You are on page 1of 17

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/282862936

Stope Blast Vibration Analysis at Dugald River Underground Mine

Conference Paper · August 2015

CITATION READS

1 818

4 authors, including:

Rhett Hassell Ernesto Villaescusa


Northern Star Resources WA School of Mines - Curtin University Australia
10 PUBLICATIONS 72 CITATIONS 141 PUBLICATIONS 756 CITATIONS

SEE PROFILE SEE PROFILE

John Roger Player


Curtin University, WASM
21 PUBLICATIONS 120 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Early strength of shotcrete View project

Design, Construction and Monitoring of Tunnels at Great Depth View project

All content following this page was uploaded by Ernesto Villaescusa on 15 October 2015.

The user has requested enhancement of the downloaded file.


11th International Symposium on Rock Fragmentation
by Blasting
Paper Number: 192.00

Stope Blast Vibration Analysis at Dugald River


Underground Mine
R Hassell1,E Villaescusa2, R de Vries3 and J Player4

1. Rhett Hassell
Geotechnical Engineer, Dugald River Mine - MMG, PO Box 69 Cloncurry QLD 4841. Email:
Rhett.Hassell@mmg.com

2. Ernesto Villaescusa
Professor and Industry Chair Mining Rock Mechanics, Western Australian School of Mine, Curtin University
of Technology, Locked Bag 30 Kalgoorlie WA 6433. Email:E.Villaescusa@curtin.edu.au

3. Rhett de Vries
Senior Geotechnical Engineer, Dugald River Mine - MMG, PO Box 69 Cloncurry QLD 484.1
Email:Rhett.deVries@mmg.com

4. John Player
Principal Geotechnical Engineer, MineGeo Tech Pty Ltd, PO Box 8054 Kalgoorlie WA 6433. Email:
JohnPlayer@minegeotech.com.au

1
ABSTRACT
Concerns regarding the uncertainty of stope dilution at MMG Dugald River Underground Mine led to a trial
stoping program to acquire full-scale comprehensive geotechnical information and to test the validity of the
geotechnical and mining parameters. Given stope production drilling and blasting is a key parameter that
influences dilution a detailed stope blast vibration project was implemented. The stoping trial consisted of 19
stopes, a combination of both sub level open stoping longitudinal and transverse stopes, of which five were
monitored for blast vibrations.
The blast vibration project utilised 20 tri-axial geophones and 5 tri-axial accelerometers installed in the
hanging wall of five monitored stopes. For each stope firing full blast waveforms were recorded along with
detailed information of the blast design. Analysis of the blasts vibrations used Holmberg-Persson approach,
with potential damage determined by the relative (Holmberg-Persson’s) K values.
The project identified the benefit of parallel, breakthrough slot holes in reducing blast vibration damage. The
results show radial rings toeing or stabbing into the stope wall leading to higher vibrational energy yield and
contributed to increased overbreak. Hanging wall overbreak, by comparison, where parallel, 1m standoff,
breakthrough blastholes combined with lower density explosive charging (0.8EP) and increased delays
limited blast vibration damage. A number of ring burdens from 1.4m to 2.5m were monitored and no
difference in terms of energy yield was identified. Changes to the stope drill and blast design are suggested
to better control the large blast vibrations and recommendations from this project have been implemented
into future mine design and planning.

INTRODUCTION
The Dugald River Zn-Pb-Ag deposit is located 85 km north-west of Cloncurry, Queensland. The deposit was
discovered pre-1880, with first systematic exploration in the 1950s. It is regarded as one of the largest and
highest-grade known undeveloped deposits of zinc, lead and silver in the world with a Mineral Resource of
63 Mt at 12% zinc, 1.8% lead and 31 g/t silver (MMG Limited). The deposit was acquired by MMG Ltd in
2009 with mine development starting in February 2012. Exposure of the geological conditions in
development and additional underground diamond drilling highlighted a more complex and challenging
mining environment than predicted in the feasibility study. A mining method review was completed in mid-
2013 which concluded there was an unacceptable level of uncertainly in regards to stope dilution. It was
recommended that a trial stoping program be completed to address the main uncertainties regarding stope
dilution. Trial stoping commenced in early 2014. Given stope production drilling and blasting is a key
parameter that influences stope dilution (Villaescusa E. , 2014a) a detailed stope blast vibration project was
implemented as part of the overall trial stoping study.

