Sub-Level Caving: Where Is It Headed?

Sub-Level Caving: Where Is It Headed?

This paper describes the challenges of mining a sill pillar under various types of paste back- ll quality. From 1994
to 2002, Louvicourt mine produced approximately 12,600,000 t grading 3.5% Cu, 1.5% Zn, 27 g/mt Ag, and 0.85
g/mt Au. The mining method is transverse blasthole stoping, mining primary and secondary stopes. The production
of this 4,300 t/d operation was accomplished using two mining horizons; from the 655 m level to the 415 m level and
from the 860 m level to the 680 m level. The main sill pillar is between the 680 m level and the 655 m level. The
second sill pillar is between the 885 m level and the 860 m level representing a quarter of the main sill pillar in size.
Since 2002, production has gradually decreased and mining activities are expected to end in mid-2005. The sill pillar
production will represent almost 25% of the total mine production. The overall recovery of the sill pillars stopes will
be discussed in this paper. To recover the sill pillars, more than 1,500 m of development was planned in or under
paste backll. The rst sill pillar stopes under paste backll were mined out in 2002 with excellent results

INTRODUCTION
Much of the theory on which SLC is based was developed in Scandinavia many years ago. It was mostly based on
bin theory and ellipsoids of movement. SLC has since fallen into disfavour.
Recently a number of Australian mines have adopted SLC as a primary method and have achieved some very good
results which has called into question the original bin theory assumptions.
It is clear that the process of choke blasting (compaction of waste, very variable fragmentation through the ring)
significantly conflicts with the assumptions of regular loose flow of material in a bin. The good results can only be
explained by using different models of behaviour, which have been borrowed from block caving. If these models are
correct then they point to a number of design and operational changes that can be applied to improve the
effectiveness of SLC.
This paper describes a new model for SLC and how this model can be used in terms of design and operation.

SUB-LEVEL CAVING
Sub-level caving is considered a low cost method by the ton moved (often referred to as a factory method) but
higher cost due to the perceived high dilution (unfavourable costs to final finished metal). but there are existing
operations that have remained faithful to slc (even after trying alternatives) and there are new operations coming on
stream using the method.
It is important to recognise both the attributes and the drawbacks of the method. It might have limited application but
it is a valid method worthy of consideration in the right context: usually strong, competent and massive deposits. The
knowledge base for SLC is comparatively small as very few mines use the method. Most of the "theory" comes out
of Scandinavian iron ore mines and is many years old and mostly based on model studies and classical bin theory.
Data on draw behaviour (what is happening in the broken mass) is difficult to come by. The other major problem
with iron ore operations is that dilution can carry very high grades so it is difficult to assess the "true" dilution
(tonnes drawn from outside the current ring). The clear advantage of SLC is that it is a very predictable "factory"
type method with high production potential, reasonable costs, "top-down" (low up-front capital, low stress) and very
little ore is at risk at any one time (a few thousand tonnes in an individual ring).
SLC suffers from apparent significant contradictions: a slice of broken material being drawn relatively clean whilst
surrounded by broken waste, figure 1.

Sub-Level Caving: Where Is It Headed?

Figure 1 - Illustration of SLC method

It seems to defy logic. For this reason it is very important that a practical, common sense model of how, or why, it
works is presented.
When a ring is blasted (typically 2,000 t) the first part of the draw is reasonably clean; then waste above and in front
of the ring starts to come into the draw point and a mixture of ore and waste is drawn; the proportion of waste
increases until cut-off is reached. When the draw point is closed there will be ore left behind. This mixes with the
previous ore/waste in the cave; the waste therefore increases in grade as the SLC matures. The objective is to keep
the waste out for as long as possible but to try and make the most of the zones of higher grade dilution.

HOW (WHY!) DOES SLC WORK?

The classical theory is based on ellipsoids of motion and isolated draw from a single draw point developed from
flow of loose material in bins (ref.1,2,3). There are concerns with this model as it tends to ignore differences in
fragmentation, material types, the significant weight of the cave which compacts the cave material, and the action of
the blast which also consolidates the waste in front of the blasted ring. All of these factors will significantly affect
the way in which choked material moves. Conventional draw analysis suggests that dilution enters the draw at very
low draws (below 40%) but recent experience suggests that it can enter at draws well above 50%.
Recent operations at an Australian gold mine appeared to be breaking some of the classic rules and included some
unusual operational changes but they were achieving much better results than suggested by classic theory. When we
analysed what we thought was happening we concluded that the critical considerations could be broken into design
and operation (layouts and operating practices) and physical (orebody characteristics):

Sub-Level Caving: Where Is It Headed?

