Interaction between the block cave and J

the pit slopes at Palabora mine o

u
r
by A. Moss*, S. Diachenko, and P. Townsend n
a
l

P
Development of the crinkle undercut a
Taking secondary breaking to new levels p
Highest lift heights yet attempted in a
Synopsis e
block cave
One of the fastest production build-ups
r
The Palabora Mining Company has successfully extended the life of
its operation by developing a 30 000 tpd block cave underground among block cave operations
mine below the now closed open pit. The process of transitioning Ramp scavenging in the pit concurrent
from surface to underground mining has been technically, with cave development.
operationally and managerially challenging. This has been due to a
number of factors, in particular, the development of a block cave in Equally important were the lessons
a very competent rock mass. The competence of the rock mass is learned:
clearly demonstrated in the open pit that preceded the block cave, The need to enhance geotechnical
one of the steepest deep open pit excavations in the world. With
knowledge, including data collection,
cave breakthrough, however, the pit was instantaneously
deepened by 400 m causing a major failure of the north-west wall ground characterization and analysis of
involving close to 100 Mt of waste material. rock mass behaviour
The limitations of the various tools
currently available to reliably predict
cave performance
The critical importance of quality and
Introduction performance monitoring
The impact of a large pit on the water
The Palabora Mining Company has management strategy
successfully extended the life of its operation The interaction between the cave and the
by developing a 30 000 tpd block cave overlying pit.
operation below the now closed open pit. The The last lesson was brought into focus
process of transitioning from surface to when, shortly after cave breakthrough,
underground mining has been technically, cracking was observed some 250 m behind the
operationally and managerially challenging. north-west wall of the pit. With time, what
This has been due to a number of factors, in initially appeared to be two separate wall
particular the development of a block cave in a failures coalesced into a single failure
very competent rock mass with limited involving some 100 Mt of material over the
experience within the industry on which to full wall height. The failure has highlighted a
base design. In effect Palabora was deficiency in our understanding of the effect
pioneering a difficult position in any that caving has on pit wall stability and,
operation. However, with planned production equally important, the impact of the failure on
rates in excess of 30 000 tpd; Palabora has cave operations.
successfully transitioned from an open pit to
an underground operation and is considered to
be at the cutting edge for this type of
development.
There were a number of important * Rio Tinto Technical Services.
achievements during the transition process. Palabora Mining Company.
These include: The South African Institute of Mining and
Safety performance during construction Metallurgy, 2006. SA ISSN 0038223X/3.00 +
0.00. This paper was first published at the SAIMM
and operations comparable with open
International Symposium, Stability of rock slopes
pits in open pit mining and civil engineering situations,
Caving of a competent rock mass. 36 April 2006.

The Journal of The South African Institute of Mining and Metallurgy VOLUME 106 NON-REFEREED PAPER JULY 2006 479
Interaction between the block cave and the pit slopes at Palabora mine
Overview wide. Drawpoints are on an off-set herringbone style. The
undercut level is located 18 m above the production level.
Project description and history LHDs dump directly into four crusher stations located along
The Palabora Mining Company (PMC) was founded in 1956 the northern extremity of the production footprint. Access to
and open pit mining operations commenced in 1966 at rate of the orebody is via a 10 m diameter service shaft. Ore is
30 000 tpd increasing to 82 000 tpd prior to closure in 2002. hoisted through a 7.5 m production shaft equipped with four
In total some 960 Mt of ore and 1 300 Mt of waste were 32 t skips.
mined. Throughout pit operations there was a history of Palabora has the highest lifts yet attempted (lift is the
innovation and optimization, for example the use of trolley vertical distance from the production level to the break
assist for 240 t haul trucks and the introduction of in-pit through point on surface or in the open pit above). This
crushing. The pit was approximately 800 m deep with inter- points the way for future block caves. High lifts allow a
ramp slope angles ranging from 37 in the upper weathered substantial reduction in capital requirements per vertical ton
lithologies to about 58 in the competent constrained ground of ore mined.
toward the base of the pit (Figure 1). Key statistics on the underground are given in Table I. A
The orebody is an elliptical shaped vertically dipping detailed description of the mine layout is given in Calder et
volcanic pipe measuring 1 400 m and 800 m in plan with al., 2000, while additional data on cave progression are given
resources to 1 800 m below surface. Transgressive and in Moss et al, 2004.
banded carbonatites form the central core of the orebody with In 2004 the underground mine had achieved one of the
the banded carbonatites and transgressive carbonatites best safety records for an underground operation anywhere
dominant in the western sector and eastern sectors of the in the world. Today with construction complete, the LTIFR
orebody, respectively (Figure 1). Barren dolerite dykes with a rate is 0.50, competing well not only with underground
steeply dipping north-east trend are present as are a number mines but also with surface operations. This is a notable
of north-west and north-east trending faults. achievement for both Palabora and Rio Tinto.
The development of a 30 000 tpd block cave operation
Cave development
was approved in 1996. Shaft construction started in 1997.
Target production was achieved in May 2005. The Palabora orebody is the strongest rock mass in which
The production level, Figure 2, is located approximately block cave mining has been attempted and there was consid-
1 200 m below the surface and approximately 400 m below erable scepticism within the industry as to whether it would
the final pit bottom while the production footprint, consisting cave effectively. The Palabora orebody is caving and
of 20 cross-cuts is 650 m long and approximately 250 m caving well.