Geological Conditions
The Dugald River deposit occurs within the eastern fold belt of the Mt Isa Inlier and the Dugald Lode is
hosted within a steeply dipping black slate sequence of low metamorphic grade. The deposit is a steep
dipping tabular orebody, extending approximately 2 km along strike and to a depth of approximately 1 km.
The orebody strikes approximately north-south (MGN), dips between 45 to 85° to the west, and varies in true
thickness from 2 to 35 m. The area of greatest ore thickness occurs in the central part of the ore body
between the depths of 300 and 600 m, which coincides with the area of flattest dip.
The hanging wall slates (HWSL) occur in the immediate hanging wall of the Dugald Lode and typically vary
in thickness of between 50 and 100m. In the area where the Dugald Lode dip flattens to <50°, the thickness
can be as little as 10m. Muscovite schist, mafic porphyry and calc-silicates units are to the west of the
hanging wall of the HWSL. The HWSL typically consist of massive, fine grained dark grey slate, spotted in
places; to laminated, fine grained dark grey to light grey slates, often with carbonate veining.
The footwall slates and limestone’s are to the footwall of the Dugald Lode. A narrow zone of footwall slates
(FWSL) grade to limestone (LMST) over a distance of up to 50m. In some areas the footwall slates are
absent and the limestone forms the foot wall contact with the Dugald Lode. The footwall slates are almost
identical in geological and geotechnical parameters to the hanging wall slates.

2
The ore body is hosted by slate and is dominated by sphalerite and pyrrhotite/pyrite with minor galena,
arsenopyrite and chalcopyrite. It ranges from massive sulphide breccias with large angular clasts of slate to
stringer veins hosted in slate.
The area selected for trial stoping at Dugald River is located at shallow depth (<250m below ground surface).
The rock mass consists of a steeply west dipping, bedded rock intersected by large scale, potentially weak,
geological discontinuities. Some of these features are sub-parallel to the ore body and sometimes located
within the ore body or the immediate hanging wall of the designed stopes. The structures show variability
along strike and down dip, even within a single stope geometry (i.e. 20m along strike). The rock mass
outside the geological structures can be rated as fair to good, as indicated by the low frequency of
discontinuities per metre and the high strength of the intact rock. Average UCS of the massive/breccia ore
body is approximately 200 MPa. The strength of the host rock slates is about 150 MPa with both rock
masses having a high modulus.

TRIAL STOPING AND MINING METHODOLOGY


The objective of the trial stoping was to acquire full-scale comprehensive geotechnical information and
learning through a cash flow neutral pre-production program, to test the validity of current geotechnical and
mining parameters. The selection of the trial stoping area, termed Zone 6, was based on the drilling
information at the time; it was seen to be representative of the likely hanging wall rock mass conditions
expected across the orebody in late 2013
Mining of the trial stopes was by Sub-Level Open Stoping (SLOS) with cemented rock fill (CRF). Sub-level
heights were 25m and stope strike lengths varied from 15m to 30m. A total of 19 stopes were taken during
the trial. A long section of the trial area is shown in Figure 1 with the 5 blast monitored stopes outlined. Due
to the variability in the lode width both longitudinal and transverse stopes were designed with a value of 12m
being the critical dimension between the transitions. In addition to the blast vibration project a significant
geotechnical instrumentation program was completed concurrently.

3
Figure 1 Long section of showing trial stopes and white outline of blast monitored stopes.

Stope Production Drilling


Stope production drilling was completed using a Sandvic DL421-15C. A blasthole diameter of 89mm was
considered suitable for the range of stope heights (floor to floor distances of 25m) being designed at Dugald
River. This allows for holes exceeding 30m (if required) to be drilled at a nominal 2.5-2.8m burden
(Villaescusa E. , 2014a). A range of ring burdens from 1.4m to 2.5m was trialled. Spacing to burden ratios of
1.4 to 1.8 was maintained where possible.
Breakthrough holes were surveyed at the collar and toe location to determine drill hole deviation. A collaring
error of less than 200mm from design location and a toe deviation of less than 2% of the bore hole length
was targeted. For the five monitored stopes the average collaring error was 228mm and the average toe
deviation was 2.17%. The level of drill hole accuracy was less than target and initiated an improvement in
operational factors such as survey mark-ups, rig positioning, calibration of the inclinometer, and drilling
speed.
Drilling was undertaken from ore drive development that was located in the hanging wall of the orebody
which was narrower than the ore body. Drilling was designed using radial holes toeing into the stope footwall
(FW) boundary but parallel to the designed hanging wall (HW). For transverse stopes the slot was drilled in
the cross-cut drive but not parallel to the dip of the orebody. Typical drill plan sections are shown in Figure 2.
Parallel HW holes were considered beneficial to reducing blast damage to the stope hanging walls and
limiting dilution. A disadvantage of radial drilling is that drilling and blasting becomes more difficult, as a
uniform explosive distribution may not be achieved.

