International Journal of Rock Mechanics & Mining Sciences 71 (2014) 208216

Contents lists available at ScienceDirect

International Journal of
Rock Mechanics & Mining Sciences
journal homepage: www.elsevier.com/locate/ijrmms

Effect of double-primer placement on rock fracture and ore recovery

Z.X. Zhang
Department of Arctic Technology, The University Centre in Svalbard, Longyearbyen, Norway

articleinfo abstract

Article history: The double-primer placement is based on the principle of shock wave collision. When two shock waves
Received 19 October 2013 meet each other, the nal pressure is greater than the sum of the initial two pressures. Stress analysis
Received in revised form indicates that this should be favorable to rock fracture and fragmentation in blasting. This double-primer
20 March 2014 placement was tested in Malmberget mine by using electronic detonators, aiming to improve rock
Accepted 21 March 2014
fragmentation. At the same time, another method, named DRB (Dividing Ring Blasting), was tested, too.
Two production drifts in an ore body were taken as test drifts. In each test drift both methods were tried.
Keywords: For comparison, two nearest production drifts to the test drifts were taken as reference drifts. The results
Blasting showed that on average the double-primer placement recovered more iron ore than either the DRB
Rock fracture method or the ordinary method used in the reference drifts. In addition, fragmentation looked much
Primer placement
ner and the eyebrow break became much less for the double-primer rings, compared with the
Ore recovery
reference rings.
Sublevel caving
Mining & 2014 Elsevier Ltd. All rights reserved.

1. Introduction
The placement of a primer (usually containing a detonator) in a
blasthole plays an important role in rock blasting. Unfortunately,
the importance of primer placement in rock fracture, fragmenta-
tion, and even ore recovery has not been well understood so far.
An improper or even wrong primer position in engineering can
still be found.
As only one primer is placed at each blasthole, the study in [1]
by means of stress wave analysis shows that the best primer
position is the middle of charged blasthole, in terms of detonation
energy efciency, rock fragmentation, and rock break in the roof of
a drift. The production blasts had well conrmed the above
theoretical analysis since the ore extraction and recovery had
been largely increased, and the eyebrow break markedly reduced
by the middle-primer method, compared with the old method
used in the mine [1].
It is common practice for some operators to routinely put two
primers into a blasthole, and their rationale is that using a second
primer is insurance against either a poor initiator/detonator or a
cutoff of the hole due to shifting rock caused by a previous delay
ring [2]. In many mines and quarries, two primers are often
placed in each blasthole. However, in many cases, one primer is
placed at the bottom and the other close to the collar of a
blasthole. The latter is usually taken as a backup in case of that a
malfunction occurs for the bottom primer. In this primer
E-mail address: zongxian.zhang@unis.no
placement, if the two primers are initiated at the same time, the
collar primer will produce serious back break and even bring
about a lot of detonation energy loss. If the collar primer is
initiated later than the bottom one, the result is not good, either.
In brief, a two primer placement with one primer close to collar
should be avoided. The above description indicates that if two
primers are placed in one blasthole, their positions are to be
chosen scientically. In this paper, the double-primer placement
means that two primers with same delay time are placed at
correct positions in a blasthole.
When this double-primer placement is applied to a blasthole, a
collision of shock waves from two primer locations will happen.
Different from elastic wave collision, a shock wave collision results
in that the nal pressure produced is greater than the sum of the
initial two shock waves, according to shock wave theory [3].
An experimental study by Dawes et al. [4] well conrmed this theory
in rock blasting. Their experiments showed that the amplitude of
stress waves in rock mass due to two-primer placement in a
blasthole was much greater than the double of the amplitude of
the waves caused by one single primer in a similar blasthole. Their
experiments indicate potential applications of a two-primer place-
ment in rock blasting. After a long time when electronic detona-
tors came into being, shock collision theory was used to improve
fragmentation at Salvador mine [5]. Even by using NONEL deto-
nators, the shock collision theory was applied to break down
remained roofs in sublevel caving mining [6]. Due to the success in
breaking down remained roofs, this theory was applied to reduce
eyebrow break in Malmberget mine by NONEL detonators [7].
In order to further improve rock fragmentation in the same mine,

http://dx.doi.org/10.1016/j.ijrmms.2014.03.020
1365-1609/& 2014 Elsevier Ltd. All rights reserved.
 Z.X. Zhang / International Journal of Rock Mechanics & Mining Sciences 71 (2014) 208216 209

from 2011 to 2012 two blast methods were tested by using

P (GPa)
electronic detonators, one is DRB (Dividing Ring Blasting) method,
60
originally developed for vibration control [8,9], and the other is
the double-primer placement. In the DRB method, one ring is 55
separated into two parts upper one and lower one in charging
and blasting. The blastholes in the lower part of a ring whose 50
upper part has been blasted in previous blasting are blasted one by 3 (P3, u3)
one with a delay time between two neighboring holes, and then in 45
the same blast the blastholes in the upper part of the next ring are
initiated one by one with a delay time. In this way, the explosive in 40