PIT BOUNDARY AT BASE

OF WEATHERED ROCK

37 ULTIMATE PIT
CREST

48 49
49
60 50
51 HIGH WATER TABLE AREA 51

37 55
53

58

CARBONATITE 53
55 37
GLIMMERITE

FOSKORITE
49 50
40 45
MICACEOUS PYROXENITE

53 53
FENITE 53

0 200m 49

SCALE
37

Figure 1General geology and pit slope geometry

480 JULY 2006 VOLUME 106 NON-REFEREED PAPER The Journal of The South African Institute of Mining and Metallurgy
Interaction between the block cave and the pit slopes at Palabora mine
J
o
u
r
n
a
l

P
a
p
e
r

Figure 2Plan of production level showing dykes and faults

Table I
Summary statistics
Rock mass quality (Q) 1015 Hydraulic radius at which caving initiated 45 m

Lift heights 400600 m Estimated swell factor 20%

Number of production cross-cuts 20 Number of drawbells 166
Drawbell dimensions 34 by 17m Average draw rate (May 2005) 116 mm

500

400
Metres

300

200

100

0
Palabora

Parkes

Henderson

EI
Teniente

El Salv

Bingham
(plan)

Figure 3Comparison of lift height

Undercutting (the process used to horizontally slot the substantial uncertainty with this prediction but that the
orebody and initiate caving) started in 1999 and was highly dimensions of the overall footprint were sufficient to cater for
successful. The undercut was advanced ahead of construction this uncertainty. This was borne out by experience as,
of the production level to providing a stress sheltering for although the dimensions required to initiate caving were
the excavation of the production level. Palabora was the first about 30% greater than that predicted, the hyrdaulic at which
block cave to use the crinkle cut undercut design. The caving occurred was well within the dimensions of the
combination of geometries proved to be an extremely production footprint. Caving initiated in April 2002 when a
efficient undercutting method taking 52 months to undercut hydraulic radius of 45 was reached.
119 000 m2 at rates reaching some 4 500 m2 per month. A number of monitoring systems including TDRs, open
During feasibility it was predicted that caving would be holes and a mine wide micro-seismic system were installed to
initiated when the undercut area reached 140 m by 140 m. It track cave propagation. Of these, only the micro-seismic
was recognized that due to lack of knowledge there was system provided a broad view of cave progression (see Glazer