4
Two types of winze designs were trialled in the blast monitoring stopes, both of 3.0m x 3.0m size. A
conventional long hole winze (LHW) design incorporating four reamers (203mm diameter) and 13 blast holes
and a raisebored (1000mm diameter) assisted winze with 12 blast holes (Figure 3).

Figure 2. Typical longitudinal drill plan (left) and transverse drill plan (right).

Figure 3 Plan views of designs for a long hole winze (A) and a raisebore assisted winze (B).

Explosives
Ammonium Nitrate Emulsion was used as the bulk explosive and was loaded into blastholes using a
hypercharge drive unit. The unit combines the non-explosive emulsion with sensitiser to deliver the water
resistant explosive to the blasthole (Orica, 2011). A nominal emulsion density of 1.09 g/cm3 was used for all
holes other than hanging wall holes where a density of 0.87 g/cm3 was used. The lower density explosive in
the hanging wall was used to aid in limiting blast damage. I-kon electronic detonators (Orica, 2014) were
utilised for all mass blasts due to their greater accuracy, increased delay interval and reduced probability of
cut-offs. Electronic detonators can be logged in delay increments of 1ms from 0 to 15,000ms. Non-electric
detonators utilising the shock tube initiation system were used only for winze firings. These detonators have
considerably larger scatter in their detonation time.

5
Sequence
Sub level open stoping (SLOS) requires sequential blasting of three areas: winzes, slots and rings. Winzes
(3.0m x 3.0m) create the initial void into which the slot (3.0m x width of stope) is then fired into. The
remainder of the stope, termed ring firings is then fired into the slot. A combination of these areas can be
fired in one blast which provides productivity efficiencies, however also increases the risk of misfires and
frozen ground. With the use of electronic detonators the risk can be greatly reduced. The limiting factor to the
size of many of the blasts is the available void. A minimum void ratio of 30% was used. Stope 2 was
sequenced with the slot fully extracted before charging and firing the rings. This created a working at heights
exposure which was eliminated in subsequent stope firing sequences by firing the winze and slot to a cap.
With the use of raisebore assisted winzes the initial winze firing was included in the undercut firing reducing
the number of firings from three to two.

Timing
The fundamental objective of the blast timing is to provide each charge column with as many free faces as
possible to break into. The cut-off slot is the most important geometry in sublevel stoping, as it provides a
free face and void for the remainder of a stope to be blasted into. Slots are critical areas where significant
rock mass damage can occur due to the high concentration of explosive energy utilized to ensure an initial
free face or void. Typically 1500-2000 ms were used to fire the winze, followed immediately by the slot holes
which were fired 100 ms apart. Between the slot and rings a delay of up to 1500ms was used to ensure the
slot material had moved from the void. The rings were fired with an inter-ring timing of 2-3 ms per metre of
spacing and between rings >40 ms per metre of burden. An additional 40 ms of timing was used for the
hanging wall holes to ensure as many free faces as possible.

STOPE BLAST VIBRATION MONITORING


In general, the objectives of a blast vibration monitoring program can focus on one of two areas, namely; a
fundamental analysis identifying the location and cause of any misfired blastholes and also a detailed
vibration analysis assessing blast damage, ultimately aiding stope stability. The approach includes several
stages as follows (Villaescusa E. , 2014b):
1. Design and installation of sensors (geophones and accelerometers) at locations appropriate for
monitoring the different blasting sequences within a stope and meeting the program objectives;
2. Monitor the level of vibration from the various stages of a stope extraction (rise, slot and production
rings);
3. Analysis of the data to determine a relationship between vibration level, distance and explosive
quantity for each of the blasts;
4. Identify a link between vibration level and overbreak or damage around a stope perimeter;
5. Modelling of the blast patterns to identify expected zones of poor fragmentation within the blast or
damage around the stope boundaries.

The key objective for blast vibration monitoring was to assess the extent of blast vibration damage and
provide recommendations on drillhole orientation, ring burdens, stand-off distances, sequencing and timing
to reduce blast damage.

