacent blastholes,
b) the (simultaneous or delayed) interaction of three adjacent blastholes arranged in a
triangular pattern, and
c) the (simultaneous or delayed) interaction of four or ®ve adjacent blastholes
forming a centered square (cage).
This complexity will be discussed in detail in Part II of this paper.
Figures 11a and 11b feature the formation of fragments between two adjacent
blastholes. Uniform fragmentation is achieved if the fragmentation pattern around the
blasthole and the wave interaction pattern completely cover the entire volume
between the blastholes. Figures 11a and 11b pertain to two adjacent charges which
initiate simultaneously and slightly delayed, respectively.
Numerous experimental laboratory studies have identi®ed tensile fracturing and
crushing as the main mechanisms of blast-wave induced rock fragmentation. During
the 1970s and 1980s dynamic fracture mechanics and wave propagation in conjun-
ction with dynamic photoelasticity have been applied to blasting problems at several
universities and other research institutions, notably at the University of Maryland and
at Vienna University of Technology. In order to record blast-induced crack-wave
interaction in a brittle material photoelastic laboratory scaled models have been fabri-
cated, dynamically loaded with small explosive charges, and sequences of photo-
128 H.P. ROSSMANITH

Fig. 11. (a) Formation of fragmentation and regions of interaction of P-waves, S-waves and blasthole crack
extension for simultaneous detonation of two adjacent explosive charges. (b) Formation of
fragmentation and regions of interaction of P-waves, S-waves and blasthole crack extension for
delayed detonation of two adjacent explosive charges.

elastic fringe pattern were recorded by means of high-speed cinematographic

cameras. Studying these fringe patterns enabled one to identify various wave-wave
and crack-wave interaction mechanisms [2, 5]. The use of polymers as a working
material was justi®ed in the fact that under short-time loading the material behavior of
the polymers used was strikingly similar to the behavior of rock. This is assisted by
the fact that similar sequences of dynamic events could be predicted (and in fact were
 THE USE OF LAGRANGE DIAGRAMS IN DELAY TIME DETERMINATION 129

Fig. 11. (c) Nomenclature and de®nition of intersection points of two interacting P-waves. (d) Nomen-
clature and de®nition of intersection points of two interacting P- and S-waves. (e) De®nition of
tensile regions and identi®cation of intersections points for two interacting P-waves.
130 H.P. ROSSMANITH

Fig. 11. (f) De®nition of regions of large tensile and shear stresses and identi®cation of intersections points
for two interacting P- and S-waves. (g) De®nition of regions of high shear stress and identi®cation
of intersections points for two interacting S-waves.

observed) for the polymers used and rock. In addition, scaling with respect to the
length of the stress pulse generated by the detonation could easily be achieved by
using small amounts of PETN or lead azide (PbN6) in the model blasts.
 THE USE OF LAGRANGE DIAGRAMS IN DELAY TIME DETERMINATION 131

For the interaction of two (P or S) stress waves the two main mechanisms
highlighted in the sequence of Figure 11a±d have been found:
P-wave interaction: In Figures 11a and 11b the P1 ±P2 interaction regime is marked
by the closed polygon represented by APP ±BPP ±CPP ±DPP .
S-wave interaction: In Figures 11a and 11b the S1 ±S2 interaction regime is marked
by the closed polygon represented by ASS ±BSS ±CSS ±DSS .
P±S-wave interaction: In Figures 11a and 11b the P1 ±S2 and P2 ±S1 interaction
regimes are visible but have not been marked.
Tensile fracturing can occur during the interaction of
a) the trailing tensile tails of the P-waves (in Figure 11c in the regime Gÿ
PP marked
‡
by: CPP ±BÿPP ±C ÿ
PP ±D ÿ
PP )
b) and to some degree during the interaction of a S-wave and the tail of a P-wave
‡
(e.g., in Figure 11d in the regime Gÿ ÿ
SP marked by: DPS ±BPS ±CPS ±DPS ).

