Bulletin of the Geological Society of Greece, vol.

XLVII 2013
Δελτίο της Ελληνικής Γεωλογικής Εταιρίας, τομ. XLVII , 2013
Proceedings of the 13th International Congress, Chania, Sept.
Πρακτικά 13ου Διεθνούς Συνεδρίου, Χανιά, Σεπτ. 2013
2013

USE OF MINE PLANNING SOFTWARE FOR THE

EVALUATION OF RESOURCES AND RESERVES OF A
SEDIMENTARY NICKEL DEPOSIT

Kapageridis I.1, Apostolikas A.2, Pappas S.2 and Zevgolis I.2

1 Department of Geotechnology and Environmental Engineering, Technological Educational
Institute of Western Macedonia, Koila, 50100, Kozani, Greece, [email protected]
2
LARCO GMMSA, 81-83 Kifisias Avenue, Athens, Greece

Abstract
The use of mine planning software in the evaluation of deposits and estimation of
mineral resources and reserves has become widespread and is considered a prereq-
uisite for the reporting of reserves according to international reporting standards
and codes. The main principles of operation and application of these codes are
transparency of reported material, the relevance of reported information (materiali-
ty) and the competence of people involved in producing the reports. The application
of specialised mine planning software described in this paper currently takes place
at operational mines of sedimentary nickel deposits in Central Evia, Greece. It in-
volves all stages of processing of exploration data including entry, analysis, pro-
cessing and modelling of data, geostatistical estimation of quantities and qualities of
mineral resources and mineral reserves, and the design, optimisation and schedul-
ing of mining operations.
Key words: geostatistics, pit optimisation, reporting standards.

Περίληψη
Η χρήση λογισμικού μεταλλευτικού σχεδιασμού για την αξιολόγηση κοιτασμάτων και
την εκτίμηση ορυκτών πόρων και αποθεμάτων είναι διαδεδομένη και θεωρείται προ-
ϋπόθεση για την αναφορά αποθεμάτων με βάση διεθνή πρότυπα και κώδικες αναφο-
ράς. Οι βασικές αρχές λειτουργίας και εφαρμογής αυτών των κωδικών είναι η διαφά-
νεια του αναφερόμενου υλικού, η σχετικότητα των αναφερόμενων πληροφοριών (ου-
σιαστικότητα) και η επάρκεια των ατόμων που εμπλέκονται στην παραγωγή των ανα-
φορών. Η εφαρμογή ειδικού λογισμικού μεταλλευτικού σχεδιασμού που περιγράφεται
στην εργασία αυτή λαμβάνει χώρα στα λειτουργικά μεταλλεία ιζηματογενών κοιτασμά-
των νικελίου στην Κεντρική Εύβοια. Περιλαμβάνει όλα τα στάδια της επεξεργασίας
ερευνητικών δεδομένων συμπεριλαμβανομένης της εισαγωγής, ανάλυσης, επεξεργασί-
ας και μοντελοποίησης των δεδομένων, της γεωστατιστικής εκτίμησης των ποσοτήτων
και ποιοτήτων των ορυκτών πόρων και αποθεμάτων, και το σχεδιασμό, βελτιστοποίη-
ση και προγραμματισμό των μεταλλευτικών δραστηριοτήτων.
Λέξεις κλειδιά: γεωστατιστική, βελτιστοποίηση εκσκαφής, πρότυπα αναφοράς.

XLVII, No 3 - 1980
1. Introduction
1.1. Geological Setting
The nickeliferous mineralization in Greece is related to the geotectonic zones of Almopia,
Pelagonian and Sub-Pelagonian, - the main metalliferous regions are situated in Locris, Euboea
and Kastoria. In Central Euboea, iron and nickel ores of Cretaceous age occur, which are of
sedimentary type and consist of stratified lenses and layers, overlain by Upper Cretaceous
limestones and underlain by ophiolites (and in exceptional cases by Jurassic limestones). The
mineralization is either pissolitic or compact with silcretes developed within the ore, the
development of lenticular intercalations or siliceous layers is also common, while silcretes are also
found in the bedrock. A large number of significant deposits exist in the Psachnon area, the Akres,
Katsikiza, Isomata and the Katavolo-Fterada in the Kimi’s area. The deposit used in the study
presented in this paper is from the Akres area (Figure 1).