Blast vibration monitoring system design


The monitoring system used at Dugald River consisted an Instantel Minimate Plus unit and an interface box.
Each Instantel device has 8 channels and allows two tri-axial sensor arrays to be used, either two
geophones, two accelerometers, or a combination of accelerometers and geophones. The tri-axial sensor
has one element oriented directly towards the stope (referred to as the radial), one sensor orientated parallel

6
to a stope hanging wall (referred to as the transverse), and a third sensor oriented vertically. The sensor
sondes were constructed specifically for the hole diameter (NQ2 diamond) and orientation.
The research project utilised 24 tri-axial geophones and six tri-axial accelerometers installed over five trial
stopes. Typically, each stope had three vibration monitoring holes drilled from a geotechnical monitoring
hanging wall drive. The centre monitoring hole for each stope has an accelerometer 2-4 metres from the
hanging wall blasthole, a geophone approximately 3 metres further away and a second geophone a further 3
metres. The additional two vibration monitoring holes for each stope have a single geophone in each hole 5
to 7 metres from the blasthole. The coordinates for each sensor location are based upon the distance along
the monitoring hole. A typical monitoring set-up used at Dugald River is shown in Figure 4, where different
monitoring holes; a geophone hole, an observation hole, and an extensometer hole are shown. Data can be
collected from closed walls (excavation not exposed), from exposed walls, adjacent to the stope void or from
a filled wall following stope extraction.

Figure 4 Typical stope hanging wall blast monitoring in section (A) and plan (B).

Accelerometers were positioned closest to the stope boundary because of their capability to record the high
frequency responses associated with very close proximity to blasthole charges. The accelerometer used was
the PCB 356A01, 1000g range with a nominal 5mV/g sensitivity. Unlike the geophone elements,
accelerometers require an excitation voltage which is provided to the unit via the interface box within each of
the units. Accounting for the maximum voltage input permissible with the Instantel units, the sensitivity is
adjusted in the interface boxes to 0.00182V/g. The geophones used were OYO MIcroSeis 101 LT with a
frequency response of 14-2000 Hz and a range of 0-2000 mm/s. The 101 LT is a versatile geophone with a
suitable frequency and amplitude range for blast monitoring (Villaescusa E. , 2014b).

Installation
Correct installation of the geophones/accelerometers was critical to the success of the monitoring program.
The geophones and accelerometers were connected to slotted PVC pipe and installed down 75.7 mm
diameter diamond drill holes. The pipes were marked and connected to ensure the instruments were
installed in the correct orientation. Following final positioning the holes were grouted with 0.35 water cement
ratio and allowed to cure a minimum of 7 days before the first firing.

Data Collection
Prior to each blast four Instantel Minimate Plus units were connected to two geophones/accelerometers
each. All instruments could not be connected for each blast and only those instruments deemed the most

7
relevant were used. To ensure the start of each blast is recorded a separate break wire was run from each
unit to the location of the electric detonator where it was joined to the detonator cord. This ensured that the
units started recording when the electric detonator was initiated, which subsequently initiated the down-hole
detonators. For i-kon blasts which do not use an electric detonator an electronic detonator timed at zero was
substituted to specifically perform the wire break.
A total of 12 out of the 14 blasts were successfully monitored. For each blast the detailed charging and
timing information for each hole was recorded, including the ring number, the hole number, the explosive
column length, the blasthole diameter, the delay time, the explosive density and the design XYZ of collar and
toe locations of the charge column.
For each instrument full waveforms were recorded and peak levels corresponding to each of the given
delays as per the nominal initiation sequence, were determined from the waveforms. This was achieved by
calculating the vector sum from the geophones tri-axial measurements and assigning the peak vector sum to
the relevant delay. Figure 5 shows the vector sums from four sensors (VSG16, VSG17, VSG12 & VSG18)
during the mass blast of Stope 9. Delay times as per charge plan were plotted. Matching the delay to peak
velocity is possible due to the high accuracy of the electronic detonator; detonator error is ±0.01% of the
delay time which is considerably less than the delays between charges. However, during the analysis is was
quickly noted that there was a standard drift with the Instanels of approximately 5 ms which needed to be
accounted for. For winze firings the long delay times used (>200ms) meant scatter with the non-electric
detonators did not affect the peak vibration selection. This methodology allowed accurate identification of the
peak velocity at the geophone for each charge, irrespective of the orientation of the blasthole relative to the
geophone location. The peak velocity for each hole was plotted at the blasthole toe location (Figure 6) which
allowed the identification of which holes were responsible for the high vibration.

Figure 5 Peak vector sum for 4 tri-axial sensors with respect to an individual stope blasting sequence.