Crushing will predominantly take place during the interaction of:

c) the leading compressive parts of the P-waves (in Figure 11c in the regime G‡ PP
marked by: A‡ ‡ ÿ ‡
PP ±BPP ±APP ±DPP ),
d) the S-waves (regime GSS marked in Figure 11b by: ASS ±BSS ±CSS ±DSS ) and
e) to some extent also during the interaction of the S-wave and the head of a P-wave
(e.g., in Figure 11d in the regime G‡ ÿ ‡
SP marked by: APS ±BPS ±BPS ±DPS )

In practice crushing and tensile fracturing will only occur if the wave amplitude
has exceeded a certain threshold. As the compression and tensile part of a stress wave
do interact differently with the rock mass, the entire stress envelope is required (in a
®rst approximation the two threshold values (the compressive fracture stress and the
tensile fracture stress) are important). Hence, crushing and tensile fracturing does not
occur immediately when the wave fronts intersect but will become operative
‡ ÿ
somewhat delayed when the appropriate threshold sub-fronts Pi , Pi and Si interact.
Hence, the aforementioned regimes must be rede®ned in terms of sub-fronts. This
revision of regimes which is shown graphically in Figures 11e and 11g may seem to
be of minor importance, but becomes increasingly important with increasing wave
length and when non-linearities play a major role! From a practical point of view, it
may be almost impossible to determine in advance the position of the relevant sub-
fronts; in this case, an iterative procedure is recommended.
Delayed tensile fracturing may occur during the interaction of
f) the trailing tensile tails of the P-waves (in Fig. 11e regime Gÿ PP marked by:
AÿPP ±B ÿ
PP ±C PP ±Dÿ
PP )
g) and to some degree during the interaction of a S-wave and the tail of a P-wave
(e.g., in Fig. 11f the regime Gÿ ÿ ÿ ÿ
SP marked by: APS ±BPS ±CPS ±DPS ).
132 H.P. ROSSMANITH

Delayed crushing will predominantly take place during the interaction of:
h) the leading compressive parts of the P-waves (regime G‡ ‡
PP marked by: APP ±
‡ ‡ ‡
BPP ±CPP ±DPP ),
i) the S-waves (regime GSS marked by: ASS ±BSS ±CSS ±DSS ) and
j) to some extent also during the interaction of the S-wave and the compressive head
of a P-wave (e.g., in the regime G‡ ‡ ‡ ‡ ‡
SP marked by: APS ±BPS ±CPS ±DPS ).

For very long waves these regimes may overlap and the stress ®eld becomes fairly
involved due to multiple superposition.
If the detonation of the second blasthole is delayed, the various regimes discussed
above move closer to the delayed blasthole. This is depicted in Figure 11b for short
wave lengths. Hence, delayed detonation is a means to control the fragmentation
pattern. On the other hand, if the stress waves become longer the regimes overlap and
the potential for enhanced fragmentation increases. Long stress wave pulses are
obtained by using explosives that are characterised by a lower velocity of detonation
and lesser brisance. Such explosives are usually associated with a larger volume of
gas produced.
Consider the case where the length of the detonation wave/pulse is of the same
order of magnitude as the spacing or even longer. This situation was observed in a
nickeliferous gold mine in Kambalda in Western Australia where the length of the
P-wave pulse was in the order of 100 m. In this case the individual waves, i.e., the
P-wave and the S-wave radiating from blasthole #2 and impinging on the radial cracks
emerging from blasthole #1, have not completely separated and the interaction
regimes overlap. This overlap can become essential and it is dif®cult to assign each
wave an individual interaction pattern. Obviously, a more homogeneous and uniform
fragmentation pattern is expected if the stress wave/pulse is very long.
Figures 12a and 12b show the delayed detonation of two adjacent blastholes, where,
for two different spacings, the length of the stress wave/pulse is chosen to be of the
same length as the spacing, respectively. It is shown that the blasthole cracks
propagate at nearly constant speed without arrest because their movement is fueled by
the successive arrival of the P-wave and S-wave of the adjacent blasthole. For the sake
of demonstration, in Figure 12a it is assumed that the cracks come to arrest after the
passage of the adjacent S-wave.
Allowing for an incubation
c time for crack generation, the arrest radius, ra , and
associated time, ta , follow from the geometry in Figure 12. One obtains the time
equivalence

