You are on page 1of 4

McCarthy P L 2001.

Mining Dilution and Losses in Proceedings Underground


Mining, in Mineral Resource and Ore Reserve Estimation The AusIMM Guide to
Good Practice (Ed: A C Edwards), pp333-336 (The Australasian Institute of
Mining and Metallurgy: Melbourne). Reprinted with permission of The
Australasian Institute of Mining and Metallurgy.

Mining Dilution and Losses in


Underground Mining
By P L McCarthy 1

Abstract
Resource modelling and Ore Reserve estimation procedures are different for open pit and
underground mines. A statistical approach to the spatial location of orebody limits is unhelpful in
underground Ore Reserve estimation, while experience with practical mining outcomes and
economics is fundamental.
Resource modelling for underground ruining relies heavily on geological interpretation and
experience. When converting Mineral Resources to Ore Reserves, the chosen grade interpolation
technique, while important, has less significance than the raining, geotechnical and economic
considerations which determine mining dilution and recovery.
Mining dilution and recovery are difficult to measure and more difficult to predict. There is no
alternative to careful measurement coupled with experience-based adjustment. It is possible that a
`Competent Person' for the purpose of preparing a Mineral Resource estimate may not be `competent'
to estimate Ore Reserves for the same deposit.

Introduction

Definitions

Conversion of a Mineral Resource estimate to an Ore Reserve


estimate is a team effort involving, at minimum, a geologist,
metallurgist and mining engineer. The factors to be considered
relate to practical mining outcomes and economics, so the
Competent Person preparing the Ore Reserve estimate must be
very familiar with the proposed mining methods. The main
considerations are the amount of lower-grade or waste material
that will become mixed with the ore (mining dilution), and the
proportion of the resource that can be economically recovered
(mining recovery).

Dilution may be defined in several ways. To the metallurgist


receiving the ore for treatment, it is the percentage of the
delivered material which is waste.
Thus;
Dilution (%) = (mass of waste) x 100 / (mass of ore + mass of waste)

(1)

The mining engineer often expresses dilution as a tonnage


increase.
Thus;

Face-to-face involvement of the geologist who prepared the


resource estimate is essential. The assumptions and limitations
inherent in the resource model must be drawn out. The
background of the resource geologist and his or her experience
with underground mining estimates (as distinct from open-pit)
should be understood.
For reasons set out in this paper, conversion of a Mineral
Resource estimate to an Ore Reserve estimate is a challenging
task. The use of `text book' factors for dilution and recovery is
likely to lead to errors.

1. Managing
Director,
AMC
Consultants
19/114
William
Street,
Melbourne
E-mail: pmccarthy@amcconsultants.com.au

Pty
Vic

Ltd,
3000.

Dilution (%) = (mass of waste) x 100 / (mass of ore)

(2)

Formulae (1) and (2) ignore the fact that `waste' may contain
payable values, so that the economic impact of dilution is less
severe. Dilution may also be expressed as a grade reduction.
Thus;
Dilution (%) = (resource grade - diluted grade) x 100 / (resource grade)

(3)

All of the above measures of dilution are acceptable so long as


they are defined before use. An example of the
misunderstandings that may otherwise arise is given by the
following example. Consider 100 t of ore of ten per cent grade
diluted with 10 t of material of four per cent grade, to give 110 t
at 9.127 per cent grade.
Equation (1) gives: 10 x 100 / 110 = 9.1% dilution.
Equation (2) gives: 10 x 100 / 100 = 10% dilution.
Equation (3) gives: (10 - 9.127) x 100 / 10 = 8.7% dilution.

Mining Dilution and Losses in Underground Mining

Mining recovery may also be expressed in a variety of ways as


follows:

What percentage of the total resource tonnage will


ultimately be mined'?

How does the diluted tonnage delivered to the mill


compare with the estimated resource tonnage?

What percentage of the total metal contained in the


resource will be delivered to the mill?

What percentage of the resource (tonnage or


contained metal) calculated at the resource cut-off
grade will be mined (or delivered for treatment) at the
chosen mining cut-off grade?

After elimination of those parts of the resource


deemed inaccessible or otherwise uneconomic (for
reasons of width, dip, deleterious elements, rock
conditions, etc), what proportion of the remainder
will be recovered after leaving supporting pillars.

There may be other ways of defining mining recovery; it is


sufficient to state accurately what is meant by the term.

