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

net/publication/312653068

Open Pit Optimisation and Design: A Stepwise Approach*

Article · January 2016

CITATION READS

1 7,127

2 authors:

Emmanuel John Andrew Appianing Daniel Mireku-Gyimah

University of Mines and Technology University of Mines and Technology
2 PUBLICATIONS   3 CITATIONS    51 PUBLICATIONS   201 CITATIONS   

SEE PROFILE SEE PROFILE

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

Performance of Gas Bags (Air Decking) in drill and Blast Operations at Goldfields Ghana Limited, Tarkwa Mine View project

THe use of GA in underground stope boundary optimisation View project

All content following this page was uploaded by Daniel Mireku-Gyimah on 24 January 2017.

The user has requested enhancement of the downloaded file.

 Open Pit Optimisation and Design: A Stepwise Approach*
E. J. A. Appianing and D. Mireku-Gyimah

Appianing, E. J. A., and Mireku-Gyimah, D. (2015), “Open Pit Optimisation and Design: A Stepwise
Approach”, Ghana Mining Journal, Vol. 15, No. 2, pp. 27 - 35.

Abstract
The erstwhile Nkroful Mining Limited (NML) had a concession with an approximate area of 47.84 km2 located at Nkroful in
the Nzema East District of the Western Region of Ghana. The exploration data, assembled from 1997 to 2003, revealed a
potential gold resource and indicated that the deposit could be mined by using open pit mining method. This paper
demonstrates a stepwise approach to design an optimal open pit to exploit the deposit by using Surpac and Whittle software.
Using the exploration data as the primary input, Surpac is used to create a block model of the deposit. The block model is
imported into Whittle for the open pit optimisation, based on geotechnical and economic parameters derived from
calculations or assembled from mines operating in Ghana. The optimal pit is then imported back to Surpac for detailed pit
design incorporating a ramp and berms. The designed pit contains 5.03 Mt of ore at an average grade of 1.7063 g/t.
However, it has 19.58 % more tonnes of waste, 16.05 % less tonnes of ore and 0.34 % lower average grade of ore than the
respective figures in the Whittle optimal pit. The differences in the figures are due to the widening of the pit bottom and the
creation of a ramp and berms in the designed pit, which resulted in the addition of some waste, ore loss and dilution.

Keywords: Optimisation, Mining, Stripping Ratio, Block Model, Pit Design

1 Introduction (ii) Exportation of the block model to Whittle

software for pit optimisation; and
The erstwhile Nkroful Mining Limited (NML), (iii) Exportation of the optimal pit from Whittle
which was a joint venture company formed by back to Surpac software for detailed pit
Minconsult Limited (ML), Union Mining Limited design.
(UML) and Samax Resources Limited (Samax),
had a concession with an approximate area of 47.84 2.1 Block Modelling of the Orebody
km2 located at Nkroful in the Nzema East District
of the Western Region of Ghana. The exploration 2.1.1 Drill Hole Data Management
data, assembled from 1997 to 2003 (Mireku-
Gyimah, 2005), revealed a potential gold resource Drill hole data processing was carried out in order
and indicated that the deposit could be mined using to present the data in a form suitable for easy
open pit mining method taking into consideration retrieval and analysis, using Surpac software. The
geological, geotechnical and economic parameters. drill hole data was organised into four text files,
This paper demonstrates how to design an optimal namely: collar, survey, assay and geology text files
open pit to exploit the deposit using Surpac and formats required by Surpac software. Tables 1 – 4
Whittle software which are well accepted in the show the arrangement of the fields and records (i.e.
mining industry. columns and rows) in each text file.

2 Materials and Methods Used Table 1 Part of the Collar Text File (Collar.txt)
Hole Maximum
The pit optimisation and design in this work are Y (m) X (m) Z (m)
Id Depth (m)
carried by out using Surpac and Whittle software. A01 31797.38 129893.50 78.0 23.1
The work is based on data gathered from the A02 31752.82 129860.41 78.0 23.1
exploration drilling programme executed by NML
A03 31698.17 129808.78 77.0 23.1
as at 2003. For matters of confidentiality, all drill
A04 31645.00 129752.00 78.0 23.1
hole identification names and numbers have been
changed. A total of 2 626 sample data from 59 drill A05 31564.00 129693.00 78.0 23.1
holes were used for the analysis. The drill holes fell
within 31 100 and 31 900 Northings, and 129 100 Table 2 Part of the Survey Text File (Survey.txt)
and 129 900 Eastings of local coordinates system. Hole Id Depth (m) Dip (º) Azimuth (º)
A01 23.1 -70 90
The work is divided into three essential steps: A02 23.1 -70 90
A03 23.1 -70 90
(i) Block modelling of the orebody, using A04 23.1 -70 90
Surpac software; A05 23.1 -70 90