each delay time can be reduced by about 50% if the blastholes in a

35
ring are divided at their middles. In order to improve the
fragmentation in the upper part of a sublevel ring, a 10 ms delay
30
time between two neighboring holes was employed in the upper
parts of the DRB rings. 25
In terms of the above description, we will briey introduce the
shock collision theory, show how the stress and energy distribu- 20
tions are changed due to shock wave collision, and analyze the
effects of shock collision on rock fracture and fragmentation. Then 2 (P2, u2) 15 1 (P1, u1)
the test results for the double-primer placement in the mine will
be presented and discussed. 10

5
2u2 2u1
2. Theory on shock wave collision 0
-3.6 -3.2 -2.8 -2.4 -2 -1.6 -1.2 -0.8 -0.4 0 0.4 0.8 1.2 1.6 2 2.4 2.8 3.2 3.6
According to one-dimensional shock wave theory [3], when u (km/s)
one shock wave with pressure P1 meets another shock with Fig. 2. Solution of shock wave collision caused by initial two shocks P1 and P2.
pressure P2, the nal shock pressure P3 produced is greater than
the sum of the pressures of the initial two shock waves, i.e.
Since P1 15GPa, we get u1 1:625 km=s from Eq. (2). The
P 3 4 P1 P2 1 Hugoniot curve for the new shock in negative direction must be
rotated around this point and will intercept the P0 0 or u axis at
This case is called shock wave collision. A shock wave collision
2u1 (see Fig. 2), and its equation is
is different from an elastic wave collision. In one-dimensional
condition, as an elastic stress wave with stress 1 meets with P C 0 2u1 u s2u1 u2 3
another elastic wave with stress 2, the nal stress 3 produced is 0 0

equal to the sum of the stresses of the initial two elastic waves, i.e.
Now we consider the Hugoniot curve for the new wave in
3
1 2.
positive-x direction. This Hugoniot is coming from state P2; u2, and
In shock wave collision, the nal pressure depends on both
that u2 is negative; see Figs. 1 and 2. This state was arrived at by
initial two shock pressures and the material. In the following, we
the initial shock B in negative-x direction, into u0 0 material. The
will see how much the nal pressure is increased by shock
Hugoniot curve of this initial shock is
collision, taking TNT (cast) as an example. Assume a shock A with
pressure P1 15 GPa travelling in positive direction in the explo-
sive (one-dimensional), as shown in Fig. 1. In the same explosive P 0C 0u 0 u 0su 0 u2 4
there is another shock B with pressure P2 15 GPa travelling in
negative direction. The Hugoniot values for TNT (cast) are Since P2 15GPa, u0 0, we get u2 1:625 km=s from Eq.
(4). The Hugoniot curve for the new shock in positive direction
0 1:614 g=cm3, C0 2:39 km=s, and s 2:05 [3]. When the two
must be rotated around this point and will intercept the P0 0 or u
shock waves approach each other head-on, the collision will cause
two new shock waves that are reected back in each direction. axis at 2u2 (see Fig. 2), and its equation is
We start with the Hugoniot curve for the new wave in
negative-x direction. This Hugoniot is coming from state P1, u1, P 0C 0 u 2u2 0 su 2u2 2 5
and that u1 is positive, see Fig. 1 and Fig. 2. This state was arrived at
by the initial shock A in positive-x direction, into u The solution for the particle velocity after the collision can be
0 0 material.
The Hugoniot curve of this initial shock is obtained from equating the two Hugoniot Eqs. (3) and (5). This
gives rise to
P 0 C0 u 0su2 2
u3 0 6

Then the pressure at the interaction can be obtained by using

P1 P2
this particle velocity in either Eqs. (3) or (5):