8
Figure 6 Plan view of the maximum recorded vibration levels with respect to toe of the holes.

BLAST VIBRATION ANALYSIS


In open stoping, equations to estimate the level of vibration very close to a blasthole (where fracturing is
occurring), must be modified to account for the long, cylindrical shape of the charges with respect to the
monitoring locations. The typical equations used for surface blasting, such as the USBM inverse distance
law, can only be applied when the distance between the blasthole and the monitoring point exceeds three
times the length of the charge length (Villaescusa E. , 2014b). For near field damage, which is the interest at
Dugald River, the most widely used equation for near-field vibration prediction is the Holmberg- Persson
approach (Holmberg & Persson, 1980)
The Holmberg- Persson models blast wave attenuation by defining two rock mass and explosive specific
constants (K and α). The Holmberg-Persson approach makes the following assumptions:
1. A radiating blast wave obeys charge weight scaling laws;
2. The peak particle velocity due to each small element of charge within the blast hole is numerically
additive;
3. For practical purposes, the velocity of detonation (VOD) of the explosive charge is neglected;
4. The effect of free face boundaries is also neglected.

9
The Holmberg-Persson approach was used to determine site specific constants K and α in order to model
peak particle velocity attenuation across a rock mass. This is achieved by analysing near field PPV
measurements for different blasting configuration and explosive types. From the vibration waveforms, the
peak vector sum vibration levels is calculated and, according to the nominal initiation time based on the blast
plans and the recorded time of the vibration pulse, a vibration level is assigned to a given blasthole. The data
is then plotted as vibration level versus the “Holmberg-Persson Term” (a term which accounts for the
explosive loading and blasthole geometry).
The data can be regressed to determine equations that relate vibration level to distance and explosive
loading. The data is fitted using a constant defining the rate of attenuation (α) and the assessment of the
potential damage from the blast is determined by the relative K values. Higher values of K reflect greater
vibrational energy yield and potentially greater zone of damage around a blasthole. A value of α equal to
0.75 was used for all the monitored blasts at Dugald River. A typical analysis result is shown in Figure 7, in
this example of the final mass blast of Stope 9 the K value was calculated at 2200, the largest K value
calculated for the monitoring. Some of the points do not fit this trend and the outlier high vibration points are
discussed later.

Figure 7 Vibration levels versus the Holmberg-Persson Tern for Stope 9, final mass blast from (Villaescusa
E. , 2014b).

After determining the site specific attenuation constants, predictions of the extent of blast damage/pre-
conditioning into stope walls can be made by applying the Holmberg-Persson model and by considering a
site specific critical PPV or damage threshold given by the following relationship (Persson, Holmberg, & Lee,
1994):

𝑻𝑽𝒑
𝑷𝑷𝑽𝒄𝒓𝒊𝒕 = Equation 1
𝑬

where T is the tensile strength of the rock (Pa), E is the elastic Young's Modulus (Pa) and Vp is the
compressional wave velocity (m/s).
Intact rock property testing of the orebody and adjacent slates provides tensile strength ranges from 10.0 to
18.5 MPa and a modulus from 55 to 73 GPa. Assuming a p-wave velocity of 5800 m/s this provides PPVcrit
range from 800 mm/s for massive breccia to 1500 mm/s for unaltered slates. For this analysis values of PPV
exceeding 1000 mm/s are considered critical.

10
Blast Vibration Analysis Results
The underlying assumption is that the assessment of the potential damage from a particular blast is
determined by the relative (Holmberg-Persson’s) K values. Higher values of K reflect greater vibrational
energy yield and potentially greater zone of damage around the blastholes. K values were determined for
each blast. However, given the complexity of the blasts with multiple blast area (winze, slot, rings) all being
fired together the blast vibration data was separated into separate blast areas and K values determined.
Given that the values have been calculated for similar rock mass conditions, stope size and monitoring
location and distances, they can be used as a relative comparison to assess performance. Due to the large
data set and subsequent time restraints the same level of analysis was not completed for each stope. The
comparative results for the Dugald River stopes are shown on Table 1.
The values show that for comparative amounts of explosive charges, most of the blast damage is occurring
at the initial stages of the void generation during the winze/slot creation (and interaction between them). The
process is complicated by the stoping geometry with both parallel and radial drilling. For example, for a
longitudinal stoping, the drilling drives are not open across the full orebody width, thus requiring radial drilling
for the slot instead of the parallel drilling. With hanging wall access drives, radial drilling will necessarily toe
into the stope FW boundaries. To understand the effect of different blasthole orientations further analysis is
required.