Suitability of the Resource Model


Resource models and the Ore Reserve estimation procedure are
different for open-pit and underground mints. To a scale of tens
of metres, the location and spatial distribution of values may be
unimportant in an open pit resource model. Provided the pins
located to access the mineralised zone, any valuable material
can be identified by grade control sampling and then marked
out for mining. Thus, the emphasis in open-pit resource
modelling is on the global accuracy of estimates of tonnes and
grade and internal variability at a scale that might affect pit
optimisation, so a statistical approach is often appropriate.
For underground mining the thickness, dip, continuity and
spatial relationship of ore zones, the regularity of wall contacts,
strength of ore and wall rocks are all critical inputs to the Ore
Reserve estimate. These are drawn from the resource model, or
from the geologist's knowledge of the deposit gained during the
data-gathering and modelling phase. Some parts of the resource
may be impossible to mine; others may be located too far from
development to be economic; others may suffer severe dilution.
A sectional interpretation by an experienced geologist at one or
more possible cut-off grades is usually the first step in
preparing the resource model. This sectional interpretation will
include features inferred from the drill logs that would not be
generated by any grade interpolation software. The geologist's
experience tells him or her how variable the ore boundaries are
in this type of deposit and what shapes the variations might
take. When the sections are linked and wireframed, then
checked and corrected in plan and back to section, the resulting
three-dimensional outlines can be used to validate the gradeinterpolation process.
The above procedure is usually necessary regardless of whether
or not the limits of the mineralisation envelope have been
interpreted (at a subeconomic cut-off) and wireframed as limits
to the grade block model. This is because the shape of the
mineralisation envelope may be quite different from the shape
of the economic material.
It may be acceptable to let the grade interpolation process
determine limits of economic mineralisation in large deposits to
be mined by caving methods, where the edge inaccuracies of
AMC Reference Library www.amcconsultants.com.au

the model become insignificant. For other cases, the geologist


must form a view about the spatial limits of ore at the chosen
cut-off grade, and must be prepared to model the ore
boundaries realistically. To do this, the geologist needs to
understand the style of mineralisation and to be able to infer
irregularities, including structural dislocations such as faults, at
a scale smaller than the drill spacing.
The mining engineer will design slopes which have geometric
limitations dictated by geotechnical factors, the economic
spacing and length of production blastholes, or the need to
combine blocks of `ore' and `waste' into mineable units. When
these shapes are overlaid on the resource model, the resource
grade is diluted and some of the resource is lost. The resource
model will contain internal dilution according to the model
block size (based on the assumed Selective Mining Unit) which
may or may not accord with the engineer's proposed method.
In an underground mine, levels are planned on the basis of
relatively coarse-spaced drilling. Stopes are designed and then
mined with limited flexibility for change. Pods of ore not
identified by drilling will be lost, even if they are expected
statistically to be present.
Proponents of geostatistics sometimes claim that a resource
model inherently contains an appropriate allowance for internal
and edge dilution. This is an obvious fallacy; the dilution
estimate must derive from mining and geotechnical
considerations. A resource model which purports to include
dilution provides an undefined starting point for the Ore
Reserve estimator, who must somehow `remove' the diluent
material from the model before adding back mining dilution.
This is an impractical task, so the only satisfactory approach is
to refuse to accept such a model as a basis for an underground
mine ore reserve.
The resource estimate for underground mining must include a
statement of:

cut-off grade,

minimum mining width,

vertical limits (top and bottom RL), and

lateral extent (plan limits).

The estimate should include a grade-tonnage curve. This


enables the mining engineer to consider strategic sloping
options (high tonnage bulk mining vs low tonnage selective
mining). The estimate should also quote resources at varying
cut-off grades and minimum widths to enable economic
optimisation.

Measurement of Dilution and Recovery


In many mines, ore from a number of sources is stockpiled and
blended before treatment, making reconciliations difficult or
impossible. Assuming reconciliation is possible, the resulting
calculations of dilution and recovery may reflect on the
accuracy of the Mineral Resource and Ore Reserve estimates
rather than on actual mining performance. As in open pit
mining there are several levels of reconciliation that may be of
interest as performance measures:

How does the material treated compare with the


Mineral Resource estimate'?

How does the material treated compare with the Ore


Reserve estimate'?
2

Mining Dilution and Losses in Underground Mining

How does the material treated compare with the slope


design estimate, which was based on infill drilling,
and perhaps a `call factor' grade adjustment`?

What was actually drilled out, charged and fired (as


distinct from design)?

What was delivered for treatment as measured by


truck factors, load cells, weightometers and grab
samples?

What was really extracted from a stope as calculated


from slope surveys and back-calculation using all
available data?

When overbreak occurs beyond the slope design line, it may


introduce unexpected high-grade ore, low-grade, waste or a
process contaminant such as graphite in a contact shear zone.
Thus there is usually no direct correlation between
measurements of slope overbreak and the variation of recovered
metal from the treatment plant. Expressed globally in relation to
the resource estimate, 'dilution' is an experience-based
adjustment that takes account of a number of subjectively
assessed variables.
Among the less obvious variables are:

mixing of waste and spillage into ore in passes and


onto stockpiles;

blasthole damage to slope walls;

turn-around time from grade control sampling to


mark-up;

selective mining by resuing or in-slope sorting; and

loss of free gold in mining and transportation.

Recently surveying instruments have become available which


enable very accurate three-dimensional profiles of a slope void
to be determined. These are invaluable for mine planning,
reconciliation and management, and their use can provide an
improved understanding and control of dilution. There is a
growing database of these measurements.

