TAYLOR, D.L. and FIRTH, LR. Utilization of blast movement measurements in grade control.

Application of Computers and Operations Research in the

Minerals Industries, South African Institute of Mining and Metallurgy, 2003.

Utilization of blast movement measurements in grade control

D.L. TAYLOR and LR. FIRTH
Department of Mining Engineering, Mackay School of Mines, University of Nevada, Reno, Nevada, USA

Metal mining operations, in particular gold mining operations, use intensive ore control
procedures in order to manage the extraction of ore zones from benches containing both ore and
waste material. Typical grade control procedures rely on blast hole samples to indicate ore grade.
Statistical techniques and geological controls are then used to create two-dimensional polygons,
which indicate zones of material designated as ore. These ore-polygons are calculated using pre-
blast material locations and are generally not corrected for rock movement caused by blasting.
This lack of correction appears to exacerbate the material mischaracterization problem of sending
ore to the waste dump and/or waste to the mill. One Nevada gold mining company has estimated
that this could be a $10 million per year problem for their operations.
U sing a simplified vector field algorithm, a method of taking point movement data and
interpolating post-blast positions for all nodes in a polygon has been developed. Currently
working in 2-dimentions (X-Y), this algorithm utilizes the movement vectors calculated from pre-
and post-blast positions of targets placed in a blast, to generate a spatial transformation function
for any point on the bench. This transformation is then applied to the nodes of the ore-polygons to
produce modified dig lines ore polygons reflecting distortion and movement resulting from the
blast. The method as implemented uses common mine planning routines and is performed using
the Surpac software suite. The method may be adapted to use any pre-and post-blast point
movement data available, and most mine planning suites include the simple interpolation
algorithms used.
Keywords: Blast movement; Ore control; Open pit blasting; Grade control

Introduction being misclassified. The classical misc1assification occurs

For open pit mining operations in which ore and waste when ore material is moved out of the ore polygon and is
materials are intimately mixed within the orebody, accurate marked as waste, and conversely, waste material is moved
separation of these materials during loading operations is into the ore polygon and is marked as ore. Hence, metal
critical. In many such operations, ore and waste materials value is lost as ore is loaded to the waste dump, and milling
cannot be identified solely on the basis of rock type, cost per unit metal recovered is increased as waste material
structure or colour, and ore waste boundaries will be is loaded to the mill.
detelmined on the basis of an economic cut-off grade. An Recently, several mining operations report the use of the
example of this type of orebody is the hydrothermally 'mine to mill' concept in optimizing total cost per unit
altered, disseminated gold deposits of Nevada and the ore- metal recovered. Part of this methodology is to produce a
bodies of the western Australian goldfield, as discussed by product in the open pit, which is designed to reduce
Shawl and Zhang2. downstream costs and total unit cost, rather than simply
The majority of these mining operations rely on blast hole minimizing the cost of the drilling and blasting operation.
sampling to identify the ore grade at a point, these data This method, while resulting in substantial savings in
arethen used to identify zones of ore and waste within a crushing and milling costs according to Bulow3 , has in
given bench horizon, termed ore/waste polygons. These several cases also resulted in increased dilution and poor
polygons are normally calculated using a computer-based reconciliation between head grade and polygon grade. This
interpolation scheme and may be of any shape and size. increased dilution is thought to be due to the increased
After the locations of ore and waste polygons in a particular movement experienced by ore polygons during blasting.
bench have been calculated, the bench is blasted and the ore As part of an ongoing study of blast movement and its
and waste is marked on the surface of the muckpile using a measurement2--7, this paper describes the development of a
system of stakes, flags, ribbon, placards and tape. It is simple, effective means of using movement data for discrete
important to note that these material polygons are two- points in a blast, to account for blast movement and re-
dimensional shapes and are marked on the muckpile using calculate ore polygon shapes post blast. The algorithm can
X-Y coordinates only, with no reference to elevation. be used with any point movement data set and can be
The location of ore and waste in a given blast is adapted to anyone of the popular geostatistical
calculated based on the position of the preblast drill interpolation methods.
samples. The polygon marked on the muckpile is also based
on preblast sample locations and as such, any movement of Background
rock in the X-Y plane due to blasting will result in material This section serves as a discussion of common approaches