Table 1 A comparison of K values at the Dugald River Mine from (Villaescusa E. , 2014b)

Stope Stope Type Firing Type Ring Burdens K Value


(m)
Winze Slot Rings
LHW Winze 1 - 567 - -
LHW Winze 2 - 1200 - -
2 Transverse
LHW Winze 3, full slot - * * -
Rings 2.0 - - 2000
LHW Winze 1 - 497 - -
5 Transverse LHW Winze 2, slot undercut - 300 1000 -
LHW Winze 3, slot, rings 2.0 850 1600 2200
LHW Winze 1 - * * *
7 Longitudinal LHW Winze 2, slot undercut - 1131 650 -
LHW Winze 3, slot, rings 1.4 & 2.4 # # 2200
Raisebore Winze 1, slot - 350-800 2000 -
9 Transverse undercut
Raisebore Winze 2, slot, rings 2.0 & 2.5 # # 2200
Raisebore Winze 1, slot - 1000- 1000 -
10 Longitudinal undercut 2200
Raisebore Winze 2, slot, rings 1.9 & 2.0 350 500 800
- Not part of firing.
* Firing occurred but no vibration data recorded.
# Firing occurred, data recorded, detailed analysis not completed.

Blasthole Orientation
To determine the source of the high vibration the results from ring firings were divided into damage potential
from breakthrough holes (B/T), FW Stabbing and radial toeing into the footwall (FW Toeing), as per Figure 8.
The blast vibration data for ring firings for Stopes 5, 7 and 9 were separated out based on this hole

11
nomenclature and K values determined. The results, shown in Table 2, suggest that most of the damage is
occurring from the FW stabbing and radial toeing into the stope footwall
The large vibrations (for the charge weight) emanating from the footwall boundaries are being monitored
within the instrumented stope hanging walls. That is, the large vibrations are being recorded even with
progressive void being formed ahead of the detonation front at those locations (i.e. centre of stope blasted
first). The increase in blast vibration damage due to confinement of the blast holes in the footwall is
considerable. Cavity Monitoring System (CMS) surveys at those stope locations indicates damage such as
over break. The benefit of parallel, breakthrough holes in reducing blast vibration damage is clear.
An exception to this was Stope 9 where high vibration was observed in the hanging wall holes (Figure 7).
This stope had a large amount of underbreak on both the hanging wall and footwall from the ring firing due to
poor ground conditions with collapsing holes leading to some holes being uncharged. It is presumed the
additional confinement provided by the unfired holes created the high vibration results as shown in Figure 7

Figure 8 Sections showing design and CMS shapes with locations for breakthough holes (B/T), FW stabbing
holes and FW Toeing holes.

Table 2 K values for ring firings by hole location (Villaescusa E. , 2014b).

Stope Firing Number Ring Burdens K Value


(m) All rings B/T FW FW
Stabbing Toeing
5 5.3 2.0 2200 1600 1600 2200
7 7.3 1.4 1000 200 1200 1000
7 7.3 2.4 2600 180 600 2600

12
9 9.1 2.5 2200 2200 1100 2200

Hanging wall blasthole stand-off distance


All stopes were designed with a 1m blasthole stand-off from the stope hanging wall boundary. The aim of the
stand-off is to limit rock breakage beyond the stope design. Reconciliations of hanging wall overbreak
following mining compared favourably with the stand-off distance with overbreak in good ground conditions
limited to <1m.
Hanging wall failure of Stope 2 in poor ground conditions nearly reached the adjacent accelerometer
position, where very high levels of vibration (5000mm/sec) were measured at about 5.5-8.5m distances from
the hanging wall holes. This level of vibration probably represents the limit of breakage for this rock mass
and stoping geometry as indicated by the CMS survey. The results agree well with the notion that PPV breakage
is about 4 times PPVcritical (Villaescusa E. , 2014b). Stope 7 accelerometer was located about 3m away from
the hanging wall hole. Back analysis of overbreak showed similar levels of PPVcritical and PPVbreakage. Based
on the accelerometer results a stand-off distance of at least a meter is required for the hanging wall holes in
order to minimise breakage beyond the stope boundary.

Ring burdens
Ring burdens ranging from 1.4m to 2.5m were monitored. Observations on the quality of the ore walls
following firing generally showed relatively smooth, clean walls that have broken back to the ring location
(Figure 9). Some cratering occurs at the collar. The monitoring data indicates a similar level of vibration over
several mass blasts (Table 2). The 1.4 ring burden displayed the lowest vibration but its design was the
result of wall overbreak in the adjacent stope and is not considered an economically viable option. The
analysis could not distinguish a significant difference from 2.0m to 2.5m ring spacing. Therefore, a 2.5m
burden is recommended for the 89mm holes, the drilling practices and the rock mass strength at Dugald
River.