c ‡ rac =cc ˆ ta ˆ
12 ‡ S2 ‡ …s ÿ rac †=cS2 …42†
from which follows for the arrest crack radius rac
ÿ1
ra ˆ …1 ‡ ÿ1
c † ‰s ‡ cS2 …S2 ‡
12
c †Š: …43†
 THE USE OF LAGRANGE DIAGRAMS IN DELAY TIME DETERMINATION 133

Fig. 12. (a) Determination of arrest radius and arrest time for blasthole cracks with spacing s.

The time ta of crack arrest after the S-wave has passed is given by
ta ˆ
c ‡ …1 ‡ c †ÿ1 ‰s=cS2 ‡ …S2 ‡
12
c †Š: …44†
Shortening the spacing of the blastholes does not essentially change the general
situation, the only difference is that the S-wave will begin to interact at an earlier time,
ti . The difference is depicted in Figures 12a and 12b for large and short spacing,
respectively. However, the stress ®elds are totally overlapping and the crack speed
will not change due to the enforcement by the S-wave, as the crack will propagate at
terminal speed at the outset from the blasthole. Similar conclusions can be drawn for
the effect of change of spacing on the arrest time for long pulse/wave length.
Similar conditions rule the game when the initiation of the second blasthole is
delayed. Only when the delay becomes very large will the stress waves (emerging
134 H.P. ROSSMANITH

Fig. 12. (b) Determination of arrest radius and arrest time for blasthole cracks for small spacing.

from blasthole #1 and impinging from blasthole #2) separately interact with the
cracks in the vicinity of the ®rst blasthole. The situation is then governed by the
previous sections which pertain to short pulse/wave length.

5. CONCLUSIONS

The ®eld of blasting is now entering a decade of complete transition from blasting as
an art ± based on experience collected through more than one century of activities in
mining and quarrying ± to become a science-based discipline, the advanced blasting
technology, which relies on modern concepts of fracture mechanics and wave
propagation [11].
When using precise initiation systems in conjunction with electronic detonators in
the drive towards optimization of fragmentation, vibration control and cost reduction,
a new body of knowledge is required. Conventional experience, as nurtured by the
large majority of blast engineers, though a valuable body of information, is of little
value to effectively put the new technology to work. The transition from conventional
blasting to advanced blasting is accompanied by a complete change of paradigm.
In the Advanced Blasting Technology, knowledge of stress waves and fracture
mechanics, together with knowledge of explosives beyond the velocity of detonation
and other factors, such as the powder factor, is required. Blast design engineers will
 THE USE OF LAGRANGE DIAGRAMS IN DELAY TIME DETERMINATION 135

have to be trained in the basics of mechanics in order to acquire a full understanding

of the physical processes that govern blasting and which lead to fragmentation.
Part I of this contribution focuses on surface blasting and features the mechanics
of wave-wave and crack-wave interaction for two adjacent blastholes. It is found
that certain short-term delays for the initiation of the detonation result in optimal
solutions because the short delays make it possible for the stress waves to interact
within the region between the blastholes where uniform and dense fragmentation is
expected.