UTILIZATION OF BLAST MOVEMENT MEASUREMENTS IN GRADE CONTROL 243

 to minimizing ore movement due to blasting, and methods Methods of measuring blast movement include the use of
used to gather point movement data of the kind required for drill tube, chain, paint, dye, chalk, mill balls, paint cans and
input to the algorithm described later in the paper. polyethylene pipe embedded in un-charged drill holes or in
For mining operations which are sensitive to movement the stemming of a blasthole and paint lines, flags, tape and
of ore polygons due to blasting, methods have been stakes on the blast sUlface. The purpose of these methods is
developed which limit movement in an X-Y plane during to provide a surveyed pre-and post-blast position for a
the blast. The most common approach to this problem is the number of points in the blast area. These data can then be
choke or buffer blast, in which the blast is fired into a used to characterize the movement experienced by the ore
buffer of previously broken rock, hence free face movement contained in the bench. Problems with these simple
is hindered. This method is relatively effective, however, methods include data loss due to destmction or movement
field studies indicate that the effect of the buffer may be of markers during blasting and loading, failure to find a
minimal for blastholes two or three rows back from the free significant number of markers and issues surrounding the
face. When choke blasting in competent rocks, increased reliability of data for objects which may not have minored
powder factor may be required and this generally leads to a the movement of the sunounding rock mass.
large amount of vertical heave. Vertical heave and its effect Zhang2, Taylor4 and Gilbride 5 report moderate success
of ore dilution is difficult to quantify and account for, and using blast hole liner bags filled with rock chippings and
should be minimized whenever possible. placed in uncharged drill holes. These bags are marked to
Blast sequence can also be used to affect the degree of indicate the hole number and the depth from the surface of
dynamic relief available to a particular blasthole or row of the hole at which they are placed. After the blast, as loading
holes. Short inter-row and inter-hole timing may be used to proceeds, bags appearing in the digging face are surveyed
create reasonable fragmentation while minimizing throw of and hence a movement vector can be calculated for each
the muckpile. Again, if the powder factor for the blast is bag. The significant disadvantage of the method is that
elevated to ensure fragmentation in a more competent rock movement data is not available until after the blast is
mass, this method may also result in a large amount of loaded. The research conducted did, however, provide
vertical heave. Several mines in the western U.S. report the valuable data in regard to characterizing blast movement.
use of row by row firing with very short inter-row delayS. Rock is proven to move towards the instantaneous free face
The role that electronic detonators will play in controlling available to each blasthole as it fires. Also, rock is found to
blast movement for the purpose of grade control is only move more at the mid elevation of a bench than in the toe
now being investigated, however, it is likely that the use of or in the stemming region.
this technology will increase rapidly as benefits are proven. Harris 6 developed a method of measuring blast movement
One interesting approach to blasting for grade control is in which magnetic targets are placed in un-charged blast
reported by Kunze lo at the Pegasus gold mine in Montana, holes and geophysical techniques are used to determine
U.S.A. Ore and waste polygons to be blasted are tied in post-blast target location. This method develops movement
such a way that each area of the blast acts as an independent data prior to loading and in general, data is gathered for all
firing pattern, and the timing of the blast is designed such targets placed in the blast.
that each area will heave to the centroid of the individual At a fundamental level, muckpile surface topography and
polygon. This method creates troughs in the muckpile along surface markers such as embedded poly pipe provide
the polygon boundaries, which are then easily identified by reasonably reliable data at minimal cost and for a minor
the operator during loading. This method as reported used expenditure of effort on the part of the ore control, blast and
detonating cord as the primary initiation system, however, it survey crews. Irrespective of the method used to measure
appears that electronic detonators would be ideally suited to blast movement, assuming point movement data of
allowing this type of complex tie-in with minimal extra sufficient number are gathered, these data can then be used
effort. to re-calculate the ore polygon position in the post blast
Other common aspects of blast designs where dilution muckpile.
due to ore movement is a concern, include the use of short
benches to account for variation in ore grade vertically,
lower powder factors where possible to reduce the energy Algorithm description
available in the blast and square blast patterns allowing The computation canied out in this analysis is to create,
easier reconciliation of sample data. Unfortunately, many of from the movement of targets included in the blast, a vector
these methods are in direct conflict with the principles of field transformation that allows the computation of post-
open pit blasting which attempt to create a well fragmented, blast position of pre-blast points. This displacement vector
loose, muckpile for a minimum cost per tonne broken. operator can then be used to map the grade outlines of an
Several authors report on intensive procedures carried ore polygon to re-establish dig lines for the loading
out at Australian mines in which ore dilution is a operation. Cunent practice in ore control is for the loader
concern 1-12. Mining is carried out using short benches, operator to work from ore-waste boundaries marked on the
typically five metres in height. Each blast hole is sampled surface of the blasted muckpile. Even though research l has
several times, typically once for the upper and once for the shown that the rock in a blast moves differently depending
lower bench. Mining is conducted using a backhoe shovel, on its elevation in the blast, the algorithm for this analysis is
situated on the top of the muckpile and the bench is mined two-dimensional, in keeping with standard loading
in flitches, typically three, two-metre flitches will result procedures in use in open pit mines in which ore polygons
from a five-metre bench after some swelling during the exist only as two-dimensional figures and ore-waste
blast. Ore control polygons are calculated for each flitch boundaries are assumed to be vertical.
based on blast hole samples and as each flitch is mined out, In order to model the displacement vector field, the
ore control is re-marked for the next flitch. Hence, ore movement of targets in the blast is broken into orthogonal
control is marked three times during the loading of the horizontal components (AN and ~). These components are
blast. then modelled separately to assign grid values for a trend