P 3 47:5GPa 7

Obviously, P3 47:5GPa is much greater than the sum

x
P1 P2 30GPa
of the initial two pressures. In brief, the nal
Fig. 1. Coordinate for shock collision. pressure caused by shock wave collision is greater than the sum of
the initial two shock waves.
210 Z.X. Zhang / International Journal of Rock Mechanics & Mining Sciences 71 (2014) 208216
3. Double-primer placement on stress and energy
distributions
3.1. Stress distribution
In rock blasting, all energy used in rock fracture and fragmen-
tation comes from the detonation wave that consists of a leading
shock wave and a rarefaction wave (or Taylor wave). In addition,
rock fracture and fragmentation depend not only on the amount of
energy applied to rock, but also on stress amplitude, stress
distribution, loading rate, etc. Because a shock wave collision can
produce a greater pressure (stress) than the sum of initial two
pressures, the nal stresses in a certain region produced by the
shock collision will be greater than the double of the stresses in
the same region caused by a single primer placement. These greater
stresses are useful for rock fracture.
Now we take one blasthole in a sublevel caving ring as example
to see the stress distribution due to shock wave collision, as shown
in Fig. 3 in which other blastholes in the ring are not shown. In this
blasthole, there are two primers at locations D1 and D2. The
charge length is F1F2 and primer positions D1 and D2 are at 1/3
and 2/3 of F1F2, respectively. There is no free surface in either left
or right side of the lower part of a sublevel caving ring. In the
upper part of the ring, there are two partly-conned surfaces. The
blasthole is fully charged. We assume that: (1) the P-waves shown
in Fig. 3 are all compressive (since detonation is just approaching
the ends of explosive charge); (2) effect of S-waves on stress
distribution is neglected since S-waves start later than P-waves
and they travel much slower than P-waves; (3) a shock wave from
a detonating blasthole quickly decays to an elastic wave in the rock
mass, so shock wave collision will be only considered in the
blasthole, while stress wave superposition will be handled in the
rock mass; (4) the detonation velocity of the explosive and the P-
wave velocity of the rock mass are equal to each other. Under
these conditions, when the two primers are simultaneously
F1
Front of P-wave
D1
E1 O B2
B1 E2
Ring boundary
D2
Front of P-wave
F2
Fig. 3. Stress distribution in the plane of a sublevel caving ring. Explosive is
charged in the blasthole between F1 and F2, and the primers are placed at D1
and D2.
initiated, the detonation from both primers will travel in two
directions: up and down in the hole. When the detonation front
from D1 propagates down to D2, its front in upward direction will
come to F1. At the same time, when the detonation front from D2
propagates up to D1, its front in downward direction will come to
F2. Now we can make a summary on the stress distribution at the
moment shown in Fig. 3, as follows:
There is a superposition region of the two P-waves starting
from the two primers, which is enclosed by curve E1D1E2D2
E1. In this region, the nal stress is greater than either of the initial
two stresses starting from D1 and D2.
In the superposition region E1D1E2D2E1, there are three
areas. The rst is the circular area the diameter of which is D1D2
or B1B2. In this area, a shock wave collision happens, and the
collision begins at location O as soon as the initial two shocks
arrive at this location. According to formula (1), this shock wave
collision results in that the nal shock pressure is greater than the
sum of the initial two shock pressures. Accordingly, the nal
stresses produced by the shock collision in this circular area must
be greater than the sum of the initial two stresses. With increasing
time, this circular area expands outward. This shock collision effect
on the increase in stresses is well conrmed by the eld measure-
ment [4], indicating that the maximum stress caused by two
primers at different positions in a blasthole, which are instanta-
neously initiated, is markedly greater than the double of the
maximum stress induced by one single primer in a similar
borehole.
In the superposition region E1D1E2D2E1, the two other
areas are E1D1B1D2E1 and D1E2D2B2D1. In these two
areas, elastic stress wave superposition occurs, i.e. the nal stress
after superposition is greater than either of the initial two stresses.
As a result, if two primers are placed at two different places in a
blasthole and they are initiated simultaneously, in the rock
surrounding the blasthole, the stresses will be efciently super-
imposed. In some regions the nal stresses by the double-primers
will be greater than the double of the stresses by a single primer;
in other regions the nal stresses will be greater than the stresses
by a single primer. We have noted that Fig. 3 is in the plane of the
ring, so each circle in the gure actually represents a spherical
region in the rock. The stress waves starting from D1 and D2 are in
fact two complex spherical waves, so is the wave (indicated by the
circle between D1 and D2) from the shock collision. The super-
position of these compressive P-waves is favourable to rock
fracture and fragmentation, since in a spherical compressive wave
the stresses in two tangential directions are often tensile.
A number of experimental studies such as in [10,11] have demon-
strated the radial cracks induced by compressive P-wave in
blasting.
In addition to the stresses on the plane of the ring, it is
necessary to know the stress distribution in the cross section