Figure 9 View towards Stope 2 hanging wall showing the condition of the ore walls after slot firing.

13
Firing Sequence
The relatively tight mass blasting sequence i.e. only two firings per stope, used for some stopes did not
seem to increase damage to the rockmass compared to those stopes with a greater number of firings. Stope
2 was extracted with four firings and by firing out the slot completely before the ring firings. This compares to
Stopes 5 & 7 with three firings and Stopes 9 &10 with two firings. The later stopes had an 8-10m cap over
the slot and which was fired with the rings during the final mass firing. The results did not show a discernable
difference between the two methods. Later stope firings during the trial that were not blast monitored
successfully fired out entire lifts (winze, slot and rings) in the one firing. It is generally considered that a more
sequential enlargement of the winze and slot reduces blast damage prior to the mass blasting. More data for
increased number of firings would have been preferred, but given the current information tight mass blasting
sequences perform well and with considerable productivity benefits.

Winzes
Two types of winzes were used; a conventional long hole winze (LHW) and a raisebore assisted winze
(Figure 3). Both winzes were 3.0m x 3.0m with the LHW containing 13 blast holes and 4 x 203mm diameter
reamers and the raise bore assisted winze containing 12 blast holes and a 1000mm diameter raisebore. K
values for LHW ranged from 300-1200 (Stopes 2, 5 & 7, Table 1) and raisebore assisted from 350-2200
(Stopes 9 & 10, Table 1). Both winze methods had comparable blast vibrations with the exception of the
winze and slot undercut firing of Stope 10.
The very high vibration levels for the raisebore assisted winze during the first firing of Stope 10 was for the
first eight holes near the raisebored relief hole. The results suggest that the use of a raisebore does not
necessarily lower the stope blast damage and that more holes may be required near the raisebore hole to
break its circularity very early on the detonation sequence. However, the raisebore assisted winzes tended to
be fired over a greater length and alternatively, shorter raise heights ahead of the slot creation could be
blasted to reduce the confinement.

The effect of backfill in limiting further damage in adjacent stopes


The effectiveness of the fill mass was investigated by monitoring Accelerometer A3 in the hanging wall of
Stope 2. Figure 10 shows the response of A3 with the three blast events on Stope 5 with K values from 549
to 850. This compares to closed wall (before extraction) K values of 550 which were recorded for other
stopes. The accelerometer data suggest that the cemented rockfill support is effectively stabilizing the filled
voids. Hence the data from filled stopes is similar to the data obtained when the stope wall is not yet open

Figure 10 Vibration levels versus the Holmberg-Persson Term for mass blast, Stope 5 as monitored from
accelerometers A3 located in adjacent filled stope Stope 2 from (Villaescusa E. , 2014b).

14
PRACTICAL APPLICATIONS OF THE BLAST MONITORING RESULTS
The key outcome of the blast monitoring trial was to confirm the benefit of parallel, breakthrough holes in
reducing blast vibration damage. Confining of blast holes by drilling of radial rings toeing or stabbing into the
rock mass leads to higher vibrational energy yield which contributes to increased overbreak. Comparison of
overbreak volumes in Table 3 provides additional confirmation. The footwall had a greater amount of
overbreak volume than the hanging wall. This compares well to the overbreak volume of all 19 stopes with
4882m3 in the hanging wall and 5655m 3 in the footwall. Typical sections showing overbreak & underbreak
have been shown previously in Figure 8.

Table 3 Overbreak volumes in the hanging wall and footwall of the stopes that were blast monitored.

Stope Hanging wall Overbreak Footwall Overbreak Volume


Volume (m3) (m3)
2 364.5 243.5
5 24.5 293.2
7 420.8 98.8
9 203.3 198.3
10 30.7 343.6
Total 1043.8 1177.4