Figure 1 – Location of Akres deposits in Euboea.

1.2. Available Data
Drilling information from 126 surface drillholes from the NE Akres deposit was provided for the
purposes of this study in the form of CSV files. This data was imported to a drillhole database in
Vulcan. The database has four tables to hold collar, survey, lithology and assay information. The
database was validated using a number of checks for overlapping intervals, numerical field ranges
(coordinates and grade fields), and numeric order (lithology and assay interval fields). The
drillhole lithological information was used to model the boundaries of the orebody in three
dimensions. The assay information was used to perform structural analysis (variography) of the
nickel spatial distribution and grade estimation on the basis of a block model. Financial and
technical information concerning the mining and processing methods used were also available and
were used in the pit optimisation and design stages of the study.
1.3. Procedure
The modelling and resource/reserve estimation study presented in this paper was performed in a
number of steps using Maptek’s Vulcan 3D software – one of the major commercial mine planning
packages implemented by LARCO SA since 2007. The implementation of industry standard
software based mine planning techniques aimed at meeting the requirements set by international

XLVII, No 3 - 1981
reporting codes (SAMREC, 2009, JORC, 2012, NI 43-101, 2011, and PERC, 2013). These
techniques regarded database generation and validation, statistical analysis and compositing of
samples, geological modelling of the orebody, structural analysis (variography), grade estimation
on a block model basis, pit optimisation and design, and reporting of resources and reserves. All
steps were supervised by qualified personnel as required by reporting codes. The implementation
of reporting code guidelines to resources and reserves estimation and reporting using mine
planning software has been examined in the past by Snowden (1996), Duke et al. (1999), and
Kapageridis (2007).

2. Geological Modelling
2.1. Orebody Boundaries Modelling
The structural roof and floor of the main orebody area and of an internal low grade envelope were
modelled using grid model interpolation based on drillhole lithological intervals. The highest
points of a particular lithology code (20) which corresponds to the main orebody area were derived
from the database and were used to generate a grid model of the roof. The same procedure was
used to model the floor from the lowest points of the same lithology code. The triangulation
algorithm with trending was used to interpolate grid points between drillholes in all cases. Figure 2
shows a section through all four structural surfaces.

Figure 2 – West-East section looking North showing roof and floor models of main orebody
area and internal low grade envelope derived from drillhole lithological intervals.
2.2. Unfolding Orebody Deformities
A number of methods for modelling deformations produced by various folding mechanisms have
been developed. Earlier methods of unfolding deposits were based on various geometrical,
mathematical or even manual techniques such as least squares, cylindrical unfolding or the fitting
of splines (Royle 1979, Dagbert et al. 1983, Dowd 1986). Other more recent methods were based
on the use of an unfolded coordinate system for the transformation of every sample and every
estimation point for variography and grade estimation (Newton 1995). Tetrahedral modelling for
variography and grade estimation was developed by Trevor Coulsen in 1995 (Maptek Pty Ltd) and
implemented in Maptek’s VULCAN 3D software package. Further improvements to the original
algorithm were made by Peter Borovina (Maptek Pty Ltd) in 2002. Tetrahedral modelling is a
method of adjusting the search ellipse used in variography and grade estimation to follow the
geometrical structure of the deposit by forming a 3D tetrahedral model of the deposit volume. For
this model to be generated, the structural surfaces of the deposit need to be modelled as surface
triangulations. In tetrahedral modelling the search ellipse is distorted to follow nominated

XLVII, No 3 - 1982
structural surfaces leading to improved estimation accuracy (Kapageridis, 2006) and improved
classification of resources.
A tetrahedral model was generated in this study using the structural roof and floor models of the
main orebody area. This model (Figure 3) captured the orientation and shape of all deformities as
represented by the structural models and was used to bring samples to an unfolded location relative
to blocks during grade estimation.