along the axis of drift, as shown in Fig. 4. At the same moment as
in Fig. 3, two compressive P-waves from D1 and D2 are indicated
by two large trimmed circles. At the same time, a small circle
between D1 and D2 is also a compressive wave region due to the
shock wave collision. These three compressive waves will be
efciently superimposed one and another in the region between
D1 and D2, and this superposition will be strengthened with time.
As mentioned above, the superposition of compressive waves is
useful for rock fracture.
As well as we know, a compressive wave will be reected into a
tensile wave at a free surface (also at a partly free surface). Such a
tensile wave is dependent on the original compressive wave, i.e.
the greater the compressive wave, the greater is the tensile wave.
Therefore, a higher nal compressive wave will be reected into
a higher tensile wave. In Fig. 4, we can see the reected wave
occupied by S1M1S3S1 and S2M1S4S2. They are well
 Z.X. Zhang / International Journal of Rock Mechanics & Mining Sciences 71 (2014) 208216
Blasthole
P1
S1 Front of P-wave
D1
S2
O
M1
S3
D2
Front of P-wave
S4
P2
Front surface Drift
Fig. 4. Stress distribution in cross section along axis of drift.
distributed in the ring. With increasing time, the collision-caused
compressive region becomes larger and larger, and relevant tensile
stresses reected from the front face will be greater and greater.
As a consequence, the rock in the ring will be highly fractured and
fragmented.
3.2. Energy distribution
The shock collision makes the total detonation time in the
blasthole shorter, resulting in a greater energy concentration.
As shown in Fig. 4, the total detonation time TD in the hole is
equal to TD lch=3D in the double-primer case shown in Fig. 4, TD
lch=D in the case of one primer at either the bottom or the
collar of the hole, and TD lch=2D in the case of one primer at the
middle of the hole. Among these three cases, the double-primer
case gives the shortest detonation time, meaning that the total
detonation energy is distributed in a shortest period of time,
compared with two other cases. This high energy concentration in
the double-primer case should be helpful for rock fragmentation.
At last, we note that at the same moment as in Fig. 4, when all
explosive in the hole is detonated, no tensile wave reaches the
eyebrow and roof of the drift. In other words, the eyebrow break
and rock fracture in the roof can be reduced by the double-primer
placement.
4. Production tests on double-primer placement
4.1. Location of tests
In order to improve fragmentation, it was planned to test the
DRB method rst and the double-primer one second. Considering
as small effect of geological factors on blasting as possible, four
neighboring production drifts in a production level were taken as
the test area. Two middle drifts 496 and 499 are planned to use
the two test methods, and two others 493 and 502 as reference
drifts. The locations of these drifts are shown in Fig. 5. In each
drift, blasting begins from hanging wall. Therefore, ring 1 (R1) is
211
Hanging wall
R1
R2
R3
DRB
Double
primers
Drift 504 Drift 502 Drift 499 Drift 496 Drift 493 Drift 491
(reference) (reference)
Foot wall
Fig. 5. Location of blast tests on a production level.
always blasted rst, and the others follow one and another in
sequence.
4.2. Blast plan
In theory, as shown in Figs. 3 and 4, in order to achieve good
fragmentation, two primers are to be placed at 1/3 and 2/3
charged parts of each blasthole. However, considering a sublevel
ring is a fan form of borehole distribution, the specic charge in
upper part of the ring is much smaller than that in the lower part.
Thus, in real blast plans, we often place the two primers at higher
positions than that shown in Figs. 3 and 4. A real charge plan is
shown in Fig. 6 where the primer positions are shown by small
squares. Also, the upper primer in the middle blasthole is at 30 m
high from the drift roof due to limited length of detonator wires.
Anyway, the position of a primer in the double-primer placement
can be changed more or less, according to practical situations.
4.3. Delay time
According to our previous blast tests with electronic detona-
tors, we chose 80 ms as the delay time between two neighboring
blastholes. In this case, when one hole is blasted, there should be
no stress wave superposition from its neighboring holes. However,
the shock collision due to double-primer placement will make the
stresses in the rock mass increased largely, as analyzed previously.
5. Results from double-primer placement in sublevel caving
5.1. Rock fracture in eyebrow region
As mentioned previously, the double-primer placement was
successfully used to reduce eyebrow break in sublevel caving by
using NONEL detonators in the same mine Malmberget [7]. The
result indicated that the eyebrow break was reduced from 3 m in
ordinary rings (in which the same primer placement was used as
in the reference rings in this study) to 2.1 m in the double-primer
rings. Note that the burden is 3.5 m in both ordinary and double-
primer rings. As electronic detonators were used in this study,
result became even better, due to the accurate initiation of
electronic detonators which makes shock collision effect better.
Fig. 7a shows the result for eyebrow break from one double- primer
ring in which a large part of burden is remained after blasting. This
result was found in most of the double-primer rings we had tested.
In this case, all of collars are visible. Therefore,
212 Z.X. Zhang / International Journal of Rock Mechanics & Mining Sciences 71 (2014) 208216