Operations
differential fragmentation: finer fragmented material flows more readily than coarse material; dilution should be
coarser than the ore; (fine material can easily flow "through" coarse material)
compaction: compacted material does not flow as readily as freshly blasted material; the blast must compact the
"waste" in front of the fresh ore so that it does not flow as readily as the freshly blasted ore
temporary arching: coarser material at the top of the ring impedes flow of waste from above; coarser material
draws over a much wider "arch" than fine material; coarse material can temporarily "hang-up" whilst the finer
material below the arch is drawn (but very coarse or unblasted material will allow waste to flow around the ore
and reduce recovery a delicate balancing act) draw coverage: the more the ore is undercut by development the
more likely it is that the ore will flow into the draw point
interactive draw: draw points drawn together along a flat face result in a much wider zone of moving material;
the increase in recovery can be significant; the material in the ring can be drawn to much flatter angles (allowing
flatter ring angles)
ground support: the support is primarily for the brow; the brow has to stay stable for that short period of draw and
then charging of the next ring; support intensity will be much more than required for normal tunnel stability; the
brow has to accept the blast damage and has an extra degree of freedom
blasting:"" have to break the rock to the right fragmentation, even but not too fine and without causing too much
damage but sufficient to fluff the ore and compact the waste; powder factors are generally more than 30% greater
than for un-choked blasting
Some of the changes in layout and operations are illustrated in figure 2.

Figure 2(a) - Effect of interactive draw

Sub-Level Caving: Where Is It Headed?

Figure 2(b) - Effect of draw point spacing

The primary objective for a SLC operation should be delayed dilution entry. Classical theory and practice has
dilution entering at 20% to 40% extraction of the ring tonnage; more recent operations have reportedly achieved 80%
and better; but we have also seen operations where the dilution was overrunning the ore at less than 10% draw. The
later the dilution enters the muckpile the greater the proportion of "clean" ore and the higher the grade factor for the
same extraction. (And to repeat: dilution in this instance is material outside the current ring - if it carries grade so
much the better).

Orebody Conditions
What the orebody will allow you to do is a function of "mining difficulty": geometry, rock mass conditions, major
structure, stresses, grade distribution etc.:
strong rock: allows small pillars; small drive cross-cut intervals increases coverage
competent rock: dependent on jointing and formation of wedges; brows are mostly disturbed by blasting and then
by gravity; competent rock masses allow wide backs; competent rock implies few joints and strong joint surfaces
(little infilling, irregular and rough surfaces); jointing may cause wedge failures in one direction but not in the
other; therefore a design decision might be to align development in a favourable direction
few major structures: prevent massive wedges and/or hole cut-offs
steep dip: keeps the low grade dilution source further away from the current draw points so most of the dilution
comes in over the top of a mixture of ore and waste from caving at much higher levels than the current extraction
level
massive: sufficiently large footprint for high production rates; most dilution comes from the boundary between
ore and waste - the more massive the deposit the smaller the proportion of material from the boundary; the more
massive the deposit the higher the development yield from waste development
fragmentation: must consider the fragmentation from blasting and the fragmentation of the material that caves;
joint frequency, joint condition and joint direction will all affect fragmentation; a competent rock mass is usually

Sub-Level Caving: Where Is It Headed?

the most suitable as larger, widely spaced holes will still result in good fragmentation but the caving will be very
coarse; this results in blasted material finer than the cave and an "even" fragmentation for the ore with a minimum
of both oversize and fines
no "muddy" material: want a minimum of very weak or rapidly weathering material to avoid problems of muck
rushes, over-compaction and hang-ups
caving: caving is seldom an issue with SLC as it usually starts out of the bottom of an open pit; lack of caving
from the hanging wall is usually an advantage as it delays the introduction of dilution; but, choke conditions must
be maintained so at dips below 50o(say) the occurrence of voids can be an issue; caving is a function of rock mass
conditions and the hydraulic radius
"Ideal" is not found in mining. Compromises are inevitable. Currently most massive deposits are either block caved
or if they will not cave readily they are open stoped with cemented fill. It would appear that the concern is with
dilution and recovery; the response being that recovery will be higher and dilution lower with a filling method. This
is partially correct but recoveries are usually well below 100% and dilution often runs at over 15%. A mature SLC
can achieve recoveries in excess of 100% and grade factors well over 80% (proportion of pure waste less than 20%).
This assumes a certain amount of cheating: it assumes that there is mineralised waste around the ore (not unusual);
and that the dilution grade steadily increases (but this is the expected attribute of SLC which must be maximised).