The Journal of The South African Institute of Mining and Metallurgy VOLUME 106 NON-REFEREED PAPER JULY 2006 481
Interaction between the block cave and the pit slopes at Palabora mine
and Hepworth, 2004). By locating events, it was possible to be safely carried out) was set at 200 m. This distance was
determine when caving was initiated, when the pillar based on an assessment of the stability of the web of ground
between the cave back and pit floor failed and to track stress between cave back and pit floor. Following on from the initial
transfer that occurred as a result of the caving process. The investigations, day-to-day geotechnical management was
micro-seismic system has also provided important infor- maintained throughout the project, comprising:
mation on pit wall behaviour. Monitoring by survey prisms, inclinometers and
As expected, the strong competent rock mass initially
geophones
resulted in coarse fragmentation with secondary breaking
Regular inspection of pit walls
being a major challenge. No cave operation has undertaken
Direction of remedial shotcreting
the amount of secondary breaking that has been required at
Approval of all blast designs
Palabora, where some 50% of the initial tonnage has had to
Inspection and geotechnical clearance after each blast.
be blasted to clear drawpoint hang-ups and blockages.
Although coarse fragmentation was anticipated, the ore has Water management
actually broken finer than anticipated; the process of safe The open pit provides a large catchment area (240 Ha),
and efficient clearing the hang-ups has required substantial funnelling surface run-off down into the underground mine
organizational effort. In doing so, however, the mine has via the cave. The average annual rainfall is 534 mm with the
achieved some notable technical successes:
majority of the rains falling during the months of December
The development of secondary breaking equipment through January as major storm events.
from prototype through to production units, including At feasibility, it was considered that storm run-off would
the considerable learning process that typically occurs be substantially attenuated within the cave, releasing flood-
with new equipment water at a rate that could be managed by the underground
Development of effective industrially engineered pumping system. The mine pumping system was designed
secondary breaking processes that fully utilize accordingly. However, experience with actual storm events
available resources. has shown that storm water reports underground rapidly (in
The cave is now fining with increasing height of draw. less than 12 hours) indicating that the cave material has a
This will result in increased draw rates and production hydraulic conductivity of around 10-2 m/sec, several orders
performance.
of magnitude greater than originally estimated. Thus, it was
realized that a significant deficit existed in the water-
Transition issues handling capacity of the underground mine. There would be
A number of issues were encountered during the transition times when the underground pumping system and water
from an open pit to an underground operation. These storage facilities as designed would be unable to cope with
included a culture change from surface to underground the stormwater inflows and sections of the mine would flood.
mining, down-sizing of the mine work force in recognition of In order to predict the probability of flood events, a series
reduced throughput and the availability of appropriate skills. of simulations was carried out using the history of storm
Technically, the three most important transition issues were events in the region. Stochastically generated rainfall data
the production shortfall that occurred with pit closure, water were used to simulate rainfall as a variable flux on the model
management within the open pit and cave pit interaction. surface. The results of the rainfall analysis show that the
worst case event of a 1:100 year storm in 24 hours has a 2%
Ramp mining probability of occurring within the next 18-months and a
PMC realized that underground production would not be 10% probability of occurring within the next ten years. This
sufficient to sustain milling and smelting operations once pit level of storm would deliver 644 000 m3 of water into the
operations ceased. The decision was made to augment open pit and would cause substantial flooding of the
production with ramp mining operations. This involved underground mine. Different stages of mine life were
mining the pit ramp in a series of small benches (5 to 7 m examined together with options for water management,
high) effectively increasing ramp gradient from 10% to 15% including:
(for more details see Whitham et al. 2004). Over a 20-month Increasing interception of rainfall in the pit.Creating
period some 10.4 Mt of ore was mined in this manner at underground water storage
production rates averaging 16 000 tpd. Minimizing recovery times from a major flood by the
From the outset, detailed geotechnical input was a key purchase of key capital spares to facilitate recovery
aspect in the design and successful implementation of ramp Establishing an insurance stock of concentrate.
mining. One of the key factors controlling the mining The analysis resulted in the following water management
sequence was determining where in the pit scavenging could strategy:
safely be undertaken, given that the cave was propagating Installation of pit sumps of sufficient capacity
upwards toward the pit floor. The size of the crown pillar (325 000 m3) to contain 60% of the projected inflow
buffer zone (the vertical distance measured from the top of Installation of water tight doors on the return air way
the cave to the lowest elevation in the pit where mining could system to provide 70 000 m3 of emergency storage.

482 JULY 2006 VOLUME 106 NON-REFEREED PAPER The Journal of The South African Institute of Mining and Metallurgy
Interaction between the block cave and the pit slopes at Palabora mine
J
This strategy was followed. However, the failure of the There was concern for other major facilities including the o
north-west wall changed the geometry of water flow into the ventilation shaft, located on the East Wall within the pit and u
underground workings. Although the failure removed one of the production and service shafts located some 90 m from the
the most productive pit sumps, it is considered that the rim of the East Wall. Thus, an extensive monitoring
r
failure itself provides very substantial storage. To date, the programme consisting of a network of GPS stations, n
mine has been able to comfortably handle all storm events. tiltmeters and crack meters was initiated, augmenting the a
Cave-pit interaction original pit system and the cave micro-seismic system. This l
was followed by a comprehensive programme of numerical
There has been interaction between the pit and the cave.
Initially, the impact of the pit on the cave was to concentrate modelling of cave pit interaction (see Brummer et al., 2006
stress within the web of ground between pit floor and cave for a detailed discussion on the failure and the modelling
process). The modelling results, once calibrated against the
P
back. The increased stress level aided cave propagation by
monitoring, indicated that the failure was associated with a
causing existing fractures to extend and creating new
fractures. More importantly, this process has affected primary pervasive jointing and the very significant stress change that p
fragmentation and substantially reduced the secondary occurred during cave breakthrough. e
breaking requirements. From the surface monitoring results three distinct r
Conversely, the cave has affected pit wall stability. movement zones were identified. As shown in Figure 5, there
Movement of all pit walls increased substantially upon cave appears to be a strong relationship between these zones and
breakthrough into the bottom of the pit. The greatest amount seismicity. Not surprisingly, a direct relationship between
occurred in the North Wall where cumulative movements of movement of the failure and the amount of material drawn
in excess of 1.5 m were measured. The first indication of a from the cave was observed.
major problem, however, was a bench failure adjacent to one The angle from undercut level to the surface limit of
of the pit sumps. This was followed by the discovery of large cracking is about 55. This shallow angle was not anticipated
cracks some 250 m back from the pit rim (note: it is not in the strong competent rock mass forming the walls of the
known if the cracking occurred before or after the initial pit. Analyses of cave pit interaction were carried out as part
bench scale failures as only once the failure occurred was a of the feasibility study. The results of these analyses were
survey made of the dense bush that surrounds that potion of
reviewed both internally and externally. The conclusion
the pit).
reached was that the developing cave would induce failures
The failure grew in size until after a period of about 18
of sections of the pit wall, but that these would be contained
months it encompassed a major section of the North Wall
within the footprint of the pit. In part, this underlines the
with the crest some 50 m back from the pit rim and the toe
limited knowledge of caving-induced subsidence, particularly
somewhere near the original pit floor. The failure dimensions
were some 800 m high by 300 m along the wall, see in combination with a deep steep pit traversed by major
Figure 4. The depth of the failure can only be surmized, but structures.
the rim is thought to be reasonably shallow, about 50 m The failure also has the potential to affect underground
thick. production. The majority of the failed rock is significantly
The failure affected several segments of the mines below cut-off grade. As the cave is pulled, this waste material
infrastructure including access and haul roads, tailings, water may move at a faster rate than the ore due to differences in
and power lines, water reservoirs and a railway line. size between caved ore and the failed waste. This could result
Fortunately, it developed at a relatively slow rate allowing the in premature dilution ingress and shortening of the mine life.
various facilities to be moved prior to the ground on which Work is underway to study this phenomenon with the aid of
they were located become enveloped in the failure. both physical and numerical modelling.