Planned hole
Actual hole
Charge

2m
2m
2m
2m
2m

Fig. 6. Blast plan of ring 496-r24. The extraction is 160.4%. Sublevel height 30 m;
drift spacing 24 m; burden 3.5 m; borehole diameter 115 mm.

chargers do not need to go to the caving area to nd the collars

during charging. In other words, the work safety for miners is
much better. As a comparison, Fig. 7b indicates the result for
eyebrow break from a reference ring in which a small part of
burden is remained.

5.2. Rock fragmentation

According to the author's eld observations, the fragmentation

from most double-primer rings looked ne, as the muckpile shown
by the picture in Fig. 8a that was taken after blast and when loading
did not start yet. We can see that the maximum fragments are Fig. 7. Remained burden. (a) ring 499-r22 (double-primer ring in drift 499);
smaller than the normal A4 paper. Such ne fragmentation was (b) ring 502-r17 (reference ring with one primer in each hole).

seldom found in the ordinary rings or reference rings in the mine.

Note that we did not increase the explosive charge in each hole in
the double-primer placement, i.e. the borehole plans were not
changed at all. As a comparison, the muckpile from one reference
ring is shown in Fig. 8b.
5.3. Ore ow and dilution
In the reference drift 502, a total of 19 rings were blasted.
In test drift 499, a total of 12 DRB rings and 7 double-primer rings
were tested. The results for ore extraction and recovery are shown
in Fig.10b, indicating both extraction and recovery in the double-
primer rings are higher than that in either reference or DRB rings.

Due to improved fragmentation in the double-primer rings, ore 6. Discussion

ow became better and dilution was low in several rings, as shown
in Fig. 9, where dilution is only 11% and 8%, respectively. From
Fig. 9b we see that the ore recovery is up to 147%, meaning that at
least 47% of the recovery is from one or more rings above ring 496-
r24. The major reason is that average sizes of fragments in the
double-primer ring became smaller, resulting in the remained ore
over this ring could ow down.
5.4. Ore recovery
Table 1 shows the results for ore extraction and ore recovery
from all of test and reference rings. A total of 17 rings were blasted
in the reference drift 493. In test drift 496, a total of 14 DRB rings
and 5 double-primer rings were tested. The results for ore
extraction and recovery are shown in Fig.10a, indicating both
extraction and recovery in the double-primer rings are higher than
that in either reference or DRB rings.
6.1. Rock fracture and fragmentation
As mentioned previously, the fragmentation from the double-
primer rings often looks ne according to our eld observations.
The major reason is that the double-primer placement shortens the
total detonation time, and largely increases the nal stresses in
rock. These make the total detonation energy highly concentrated
in a shorter time and with higher amplitude, compared with a
single-primer placement. As a result, rock mass is highly shattered
in the double-primer placement.
On the other hand, one may inquire what role loading rate may
play in the double-primer placement. It is no doubt that the
loading rate in the double-primer placement is higher than that in
a single-primer case, since detonation time is largely shortened. It
is also known that fracture toughness, compressive strength and
tensile strength of rock increase markedly with increasing loading
 Z.X. Zhang / International Journal of Rock Mechanics & Mining Sciences 71 (2014) 208216 213

Production drift: AL992-499; Ring No.22

Average iron content in last 25% extraction

Iron content (%)

Iron content (%)
Extraction (%)

Production drift: AL992-496; Ring No.24

Average iron content in last 25% extraction

Iron content (%)

Iron content (%)

Extraction (%)
Fig. 9. Iron content vs. ore extraction. (a) ring 499-r22. Ore recovery 90%;
extraction 101%; waste rock 11%. (b) ring 496-r24. Ore recovery 147%; extraction
160%; waste rock 8%.

Table 1
Ore extraction and ore recovery from all of test rings and reference rings. Rings 6
19 in drift AL496 with DRB, and rings 2024 with double-primer; Rings 819 in
drift AL499 with DRB, and rings 2026 with double-primer. Drifts AL493 and AL502
Fig. 8. Muckpile after blasting. (a) ring 499-r25 from double-primer placement. are for reference.
Middle part to upper part of the muckpile (the paper is a piece of normal A4);
(b) ring 493-r20 from ordinary blasting (the paper is 40 cm ~ 50 cm). Ring no Ore extraction (%) Ore recovery (%)

AL493 AL496 AL499 AL502 AL493 AL496 AL499 AL502

rate under dynamic loads [1216], indicating rock is stronger in

6 60 135 51 130
higher loading rate. However, although rock becomes stronger or 7 79 79 69 56
more difcult to fracture, the sizes of fragments produced decrease 8 56 106 112 95 46 78 88 87
with increasing loading rate under dynamic loading condition 9 77 86 101 73 57 70 74 62
10 100 97 86 80 85 86 63 70
[15,17], and the internal cracks produced increase with increasing
11 86 77 68 91 60 61 60 65
loading rate, too [18]. Therefore, a higher loading rate in dynamic 12 115 83 123 84 68 75 102 70
loading condition is favorable to rock fragmentation, if the effect of 13 96 93 70 84 58 86 64 75
loading rate on energy efciency is not considered in blasting. 14 99 95 113 95 57 78 82 74
A good fragmentation, i.e. smaller average sizes of fragments, 15 141 136 84 82 97 94 72 71
16 116 90 86 116 86 71 65 104
can delay the moving down of the waste rock above the ring in
17 135 99 85 107 90 76 60 92
loading and therefore increase ore recovery and reduce dilution, 18 92 138 108 92 54 95 82 73
according to previous tests in the mine [1,20]. In addition, a good 19 119 123 83 107 91 93 66 89
fragmentation can save a signicant amount of energy in down- 20 105 95 111 105 88 71 90 91
21 112 158 111 120 67 127 90 97
stream operations such as crushing and grinding. Furthermore,
22 199 132 101 75 170 65 90 66
a good fragmentation can increase extraction speed or productiv- 23 125 128 115 94 108 93
ity, according to production tests [19,20]. In this direction, it is 24 160 144 100 147 91 86
better to measure or evaluate fragmentation and follow up loading 25 92 268 55 220
operation in the double-primer rings in the future. 26 233 129 172 108