244 APCOM2003
 surface covering the area of interest. The separate values
are then used to compute a resultant displacement vector for
each grid point in the area, which is then stored in a cell of
a block model. The transformation of an ore-waste
boundary is calculated by detelmining for each vertex in the
ore polygon the displacement vector of the cell containing
that point and applying that displacement to the point.
The method of modelling the displacement trend can be
carried out using any of the standard estimating techniques, sequence
such as Triangulation, Polynomial Surface Fit, Inverse
Distance or even Kriging. For the purposes of this analysis,
a simple Inverse Distance Squared (ID2) model was
calculated using the standard Surpac Grid Tools. This
method has the advantages of allowing extrapolation
beyond the convex polygon of the original targets,
honouring the original data points and satisfying the
intuitive notion that the influence of a data point diminishes
as a function of the distance between it and the point to be
estimated. In addition, this model is computationally simple Inter Hole
to implement. Further field study is required to determine Delay,9ms
the 'best' fit method, which is expected to be site specific.
After the vector model has been constructed and the
displacement applied to the ore polygon, an estimate can be ••
made as to the cost of ignoring the ore movement. That is, Inter Row
if the ore block is found to move in the blast and the Delay, 17ms
original dig line positions are not amended, two
classification errors will occur: Ore will be sent to the waste Broken Face
dump and waste will report to the processing stream. In the
former, the cost to the operation is the lost revenue from the
Point ofInitiation
misclassified ore and in the latter, the cost is the
unnecessary cost of processing the misclassified waste.
Surpac allows the rapid calculation of these
misclassifications, by creating 'outer-sections' between the Figure 1. Schematic diagram of blast layout and sequence
original and transformed ore polygons. The area of these
sections is easily determined and hence the volume,
tonnage and contained metal can be calculated.
':I? 1 nn N
Example results
This algorithm was used to analyse a blast at an open pit .. +26
gold mine that occurred in Fall, 2001. Table I summarizes ,,;

the blast parameters.