Total Movement October Week 4 1.25

mm/day 1.2
1.15
1.1
1.05
1
0.95
0.9
0.85
0.8
0.75
0.7
0.65
Ground stressed 0.6
Approximate 0.55
0.5
significant 0.45
instability 0.4
0.35
0.3
Ground relaxed as stress Mica Fault 0.25
transferred to west of 0.2
Mica Fault 0.15
0.1
0.05
0

Figure 4Zone of slope instability

The Journal of The South African Institute of Mining and Metallurgy VOLUME 106 NON-REFEREED PAPER JULY 2006 483
Interaction between the block cave and the pit slopes at Palabora mine
12000 12500 13000 13500 14000 14500
-23000
This region is compressing
horizontally E-W, or
caving into void below

0.04
This region is 3
moving inward, 1.9
This region is on vector 2.12
moving inward, on Azimuth 210 1.59

vector Azimuth 130 1.003.0

3.05.0
5.08.0
> 8.0
Size Magnitude
-24000 0.0
1.0
2.0

Centre
-24500 24018
-13011
-804

Azimuth=0

-25000

Figure 5Movement zones and seismic activity

Conclusions Palabora Underground Mine Project and the role of the

Palabora Mine has transitioned from an open pit to an authors has simply been to report this work to a wider
underground project to a block caving operation with audience.
production having reached the targeted rate of 30 000 tpd. A The permission of Palabora Mining Company and Rio
number of firsts have been achieved during this transition Tinto Technical Services to publish this paper is gratefully
process: acknowledged.
Successful caving beneath one of the deepest and
steepest pits in the world
Successful ramp scavenging above an active block cave
References
Successful development of a high lift cave
Development of methods to effectively secondary break BRUMMER, R., LI, H., and MOSS, A. The Transition from Open Pit to
in excess of 50% of drawn tonnage. Underground Mining: an Unusual Slope Failure Mechanism at Palabora.
Several major lessons have been learned: Proceedings Int. Symposium on Stability of Rock Slopes. Cape Town, 2006
Block caving can be a viable method of prolonging
mine life
CALDER, K., TOWNSEND, P., and RUSSELL, F. The Palabora Underground Mine
However, cave pit interaction, in particular the complex
Project. Proceedings MassMin 2000. G Chitombo (ed.)Brisbane 2000.
interaction between induced stresses and structure that
pp. 219225.
occurs during cave progression and breakthrough, is
poorly understood, resulting in poor prediction of rock
mass performance. Given the level of up-front GLAZER, S. and HEPWORTH, N. Seismic Monitoring of Block Cave Crown Pillar.

investment in a block cave, it is extremely important to Proceedings MassMin 2004. Santiago.

develop reliable predictive tools

The permeability of coarse caved ground is high and MOSS, A. RUSSELL, F., and JONES, C. Caving and Fragmentation at Palabora:
must be catered for by appropriate water management. Prediction to Production. Proceedings MassMin 2004. Santiago.

Acknowledgements WHITHAM, M., TOWNSEND, P.A., DENNIS, B., and MOSSOP, D. Geotechnical Know-
The paper is the result of all those who participated in the How Aids Ramp Scavenging at Palabora. E&MJ July 2004.

484 JULY 2006 VOLUME 106 NON-REFEREED PAPER The Journal of The South African Institute of Mining and Metallurgy