Predicting Dilution
For ore reserve purposes dilution must be estimated from data
obtained from drilling and development, and from experience.
Key variables are:

the mining method and size of equipment;

grade variability at the resource boundary;

ore width, dip, geometry and continuity;

grade control method and proposed mining rate, and

slope design criteria, including hydraulic radius, RQD


and pillar dimensions

The availability of digital resource models has led some


practitioners to calculate diluted grades based on an assumed
average thickness of overbreak. For example, 0.5m on each
wall of a 3m wide stope represents a 33 per cent tonnage
increase. Grades from assays or composites within this
envelope are used to interpolate a diluent grade. Caution is
needed, as the search ellipsoid used for `ore' may have already
considered this material, or conversely the grade of this
material may be related to the sample grades lying outside the
diluent boundaries in `ore' or `waste'.
AMC Reference Library www.amcconsultants.com.au

In practice, slope overbreak usually takes an arcuate shape,


deepest at the mid-point of the slope and minimal al the pillar
sides. In large open stopes (20m spans), the `normal' arch may
be 3m deep at mid-span. If drill assays outside the stope (or
ore) limits are statistically analysed to calculate diluent grade,
then this shape must be allowed for. The shape of overbreak
may be predicted using techniques such as the Radius Factor
(Dunne and Pakalnis, 1996).
Where there is a sharp geological cut-off between ore and
waste, simple geometric analysis, assuming dilution at zero
grade, is often satisfactory. Where the boundary is gradational
(a fiat grade-tonnage curve at the chosen cut-off) then some
credit should be given for values in the diluent.
Dilution is greatest in narrow ore zones with sharp contacts,
and least in massive ore with gradational boundaries. Dilution
from backfill may be significant. If pillars are to be extracted
against freestanding cemented fill in open slopes, then the
stability and likely frequency of fill failure must be considered,
even if rigorous control procedures are in place. In cut-and-fill
mining, more dilution may be experienced if the slope miners
are paid on a piecework (tonnage) contract because they may
dig deeper into the fill floor.
Dilution can be reduced over time as experience is gained and
the mining method is optimised. Decisions about the level of
slope-wall support using cable dowels are based on cost-benefit
analysis, and will affect dilution. As a general guide the
following suggestions are made for dilution expressed using
Equation 3 above:

dilution is not less than five per cent unless an error


was made in the resource estimate;

for selective methods (eg cut and fill), dilution is


typically ten per cent;

for open sloping dilution is typically 15 to 20 per cent


but can be more;

for caving methods dilution is 20 to 30 per cent; and

for narrow vein mining, dilution of 50 to 100 per cent


is not uncommon.

Exceptions can be found to the above guidelines, and will be


dependent on ore width, dip and stratigraphy. Improvements are
possible with good mining practice. Contract mining on
tonnage-based and metres-based schedules of rates may require
more rigorous management to ensure control of dilution.
It may be useful to use dilution reconciliations for a similar
orebody and mining method as a check. This should be done
carefully with regard to the definition of dilution, the use of
hidden `call factors', and the methods of grade control and
ground support employed.
The proportion of a resource that can be recovered is typically
70 to 90 per cent after removing `inaccessible' or uneconomic
blocks. The higher recoveries can be justified using more
selective mining methods in ore of higher unit value. With all
methods, some resource that would otherwise be classified as
ore will be left in pillars or abandoned due to premature ground
failure.
Pillar recovery may be justified as part of the on-going mining
process or as a retreating salvage operation at the end of mine
life. For example, pillars were reduced on retreat in the
Cadjebut room and pillar operation, giving an improved
3

Mining Dilution and Losses in Underground Mining

recovery compared with the initial Ore Reserve assumptions. In


general, resources remain 'open' along strike or at depth for
most Ore Reserve estimates (i.e. there are Inferred Resources),
so that any error in the recovery estimate is rendered
inconsequential in time after further exploration and conversion
to Ore Reserves.
In preparing a feasibility study it is critical to estimate mining
recovery accurately so that the tonnes of ore delivered to the
mill in the life-of-mine schedule relate to the expenditure on
development and the amortisation of capital. As the mining
recovery is increased, less capital and operating costs are
incurred in accessing each tonne of ore. Whether the increased
recovery is desirable depends on how quickly the
corresponding sloping costs increase in achieving the higher
recovery.

Conclusions
The Ore Reserve estimate derives from a Mineral Resource
estimate. For an underground mine, particular limitations are
placed on the resource modelling technique. In particular,
geostatistical models which purport to include dilution are
likely to lead to errors in estimation.
The conversion of a Mineral Resource estimate to an Ore
Reserve estimate for an underground mine requires
consideration of mining dilution and mining recovery. These
two variables are the result of a multitude of factors that are
difficult to assess. Thus careful measurement, management,
judgment, experience and a thorough understanding of the
proposed mining method are required. The use of `textbook
factors' by inexperienced practitioners should he avoided. It is
possible that a `Competent Person' for the purpose of preparing
a Mineral Resource estimate may not he 'competent' to prepare
an ore reserve estimate for the same deposit.
There are several ways of expressing mining dilution and
recovery, all of which are valid. It is essential that these terms
be defined wherever they are used.

References
Dunne, K and Pakalnis, R C, 1996. Dilution aspects of sublevel
retreat .stope at Detour Lake Mine Rack Mechanics, (Eds:
Aubertin, Hassani and Mitri) (Balkema).

AMC Reference Library www.amcconsultants.com.au

You might also like