Mean (%) 105.1 110.9 107.3 106.2 76.1 87 82.8 89.1

6.2. On rock fracture in roofs

It has been proved that the double-primer method can ef- also the previous tests [7]. Therefore, the double-primer method
ciently reduce rock fracture in a drift roof or eyebrow break in can be certainly used to reduce eyebrow break or rock fracture in a
sublevel caving, according to not only the tests in this study but drift roof.
214 Z.X. Zhang / International Journal of Rock Mechanics & Mining Sciences 71 (2014) 208216

Production drift:AL992-499; Ring No.25

Extraction and recovery from 496& ref. 493
Extraction
140
134%
Average iron content
Recovery

Iron content (%)

Iron content (%)

in last 25% extraction
120

105.1% 102.6%
100.8%
100

82.1%
80 76.1%

60

Extraction (%)
40
Fig. 11. Iron content vs. ore extraction for ring 499-r25. Fragmentation is ne, but
extraction is low.
20

0
Ref. DRB 2 primers Ref. DRB 2 primers
493 in 496 in 496 493 in 496 in 496

Extraction and recovery from 499 & ref. 502

Extraction
140
131.4%

120
Recovery
106.2%
100 99.4%
93.3%
89.1%

80
73.2%

60

Fig. 12. A fault slides toward hanging wall behind ring 499-r24.
40

20
0 The second phenomenon is that the ore extraction was stopped
Ref. DRB 2 primers Ref. DRB
493 in 496 in 496 493 in 496
Fig. 10. Ore extraction and recovery. (a) from test drift 496 and reference drift 493;
(b) from test drift 499 and reference drift 502.
6.3. Special phenomena
During the tests for the double-primer method, some special
phenomena were found. The rst phenomenon is that sometimes
when fragmentation is good, ore recovery is not. For example, two
rings 499r24 and 499r25 showed a very good fragmentation, as
shown in Fig. 8a. However, the ore recovery in both rings is not
good, as shown in Fig. 11 and Table 1. Fig. 11 indicates that after
10% of extraction, a great amount of waste rock must have come
down. The major reason is that as ring 499-r24 was charged, a
fault was found after 499-r25. The fault slid toward the hanging
wall, as shown in Fig. 12 in which the left side is toward hanging
wall. During blasting in the two rings, the fault might slide a
certain distance to the hanging wall. This may cause the unblasted
ore mass behind ring 499-r24 or r25 and over the fault move a
distance toward the hanging wall. This movement pushed ore
fragments forward and reduced the width of ore ow in horizontal
direction. In this case, if loading continues, there must be much
more waste rock coming out. From Fig. 11 we can see that the iron
content is only about 40% after extraction is over 10%, indicating
that a great amount of waste rock has been loaded out. A similar
result is found in ring 499-r24, i.e. dilution is also high. In this case,
a special blast plan is to be developed in the future.
2 primers
in 496 incorrectly in a number of rings (particularly in many DRB rings),
as shown in Fig. 9, although the extraction and recovery are
already high. A common character is that a suddenly hanging up
is formed by several boulders that block the draw point. In this
case, especially as iron content is still very high, extraction should
continue, for example, by breaking down the hanging up. Other-
wise, ore loss will be unnecessarily increased.
Another phenomenon from DRB rings is that fragmentation
often looked ne and the loading in the rst 30% extraction was
almost no waste rock. However, from about 30% (sometimes 40%)
extraction there were usually some waste boulders came down.
A possible reason is that the short delay times (10 ms in upper holes
and 2060 ms in lower ones) between two neighboring holes might
give rise to a too strong fragment throw, making the waste rock
boulders dropping down during the throw. This phenomenon is
worthy of being considered in a further study on the double-primer
placement since we are not sure whether 80 ms is or not the optimum
delay time used in this study, at least in theory.
6.4. Ore recovery with double-primer placement
As reported in [7], the ore recovery was increased by 3% in
fourteen double-primer rings (average recovery 97%) by using
NONEL detonators in Malmberget mine, compared with that in
seventeen ordinary rings (average recovery 94%).
 Z.X. Zhang / International Journal of Rock Mechanics & Mining Sciences 71 (2014) 208216
In this study, the results for test drift 496 and reference drift
493 showed that the ore recovery was 108%, 76.1%, and 82.1% from
the double-primer rings, reference rings, and the DRB rings,
respectively. Similarly, for test drift 499 and reference drift 502,
the recovery was 99.4%, 89.1%, and 73.2% from the double-primer
rings, reference rings, and the DRB rings, respectively. On average,
the double-primer placement shows more than 10% ore recovery
than either of two other methods.
We note that the ore recovery in the last few rings in either
reference drifts or test ones is high, compared with that in other
places. This is because these rings are located close to the foot wall
where it is easy for the remained ore in upper levels to move down