Figure 3 – 3D representation of tetrahedral model used in unfolding.

3. Variography - Grade Estimation

3.1. Statistics and Sample Compositing
General statistics and histograms of nickel grade were calculated to gain a better understanding of
its distribution characteristics. Original assay intervals were then composited to standard length
composites (1m) to derive a set of equal support values to use for variography and grade
estimation. The choice of composite length was based on the selectivity of the mining method used
in this deposit and also on the effect of the compositing to the statistical characteristics of the
composited values.
3.2. Variography – Structural Analysis
In order to model the structure underlying the Ni grade distribution represented by the available
composites, fix the search region to be used during sample selection, and provide a model to
ordinary kriging for weights calculation, structural analysis was performed using only the
composites from the main orebody area. The procedure applied to derive the final variogram
model is discussed in the following paragraphs.
3.2.1. Downhole Variography – Nugget Effect Approximation
Downhole variography was performed in order to set the nugget effect presented by the
composites used in this study. The standard semivariogram was used in this study as the mode for
experimental variography. The downhole variogram was calculated using nickel composited
values, a lag of 1m to match the composite length, 30 lags, and a lag tolerance of 0.3m. Figure 4
below displays the experimental downhole semivariogram and the fitted model. The range and sill
displayed was also used for the final variogram model as the drillholes are all vertical, hence
correspond to a particular direction in space. The fitted value for the nugget effect was 0.002. This
value was used in the final variogram model fitted in different directions.

XLVII, No 3 - 1983
 Figure 4 – Downhole experimental semivariogram and fitted exponential model.
3.2.2. Directional Variography
Directional experimental variograms of the Ni composited values were calculated in various
directions and were used to fit a model. A lag size of 10 was used in all directions and a total
number of 15 lags were calculated, i.e. a semivariogram value was calculated every 10m and up to
150m. A lag tolerance of 2m was used to allow for irregularities of the drilling pattern. Azimuth
and plunge tolerances were set to 20o and a horizontal and vertical bandwidth of 20m was applied.
Figure 5 explains in graphical form the way these parameters are applied when searching for pairs
of samples to calculate experimental variogram points.

Figure 5 - Explanation of lag and direction setup parameters for directional variography.
3.2.3. Variogram Model
A variogram model consisting of a single exponential structure was fitted to the experimental
variogram as shown in the two indicative directions in Figure 6. The exponential model was

XLVII, No 3 - 1984
chosen as it seemed to match better the curvy shape of the experimental variogram at distances
between 0 and 50 meters. The spherical model which is the most common choice in similar studies,
has an almost linear beginning that did not fit well in this case. Table 1 gives the parameters of the
final variogram model used for grade estimation.

Figure 6 – Experimental semivariograms and fitted exponential model (point labels are pair
counts).
Table 1 – Fitted Variogram Model Parameters.

Experimental Variogram Type Standard Semivariogram

Fitted Model Type Exponential
Nugget Effect 0.002
Sill Differential 0.04
Bearing 45
Plunge -15
Dip 0
Major range 65
Semimajor range 42.85
Minor range 8.2

XLVII, No 3 - 1985
3.3 Grade Estimation and Resource Classification
A block model with 10x10x12m main blocks and a minimum sub-block size of 1x1x1m was
generated covering the entire area of the deposit. The main and sub-block sizes were set to reflect
pit geometry and mine method selectivity. The model was not rotated around any axis in order to
be used for pit optimisation. Ordinary Kriging (OK) was used as the method for interpolating Ni
values to the blocks coded as part of the main orebody. OK used the fitted variogram model
consisting of a single exponential structure as described in Section 3.2. Three separate estimation
runs were performed, each with different sample selection strategies and varying search ellipsoid
dimensions, corresponding to three different resource classifications. In each run the blocks
estimated received a flag in the class variable of the block model according to the classification: 1
for measured, 2 for indicated, and 3 for inferred. Blocks estimated with the 1st run (measured)
were excluded from the 2nd run (indicated), and blocks estimated with the 1st and 2nd run were
excluded from the 3rd run (inferred). Table 2 summarises the parameters used in each run and then
number of blocks estimated.
Table 2 - Summary of Grade Estimation Parameters.