The blast included approximately 300 holes and was fired
1? n'i n N
using shock tube initiation products on surface with
detonating cord and Pentolite primers in the hole. The blast £+23
was fired to the south into a buffer of broken material, lA'
.-*40
using an approximate VI effective sequence, Figure 1.
A total of 9 magnetic targets were included in uncharged JO, .. +32
holes in the blast and located afterward using magnetic 1? n n n N .CC
J!;
geophysics 7 . The positions of the holes, pre-and post-blast
are shown in Figure 2. (Note that the grid coordinates on all
"""'-;-49

Table I 11 ql)n N
Blast parameters for case study JO" 'ol'l
. +E 8
lA

Parameter

Burden and spacing, m 4.27

Pattern Square t87
l1qnnN
Bench height, m 6.00 .w .w

Sub-drill, m 1.2
""
~ ~ ~
~
Hole diameter, mm 171
Stemming, m 4.25 )j ~ ~
Inter-hole delay, ms 9
Inter-row delay, ms 17
Powder factor, kg/bcm 0.7
Figure 2. Movement of blast targets, showing hole ID#, pre-blast
position (+) and post-blast position (il)

UTILIZATION OF BLAST MOVEMENT MEASUREMENTS IN GRADE CONTROL 245

Figures are in V.S. survey feet, as that is the surveying
standard for the mine.) The targets show a generalized .1 '2MJ>L
movement to the south-west, which is consistent with the
blast firing sequence. These displacement vectors were then
used to compute the displacement vector field for the blast
using an ID2 algOlithm. ':I 11':; n N
(( \

Figure 3 shows the location of the ore polygon for the \\

blast as determined from kriged blast-hole assays. (Ore is
inside the polygon.) This ore-waste boundary is flagged on
/
1 ) 1 nn N 1
the muckpile surface after the blast, based on the pre-blast 't
location of the polygon, and flags are used to indicate \1
digging limits for the loader operator. No attempt is made \\A +26
to account for movement of the polygon due to blasting. '< )n11 n N J
Figure 4 shows the results of applying the computed sf ~23 40
displacement to the Oliginal ore polygon.
Figure 5 shows the results of not adjusting the ore-waste
boundary for the blast movement. The lighter shaded area 3 '1LIlR 1i
~ of 32 L<C:'

A+4~/
to the NE of the map indicates waste that has moved into
the marked ore zone and will be processed unnecessarily. (ill
The darker shaded area to the SW represents ore that has
moved out of the marked ore zone and so will report to the ':I ,QI1 n N '~
~
~'J
waste dump and be lost. pS

Discussion and conclusions

Average resultant movement for the case study was
':I IQ n n N
i-'-' f-'-'\
\ f ~7 F>'-'
l~
f-U i-'-'

approximately 3.0 metres in X- Y space. Table II

summarizes the calculated volume and tonnage of matelial
~
~
I='
~ )
R!
51:
b1
~
to
~
11
--'
~
t;;p:5
') ~<c:n
from the ore polygon that would be misclassified if this
movement is not accounted for when the ore control
Figure 4. Calculated post-blast ore polygon (left) overlaid on
polygon is marked on the post-blast muckpile. pre-blast polygon (right)
Tonnage has been calculated assuming a density of 2.7
tonnes/m3 and a bench height of 6.1 ill. Pre-blast or bank
volumes and density are used. The average gold grade for

1 P? nn N
\

':l ) 1':; n N

\
':I 1 nn N

\ +26
':l 'n t:; n N I
./23
4U

':l 'n n n N
-f; 32 ()<C:

'< . qt:; n N
+49
~
'< Iq nn N
If ~ pS ~ )~

~
~
i-'-'
~)
..LJ i-'-' i-'-'
~
~
~
~ ~
~
to 1 ~~
" ~"n ~

Figure 5. Material miscIassification. Dark =ore to waste,

Figure 3. Pre-blast ore polygon calculated from blasthole samples Light = waste ore