Both the hanging wall and footwall exposed similar ground conditions with an average hanging wall Q’ of
10.1 and an average footwall Q’ of 9.3. Both Q’ values are for the first 5m of the hanging wall or footwall. The
footwall is also kinematically more stable with failure by sliding as opposed to slabbing and gravity fall. The
hanging wall did have the benefit of cablebolting from the ore drive which controlled overbreak. However, the
evidence strongly suggests that the high vibrations created in the footwall of the stopes are a major control
on overbreak.
Hanging wall overbreak, by comparison, where parallel, 1m standoff, breakthrough blastholes combined with
lower density explosive charging (0.8EP) and increased delays limited blast vibration damage. This was
shown in the blast vibration results and back analysing stope overbreak where non-structural related
overbreak was limited to <1m. This system should be incorporated into future stoping at Dugald River.
To reduce the level of footwall blast damage a solution is to drill parallel footwall holes particularly within the
slot. However, there are some practical considerations to these recommendations. For transverse stopes
with a cross cut drive the winze and slot can be drilled parallel without any changes to the mine design.
Where the orebody dip is less than 70° there will need to be some consideration on the rilling of the fired ore
and potential confinement of the footwall holes.
To achieve parallel drilling with the rings additional footwall ore drive development is required. This adds a
significant additional cost for each stope with negative mine scheduling implications. However, a large
enough orebody width is required to achieve this which is not common for Dugald River.
For longitudinal stopes the ore drill drive will be required to be opened to the full orebody width. Current ore
drives are 5.5m wide which allow for parallel drilling of the footwall and hanging wall for orebody widths less
than 3.5m. Increased development and ground support costs, variable orebody widths and decreased
stability from larger excavations all need to be considered. Beyond a 7m orebody width this will not be
practical and a cross-cut drive is required to ensure parallel slot holes as a minimum.
Two types of winze were trialled, conventional long hole winze and raisebore assisted winzes. Results show
the introduction of the raisebore does not directly result in an automatic reduction on blast damage. With
winzes not parallel to the hangingwall (and particularly when vertical) resulting in high vibration due to radial

15
drilling within the slot. Preferences for either winze type should be based on scheduling and operational
constraints to achieve parallel holes within the slot.
The results showed that larger ring burdens and reduced number but larger stope firing sequences do not
appear to increase the overall damage to the rock mass when compared to smaller ring burdens and an
additional staged firing sequence. Both outcomes improve the productivity of the stope cycle and confirm
their low influence on blast damage.

CONCLUSIONS
A highly successful blast vibration monitoring program was completed at Dugald River which greatly
enhanced the understanding of the main controls on blast damage. The effectiveness of the program relied
on clear objectives, strong management support, good instrument design, a high level of oversight during
instrument installation and data collection, with detailed analysis techniques to understand what the data is
saying. Such programs are difficult to implement at an operating underground mine but the benefit of actual
data over opinions and rules and thumb cannot be overstated.
The main parameters that cause high blast vibrations leading to stope wall damage were identified with the
advantage of parallel breakthrough holes being clearly proved. Changes to the stope drill and blast design
are suggested to better control blast vibrations. These controls have impacts beyond simple ring design and
timing and a cost benefit analysis of the options was not part of the scope of the project but would greatly aid
in determining the balance between cost, productivity and managing overbreak from blast vibrations.

ACKNOWLEDGEMENTS
The authors would like to acknowledge the help of Tom Williams, Jack Carswell and Andrew Rajapakse
Geotechnical Engineers at Dugald River, John Heilig from Heilig and Partners and Italo Onederra from
JKMRC for aiding in the completion of the project. The authors would also like to thank MMG for permission
to publish this paper.

REFERENCES
Holmberg, R., & Persson, P. A. (1980). Design of tunnel perimeter blast hole patterns to
prevent rock damage. Transactions of the Institute of Mining and Metallurgy,
89:A37-A40.
MMG Limited. (n.d.). Duglad River Fact Sheet. Retrieved March 2015, from
http://www.mmg.com/en/Our-Operations/Development-projects/Dugald-River.aspx
Orica. (2011, October 6th). Subtek Eclipse. Retrieved March 2015, from
http://www.oricaminingservices.com/au/en/page/products_and_services/bulk_syste
ms/bulk_systems
Orica. (2014, April). i-kon II Detonator. Retrieved March 2015, from Orica Mining Services:
http://www.oricaminingservices.com/au/en/page/products_and_services/electronic_
blasting_systems/i-kon_ii/i-kon_ii_-_the_next_generation
Persson, P., Holmberg, R., & Lee, J. (1994). Rock blasting and explosives engineering.
CRC Press.
Villaescusa, E. (2014a). Geotechnical Design for Sublevel Open Stoping. CRC Press.
Villaescusa, E. (2014b). Blast Monitoring of Sublevel Open Stoping Produciton Blasting,
Research Project for MMG Dugald River. Western Australian School of Mines,
Curtin University of New Technology.

16

View publication stats

You might also like