Measured Indicated Inferred

Bearing 45 45 45
Plunge -10 -10 -10
Dip -25 -25 -25
Major 32.5 65 65
Semi 21.4 42.8 42.8
Minor 8.2 8.2 8.2
Minimum Samples 8 4 4
Maximum Samples 16 16 16
Octant Based Search Yes Yes No
Maximum Samples per Octant 2 2 -
Minimum Octants 4 - -
Minimum Samples per Octant 2 - -
Blocks Estimated 9958 27528 1

4. Pit Optimisation and Design

4.1. Estimation and Application of Technical and Financial Parameters
Technical and financial parameters related to the mine geometry and location, mining method,
haulage and ore processing were calculated for each block individually using a script written in
Perl and utilising Vulcan’s Lava extensions to this popular scripting language. The parameters set
by the script and the values used for the purposes of this study are shown as entered in the panel in
Figure 7. The calculated values for the various parameters were used to estimate a total block
value that was used for pit optimisation using the following equation:
block_value = revenue – mining_cost – haulage_cost – crushing_cost – metallurgy_cost –
sea_transport_cost – other_block_cost
The values for the each of the parameters were also stored in separate block model variables for
validation purposes. Through this process, all the applicable resources to reserves conversion

XLVII, No 3 - 1986
parameters according to reporting standards were applied to the resource block model before using
it for pit optimisation.

Figure 7 – Configuration panel for pit optimisation mining and financial criteria generated
by the script.
4.2. Pit Optimisation
The block value calculated in the previous step was passed to Vulcan’s Pit Optimiser module – an
implementation of the Lerchs-Grossman algorithm. Two different angles were used to control the
slope of the final pit limits, matching the bench angle and height, and berm width configuration
and also allowing for the presence of haul road on one side of the pit. A 45 o pit slope was used in
the East side of the pit and a 25o slope in the west side (the side of the haul road). The pit slope
was also controlled by a lithology block model variable. The result of pit optimisation was stored
as a code in a block model variable. Blocks that belong to the optimum pit (including ore and
waste blocks) receive a different value from those that do not. This way, a subset of the resource
blocks were converted to potential reserve blocks. However, the optimum pit as presented by the
contours shown in Figure 8 is not a complete pit design and can only be used as a guide for the
next important step of designing a feasible pit.
4.3. Pit Design
Pit design was performed using the optimum pit limits as a guide (Figure 8). The top part of the pit
from the lowest open bench, i.e. the lowest bench that could be accessed from the surface, was

XLVII, No 3 - 1987
designed with projections upwards and outwards of the corresponding optimum pit limit to
generate the toe and crest polygons without a haul road. The bottom part of the pit from the highest
closed bench, was designed using a more complex function in Vulcan that inserts the haul road and
then projects the toe and crest polygons for each bench. Two switchbacks were introduced to
maintain the haul road only on the west side of the pit as this was the side with the lower pit slope
set in pit optimisation. The benches were 12m high with 8m berms and a 70 o batter angle. The haul
road was 12m wide with a 10% grade applied to the shortest side. The haul road polygons were
adjusted to provide bench access both inside and outside of the road. The complete pit design was
intersected and cropped with current topography.

Figure 8 – Optimum pit contours and final pit design.