246 APCOM2003
 Table 11 J.J.K. MOUSSET-JONES, P. The impact of blast
Misdassification summary induced movement on grade dilution in Nevada's
precious metal mines. Proceedings of the Fifth
Area, Tonnage, Recoverable Value, Processing International Symposium on Rock Fragmentation by
m2 T Au, g USD Cost, USD Blasting-Fragblast 5, Montreal, Quebec. Mohanty
Original 1550 26,000 64,500 682,000 426,000
(ed). Balkema, Rotterdam. 1996. pp. 407-413.
polygon 5. GILBRIDE, L., TAYLOR, S., DAEMEN, IJ.K. and
Ore to waste 420 7,000 17,500 184,000 N/A MOUSSET-JONES, P. Rock movement induced by
Waste to ore 360 6,000 0 0 N/A bench blasting. Proceedings of the 1996 International
Symposium on Mining Science and Technology. Guo
and Golosinski (eds). Balkema, Rotterdam, 1996.
pp. 667-676.
the polygon was 0.1 oz/ton or approximately 3.1 g/tonne.
6. HARRIS, G.W., MOUSSET-JONES, P., and
Recoverable gold mass assumes a recovery of 80% and the
DAEMEN, J.J.K. Measurement of blast induced rock
gold price used is 300.00 U.S. dollars (USD) /oz or
movement in surface mines by application of
approximately 10.58 USD/g.
magnetic geophysics. Transactions of the Institution
Gold which is effectively lost as it is loaded to the waste
of Mining and Metallurgy (Section A: Mining
dump, is valued at USD 184,000. The effect of the waste-
Industry). Vol. 108, September-December. 1999.
to-ore misc1assification is to increase the overall processing pp. AI72-180.
cost per gram for this blast from 6.61 USD/g to 9.06
USD/g, or a 37% increase in unit processing cost. 7. FIRTH, I.R., MOUSSET-JONES, P., and DAEMEN,
It can be argued that for an orebody with dramatic grade J. Blast movement measurement for grade control.
gradients at the boundary of ore outlines, it may be Proceedings of the 29th Annual Conference on
necessary to smear sample data such that ore value is not Explosives and Blasting Technique. ISEE, Las Vegas,
loaded as waste, even if blast movement does occur. An Nevada. 2002. pp. 55-68.
alternative approach would be to account for ore movement 8. FIRTH, I.R. and MOUSSET-JONES, P. Drilling and
and carefully reconcile short-range planning/ore control blasting practices in Nevada's open pit mines. In
grades with mill head grades and, where possible, realized press, to be published in Mining Engineering. Society
metal values. of Mining, Metallurgy and Exploration, Littleton, Co.
2002.
References 9. KUNZE, P. and SHORT, T. The new D&B
1. SHAW, W.J. Grade control for gold mining at Mt Dimension at Pegasus Gold. Engineering and Mining
Magnet, Reedy, Rothsay and Y ouanmi in the Journal. vol. 187, no. 5, May. 1986. pp. 18-22.
Murchison District, W A. Proceedings of the Aus. 10. DEVLIN, S.P. Grade distribution and grade control
IMM Mine Geologists Conference. Mount Isa, procedures, Great Lady gold deposit, Eastern
Queensland. Aus. !MM. 1990. pp. 83-92. Goldfields, WA. Proceedings of the Aus. IMM Mine
2. ZHANG, S., HURLEY, J., MOUSSET-JONES, P., Geologists Conference. Mount Isa, Queensland. Aus.
and DAEMEN, J. Blast rock movement and its impact !MM. 1990. pp. 93-97.
on ore grade control at the Coeur Rochester Mine. 11. FARRELLY, C.T. The mining geology and grade
Proceedings of the 20th Annual Conference on control practices of the Batman Deposit, Mt Todd
Explosives and Blasting Technique. ISEE, Austin, Project. Proceedings of the Aus. IMM Annual
Texas. 1994. pp. 215-228. Conference. Hallenstein (Ed). Aus. IMM. 1994. pp.
3. BULOW, B., CARATTI, J., HENLEY, K., SPITERI, 103-107.
M., and WOTHERSPOON, S. Ore dilution versus 12. LITTLE, T.N. and VAN ROOYEN, F. The Current
mine to mill optimization at KCGM. Aus. IMM State of the Art of Grade Control Blasting in the
Bulletin. no. 6, November/December, 2001. Eastern Goldfields. Proceedings of the Aus. IMM
pp. 31-39. Explosives in Mining Workshop. Aus. IMM.
4. TAYLOR, S.L., GILBRIDE, L.J., and DAEMEN, Melbourne, VictOlia. 1988. pp. 87-95.

UTILIZATION OF BLAST MOVEMENT MEASUREMENTS IN GRADE CONTROL 247

248 APCOM2003