EVALUATION
Like all mining methods, the costs per unit are easy to calculate. The difficulty is in predicting productivity per unit
(primarily the production rate per draw point) and the head grade (a function of planned and unplanned dilution,
which in turn is dependent on the point of dilution entry and the degree of mixing and the grade of the dilution). To
add to this difficulty, there are very few SLC operations. And none of these operations have either fully monitored or
published their characteristic draw behaviour (as far as we know). But this should not detract from what is
potentially a very productive and cost effective method - a method that can deliver the lowest cost-to-metal under
certain circumstances. We are sorely in need of data. But estimations and predictions must still be made:

Grade Prediction
This is the key concern and the one in which current predictive methods do not recognise the benefits of interactive
draw. The following draw model illustrated in figure 3 has been based on models developed for block caving (ref 4)
as well from previous experience with SLC operations. The results from the model agree reasonably well with the
achievements of more recent SLC operations but none of these operations is technically or operationally ideal. We
believe that a well engineered and operated SLC in ideal circumstances (geometry and rock mass) can deliver better
results than predicted by the model.

Sub-Level Caving: Where Is It Headed?

Figure 3 - Mixture of ore and waste with draw

The model shows the proportion of ore (below the line) and dilution (above the line) as the extraction increases
(along the x axis). It can be seen that very high extractions can be achieved dependent on the shape of the curve and
the grade of the ore as well as the grade of the waste. Ideally we should also have a model that estimates the grade of
the dilution (the mixture of pure waste, mineralised waste and the ore left behind from each recovered ring). A
methodology to estimate the dilution bin is illustrated in figure 4.

Figure 4 Mechanics of dilution grade estimation

Sub-Level Caving: Where Is It Headed?

This is very simplistic as we know that the dilution grade will decrease as the ring is drawn and the waste comes
from further away. But in the absence of real data this methodology was used on a recent project and allowed
reasonable sensitivities to be run.

Production Rate
We have used the draw point as the key operational indicator; the number of available draw points will determine the
overall production rate. The longer term rate per draw point (in effect a production drift or cross-cut) will depend on
the time taken for a series of activities: drilling, charging and blasting (including any brow repairs and re-drilling);
mucking and any secondary breakage and any hang-up clearance. Keeping the face flat and achieving interactive
draw will also conflict with the availability of a draw point.
Typical results for reasonably well run operations are 500 to 600t/dy per available draw point. A large LHD will
therefore require some four to five draw points to keep it supplied with broken muck. If interactive draw is also
considered and the draw is across a group of four to five draw points then at least ten draw points may be required
per LHD. It is critical to have as large an LHD as possible operating.
The overall maximum production rate of any mining method is very difficult to estimate on paper but typical
ratesof-fall through the deposit in good conditions have been in the region of 65m per year.

DISCUSSION
There is clearly a conflict between the grandiose title of this paper and the reality of fitting it all into eight pages
including figures. We have not been able to do justice to the subject. But we hope that we have shown that there are
situations where SLC is an appropriate method. In a recent project it showed very superior economics over more
conventional methods achieving over 100% recovery with grade factors over 80% and operating costs of less than
$12/t. And SLC can be very productive in terms of rate-of-fall through a deposit generally double what could be
achieved with filling methods.
The only methods that can compete on cost-to-finished-metal are possibly block caving and partial extraction. But
very few ore bodies are really suitable; they need to be:
strong and competent
have a large footprint with very steep dip
and preferably have mineralised waste
Coupled with this is the uncertainty of outcome as there is very little documentation on draw behaviour on which to
base reasonable draw models. But we are reasonably certain of what contributes to efficient achievements. And
currently there is work being done to confirm and improve the current rather simplistic and rough models illustrated
in this paper.
The major advances are with interactive draw and the understanding of what contributes to the success of SLC and
how to ensure that this success is achieved.

REFERENCES
KVAPIL, R.1982 The Mechanics and Design of Sublevel Caving Systems. Underground Mining Methods
Handbook, SME, pp.880-897
LAUBSCHER, D.H. 1994 Cave Mining The State of the Art, Journal of the SAIMM, Oct 1994, pp 279-293

Article Sources and Contributors

Article Sources and Contributors

Sub-Level Caving: Where Is It Headed? Source: http://www.minewiki.org/index.php?oldid=4920 Contributors: Adasys, Venteditor

Image Sources, Licenses and Contributors

Image:SubLevelCaving01.png Source: http://www.minewiki.org/index.php?title=File:SubLevelCaving01.png License: unknown Contributors: Venteditor
Image:SubLevelCaving02.png Source: http://www.minewiki.org/index.php?title=File:SubLevelCaving02.png License: unknown Contributors: Venteditor
Image:SubLevelCaving03.png Source: http://www.minewiki.org/index.php?title=File:SubLevelCaving03.png License: unknown Contributors: Venteditor
Image:SubLevelCaving04.png Source: http://www.minewiki.org/index.php?title=File:SubLevelCaving04.png License: unknown Contributors: Venteditor
Image:SubLevelCaving05.png Source: http://www.minewiki.org/index.php?title=File:SubLevelCaving05.png License: unknown Contributors: Venteditor

License
Attribution 3.0 Unported
http:/ / creativecommons. org/ licenses/ by/ 3. 0/