5. Resources / Reserves Reporting

In order to compare the estimated reserves with the production figures for the period up to the end
of 2012, a solid triangulation model was generated representing the volume of the actual
excavation at the end of this period. Table 3 gives a summary of the total estimated resources for
the NA Akres deposit, the estimated reserves inside the solid model and the actual reserves mined
from the same volume. There is some over-estimation of tonnage combined with an under-
estimation of grade which translates to very similar nickel content in both cases (6,005t actual
versus 6,344t estimated).

6. Conclusions
This paper discussed most aspects of the application of mine planning software to the evaluation of
mineral resources and mineral reserves of the sedimentary nickel deposits in Euboea. The
implementation of commercial mine planning software produced results that increased confidence
as to the available resources, aided the configuration of the mining methods applied, and helped in
planning future mining operations by developing different mining scenarios with speed and clarity.
The adoption of a geostatistical approach to grade/reserves estimation increased confidence in the
produced results and reduced the risks associated with the estimates. The use of a pit optimisation

XLVII, No 3 - 1988
tool helped convert resources to reserves with more confidence and in a more standardised fashion
that is widely accepted by the mining industry. The procedures described in this paper have been
adopted with minor adjustments to other nickel deposits exploited by LARCO SA in other areas of
Greece.
Table 3 – Summary of ore resources and reserves from NA Akres deposit.
Actual
Total Estimated Estimated Reserves Reserves
Resources (2011- 2012) (2011-
2012)
Measured Indicated Measured Indicated Total Total
Ore Tonnage 587,812 1,108,906 96,851 479,925 576,777 508,926
Ore Ni Grade % 1.07 1.08 1.15 1.09 1.10 1.18

7. References
Dagbert M., David M., Crozel D. and Desbarats A. 1983. Computing variograms in folded, strata-
controlled deposits, Geostatistics for Natural Resource Characterisation, NATO A.S.I. Se-
ries C, Volume 122, Part 1: 71-90.
Dowd P.A. 1986. Geometrical and geological controls in geostatistical estimation and orebody
modelling, 19th International Symposium on the Application of Computers and Operations
Research in the Minerals Industries (APCOM), Colorado, Chapter 8: 81-94.
Duke J. and Hanna P. 1999. Computer-based Resource Estimation in Accordance with the 1999
JORC Code, White paper, ECS International Pty Ltd.
Joint Ore Reserves Committee. The Australasian Code for Reporting of Exploration Results, Min-
eral Resources and Ore Reserves (JORC 2012).
Kapageridis I. 2006. Use of Tetrahedral Modelling for Variography and Grade Estimation of a
Structurally Deformed Phosphate Deposit. 11th International Congress for Mathematical
Geology (IAMG’06 – Quantitative Geology from Multiple Sources), International Associa-
tion for Mathematical Geology, Liege.
Kapageridis I. Implementation of Reporting Code Guidelines for Mineral Resources and Mineral
Reserves Using a General Mine Planning Package, in: 3rd International Conference on
Sustainable Development Indicators in the Minerals Industry (SDIMI) - Entering the 'Inter-
national Year of Planet Earth', Milos Conference Series, Milos 2007.
National Instrument 43-101, Standards of Disclosure for Mineral Projects, Canadian Securities
Administrators, 2011.
Newton M.J. 1995. Structural analysis for folded deposits. Mining Magazine, August 1995.
Pan-European Reserves and Resources Reporting Committee, Pan-European Standard for Report-
ing of Exploration Results, Mineral Resources and Reserves (PERC Standard 2013).
Royle A.G. 1979. Plane projections of tabular orebodies for evaluation purposes, Transactions
IMM, Volume 88: A87-A91.
Snowden D.V. 1996. Practical Interpretation of Resource Classification Guidelines, AusIMM An-
nual Conference “Diversity, the Key to Prosperity”, Perth.
SAMREC/SAMVAL Committee. South African Code for Reporting of Exploration Results, Min-
eral Resources and Mineral Reserves (SAMREC 2009).

XLVII, No 3 - 1989