during loading. Considering this situation, let us compare the
recovery from all of the double-primer rings and corresponding
reference rings. We consider ve last rings in reference drift 493
and seven last rings in reference drift 502, so the total reference
rings are twelve. We take away one ring AL502-r25 with max-
imum recovery and another ring AL493-r18 with least recovery
from these twelve reference rings. At the same time, we take away
one double-primer ring AL499-r26 with maximum recovery and
one ring AL499-r25 with least recovery from twelve double-
primer rings. Then the result is that the average recovery from
10 double-primer rings is 97.3% and that from 10 reference rings is
95.7%, indicating a small increase in the recovery of the double-
primer rings. This result does not consider the negative effect of
the fault on the recovery of double-primer ring AL499-r25, as
discussed above. Anyway, although the ore recovery from the
double-primer rings is increased to a certain extent, a further
study on a large quantity of double-primer rings is necessary in
the future since the number of the double-primer rings in this
study is still small, only twelve rings.
6.5. High stresses in structures nearby and high vibrations in far-eld
As shown in Fig. 4, the shock collision induced high stress wave
in the half-circle with radius OD1 will completely propagate into
the rock mass behind the blasting ring (right side of picture).
As analyzed in Sections 2 and 3, the maximum stress in this wave
can be up to more than double of the peak stress from a single
primer placement. If this wave is very strong, a possible damage in
the rock structures nearby and very high vibrations in the far-eld
may be caused. After that moment, as shown in Fig. 4, the reected
waves will immediately follow the collision-induced compressive
waves, i.e. a superposition between the compressive waves and
the tensile waves (reected ones) may happen. At the same time,
rock fracture occurs and the extending cracks interact with the
stress waves in the collision-affected region. These of course
consume wave energy. As a consequence, the waves going into
the surrounding rock mass in the near or far eld will be reduced
more or less. In addition, approximately, the blasting-induced
stress waves are a spherical wave the geometrical attenuation of
which is great, i.e. with increasing distance the amplitude of the
wave decreases fast. Thus, it is possible that more detonation
energy might be well consumed in shattering the rock mass
surrounding the primers in the double-primer case, while less
detonation energy could be transported to the rock structures
nearby or the far-eld.
For the nearby rock structures, the maximum vibration induced
by the double-primer placement could be more than double of
that caused by a single-primer placement. But in sublevel caving
the nearby rock structures are the production drifts. A two times
high vibration might not seriously undermine the safety of the
drifts. In Malmberget mine, since year 2006 a total of thirty-seven
double-primer rings have been tried in different ore bodies, but no
any evident rock damage in the drifts has been found, as seen in
Fig. 7a. Therefore, for underground mining, especially sublevel
215
caving, the double-primer method would not cause a marked
damage to the nearby rock structures. But when this method is
used in other engineering projects, for example in the case of a
weak nearby structure, it cannot be excluded that the double-
primer placement may cause damage in the structure. Therefore, a
careful blast design together with some measurements, for exam-
ple, on vibrations is necessary. Even for the future study and tests
on the double-primer placement in sublevel caving some mea-
surements and deeper theoretical studies are recommended.
For the vibrations in the far-led, the double-primer method
could cause a very high vibration. In this study, the test area is
much far from the Malmberget city, so the blast-induced vibra-
tions in the city never cause any problems. But in the mine, one
production area named Johannes is the nearest one to the city.
Considering this situation, the double-primer method has been
never tried in Johannes. However, when two middle holes in a ring
were simultaneously initiated or two rings from two different
drifts in the same ore body were initiated at the same time, a
nearly two-time high vibration was monitored in the city, com-
pared with the normal blasting when one hole or one ring is
initiated at each time [9]. Therefore, in this case or a case similar to
this, the double-primer method is not recommended. Otherwise, if
this method is needed to be tried, some pre-tests and relevant
measurements are necessary.
6.6. Comments on the double-primer placement
The double-primer placement has been proved to be successful
in reducing eyebrow break, improving rock fragmentation and
increasing ore recovery in the Malmberget mine. So far we have
not found any negative effects of this method on mining produc-