REFERENCES

1. Lopez, C.J., Lopez-Jimeno, E. and Ayala Carcedo, F.J.: Drilling and Blasting of Rocks. A.A. Balkema
Publishers, The Netherlands, 1995.
2. Rossmanith, H.P.: Collection of Reports Published by the Fracture and Photo-Mechanics Laboratory
of the Institute of Mechanics at Vienna University of Technology (For a list of papers and purchase of
selected reports please contact Prof. H.P. Rossmanith via email: hans.peter.rossmanith@tuwien.ac.at),
1978±2002.
3. Rossmanith, H.P. (ed.): Teaching and Education in Fracture and Fatigue. E&FN SPON (a Chapman &
Hall Publication), London, 1996.
4. Rossmanith, H.P. (ed.): Rock Fracture Mechanics. CISM Course # 275. Springer-Verlag, Vienna, New
York, 1983.
5. Fourney, W.L. et al.: Collection of Reports Published by the Photomechanics Laboratory of the
University of Maryland. (For a list of reports and purchase of copies of selected reports please contact
Prof. W.L. Fourney via email: four@eng.umd.edu), 1973±1987.
6. Thiess, Co.: Private communication, 1999±2000.
7. CODELCO/Enaex: Trial Tests With Electronic Detonators. Private communication. Chuquicamata
Mine, Chile, 2000.
8. Blair, D.P. and Armstrong, L.W.: The Spectral Control of Ground Vibration Using Electronic Delay
Detonators. Fragblast 3 (1999), pp. 303±334.
9. Cunningham, C.V.B.: The Effect of Timing Precision on Control of Blasting Effects. In: Proc. 1st
World Conference on Explosives and Blasting Techniques ``Explosives and Blasting Technique'',
Munich, Germany, 2000, pp. 123±128.
10. Holmberg, R. (ed.): Explosives & Blasting Technique. In: Proc. 1st World Conference on Explosives
and Blasting Techniques ``Explosives and Blasting Technique'', Munich, Germany, 2000.
11. Rossmanith, H.P.: The In¯uence of Delay Timing on Optimal Fragmentation in Electronic Blasting. In:
Proc. 1st World Conference on Explosives and Blasting Techniques ``Explosives and Blasting
Technique'', Munich, Germany, 2000, pp. 141±147.
12. Yamamoto, M., Ichijo, T., Inaba, T., Morooka, K. and Kaneko, K.: Experimental and Theoretical Study
on Smooth Blasting With Electronic Delay Detonators. Fragblast 3 (1999), pp. 3±24.
13. Graff, K.F.: Wave Motion in Elastic Solids. Oxford University Press/Dover Publications, Oxford, 1975.
14. Rossmanith, H.P. (ed.): Fracture Research in Retrospect. Balkema, The Netherlands, 1997.
15. Rinehart, J.S.: Stress Transients in Solids. Hyperdynamics, Santa Fe, NM, USA, 1975.
16. Rossmanith, H.P. and Fourney, W.L.: Fracture Initiation and Stress Wave Diffraction at Cracked
Interfaces in Layered Media: I.- Brittle-Brittle-Transition. Rock Mechanics 14 (1982), pp. 209±233.
17. Kanninen, M.F. and Popelar, C.H.: Advanced Fracture Mechanics. McGraw-Hill, New York, 1985.
18. Atkinson, B.K. (ed.): Fracture Mechanics of Rock. Academic Press, New York, 1987.
136 H.P. ROSSMANITH

19. Clark, G.B.: Principles of Rock Fragmentation. Wiley, New York, 1987.
20. Broek, D.: The Practical Use of Fracture Mechanics. Kluwer Academic Press, Dodretch, 1998.
21. Daehnke, A., Rossmanith, H.P. and Kouzniak, N.: Dynamic Fracture Propagation Due to Blast-Induced
High Pressure Gas Loading. Rock Mechanics Tools & Techniques. In: Proc. 2nd North American Rock
Mechanics Symp., Montreal. Balkema, The Netherlands, 1996, pp. 619±626.
22. Daehnke, A., Rossmanith, H.P. and Schatz, J.F.: On Dynamic Gas Pressure Induced Fracturing.
Fragblast 1 (1997), pp. 73±98.
23. Rossmanith, H.P., Uenishi, K. and Kouzniak, N.: Blast wave propagation in rockmass ± part I:
Monolithic medium. Fragblast 1 (1997), pp. 317±359.