tion and safety. Since the double-primer rings tested in this study
are not many and no measurement on fragmentation and blast
parameters was done due to limited research resource during this
study, more tests with relevant measurements are necessary.
In sublevel caving or other similar situation, the delay time
between neighboring holes needs to be carefully chosen. A very
short delay time may not produce a high ore extraction and
recovery, even though fragmentation is good, according to the
DRB test (some blastholes in the lower part of ring also have
double-primers) in this study. The reason is that a too short delay
time may give rise to a strong fragment throw and compaction to
the materials (mostly waste rock) in the front of the blasting ring.
As a consequence, this may cause the waste rock above the ring to
drop down during the throw. In addition, this may make ore ow
become worse. Therefore, a study on the relation between ore ow
and blast parameters such as delay time and primer placement is
needed in the future.
7. Conclusions
The double-primer placement based on shock wave collision
theory is successful in reducing eyebrow break in sublevel caving
and rock break in the roof of a drift, by using either electronic or
NONEL detonators. By employing the double-primer placement
with electronic detonators, rock fragmentation looks much ner or
better according to eld observations. The double-primer rings
result in a certain increase in ore recovery, compared with the
reference rings and the DRB rings.
Acknowledgements
The tests on the double-primer placement were done in LKAB
Malmberget mine. The author is grateful to the R & D and
216 Z.X. Zhang / International Journal of Rock Mechanics & Mining Sciences 71 (2014) 208216
production departments, especially for the chargers in the mine
for their active support.
References
[1] Zhang ZX. Increasing ore extraction by changing detonator positions in LKAB
Malmberget mine. Int J Blast Fragment 2005;9:2946.
[2] Konya CJ, Walter EJ. Rock blasting and overbreak control. US Department of
Transportation, Federal Highway Administration; 1991.
[3] Cooper PW. Explosives engineering. New York: Wiley-VCH; 1996.
[4] Dawes JJ, McKenzie CK, Liddy TJ. Interaction between blast design variables:
experimental and modelling studies. In: Proceedings of the rst international
symposium on rock fragmentation by blasting, Lule, Sweden; 1983. Vol. 1,
p. 265287.
[5] Quezada DB. Economic benets associated with Fragmentation using elec-
tronic detonators. In: Proceeding of the eighth international conference on
rock fragmentation by blastingFragblast-8, Santiago, Chile; 2006. p. 221-222.
[6] Zhang ZX, Naarttijrvi T. Applying fundamental principles of stress waves to
production blasting in LKAB Malmberget mine. In: Proceeding of the eighth
international conference on rock fragmentation by blastingFragblast-8,
Santiago, Chile; 2006. p. 369374.
[7] Zhang ZX. Reducing eyebrow break caused by rock blasting in Malmberget
mine. Blast Fragment 2011;5:110.
[8] Zhang ZX, Naarttijrvi T. Reducing ground vibrations caused by underground
blasts in LKAB Malmberget mine. Int J Blast Fragment 2005;9:6178.
[9] Zhang ZX. Controlling vibrations caused by underground blasts in LKAB
Malmberget mine. In: Proceeding of the tenth international conference on
rock fragmentation by blastingFragblast-10, New Delhi, India; 2012. p. 249
254.
[10] Johansson CH, Persson PA. Detonics of high explosives. London: Academic
Press; 1970.
[11] Olsson M, Ouchterlony F. New formula for blast induced damage in the
remaining rock. . Stockholm: SveBeFo Rapport; 2003.
[12] Lindholm US, Yeakley LM, Nagy A. The dynamic strength and fracture
properties of dresser basalt. Int J Rock Mech Min Sci 1974;11:18191.
[13] Cho SH, Ogata Y, Kaneko K. Strain-rate dependency of the dynamic tensile
strength of rock. Int J Rock Mech Min Sci 2003;40:76377.
[14] Dai F, Xia K, Tang L. Rate dependence of the exural tensile strength of
Laurentian granite. Int J Rock Mech Min Sci 2010;47:46875.
[15] Zhang ZX, Kou SQ, Yu J, Yu Y, Jiang LG, Lindqvist P-A. Effects of loading rates on
rock fracture. Int J Rock Mech Min Sci 1999;36:597611.
[16] Olsson WA. The compressive strength of tuff as a function of strain rate from
10 6 to 103/s. Int J Rock Mech Min Sci 1991;28:1158.
[17] Liu S, Xu J. Study on dynamic characteristics of marble under impact loading
and high temperature. Int J Rock Mech Min Sci 2013;62:518.
[18] Zhang ZX, Kou SQ, Jiang LG, Lindqvist PA. Effects of loading rates on rock
fracture: fracture characteristics and energy partitioning. Int J Rock Mech Min
Sci 2000;37:74562.
[19] Orlandi C. MeKenzie C. The impact of blasting on the business of mining. In:
Proceeding of the eighth international conference on rock fragmentation by
blastingFragblast-8, Santiago, Chile; 2006. p. 2026.
[20] Zhang ZX. Impact of rock blasting on mining engineering. In: Proceedings of
fth international conference & exhibition on mass mining, Lule, Sweden;
2008. p. 671680.