*Manuscript received August 23, 2015

Revised version accepted November 17, 2015 27 GMJ Vol. 15, No. 2, December, 2015
Table 3 Part of the Assay Text File (Assay.txt) zone string files which were saved and given a
location name and ID range.
Hole Sample Depth Depth Assay
Id Id From(m) To (m) Values (g)
A01 A0116 13.4 14.4 0.12 2.1.4 Creating the Block Model
A02 A0211 8.6 9.6 1.32
A02 A0212 9.6 10.5 1.39 The following steps were taken to create the block
A02 A0213 10.5 11.6 2.08 model:
A03 A0317 15.4 16.5 1.05
(i) Create an empty block model;
Table 4 Part of the Geology Text File (Geology.txt) (ii) Add constraints; and
(iii) Fill the created model with attribute values.
Hole Sample Depth Depth Rock
Id Id From (m) To (m) Type
A01 A0114 11.8 12.6 PH 2.1.4.1 Empty Block Moduling
A01 A0115 12.6 13.4 PH
A01 A0116 13.4 14.4 GQTZ An empty block model was created with the
A01 A0117 14.4 15.5 GQTZ following information: block model identification
A01 A0118 15.5 16.3 GQTZ name, origin, extent, and block size Fig. 3. A user
block size of 10 m × 10 m × 5 m was used to
2.1.2 Loading and Creation of Database conform with mining bench width and height.

This is a prerequisite for database processing and 2.1.4.3 Adding Constraints

also constitutes a data validation procedure
whereby any data input whose description or The addition of constraints was primarily to control
structure is inconsistent with the definitions made the selection of blocks from which interpolations
at the database creation stage is rejected. When the were made or from which information was
database is loaded, it is then ready for data obtained. Blocks falling within the solid model
processing, extraction, plotting and reporting. were ore and waste while blocks falling outside the
There is also room for data update. topography were air blocks. Blocks within the
model that had a grade value below 1g/t were
After loading the database, drill hole layout and considered as waste and those with values of 1g/t
sections of the deposit were extracted from the and above were considered as ore.
database for plotting and display. The drill hole
layout serves two significant functions: 2.1.4.4 Filling the Model with Attribute Values

(i) It assists mining engineers and geologists to At this stage, the empty block model (Fig. 3) was
know the pattern of the drill holes and filled with attribute values. The attributes are the
decide as to which planes to take the properties to be employed during the optimisation
sections through; and process in Whittle software. These were grade,
(ii) It assists mining engineers and geologists to specific gravity (rock density), Mining Cost
check drill hole collar coordinates against Adjustment Factor (MCAF), Processing Cost
manually prepared maps as a way of Adjustment Factor (PCAF), air, waste and rock
verifying the data. type. Some values were assigned directly and
others like the grade by interpolation. The assigned
In all, twelve sections (Fig. 1) were taken; each attributes are in Table 5.
section had an influence of 50 m on either side of
the section line. The sections served as an aid for
the demarcation and digitising of the ore zones or Table 5 Assigned Attributes
orebody outlines using 1g/t as cut-off grade.
Attribute Name Value
MCAF 1.00
2.1.3 Creation of Solid 3-D model PCAF 1.00
Rock Code Air, ore and waste
The demarcated ore zones in the sections were Rock Density (g/cm3) 2.73
digitised on-screen in clockwise direction to form
closed segments that were stitched together to form
a wireframe model. This wireframe model was then
validated to form a solid model (Fig. 2). The output
files formed at the end of digitising became ore

28 GMJ Vol. 15, No. 2, December, 2015

Fig. 1 Digitised Ore Zone Sections

Fig. 2 Solid Model of the Orebody

29 GMJ Vol. 15, No. 2, December, 2015

 LEGEND
Grade >2g/t
1≤ grade ≤2g/t
0< grade < 1g/t
Grade = 0

Fig. 3 Block Model

2.1.4.4 Block Grade Estimation directional semi variographs were drawn and
analysed.
The grade of each block in the block model was
estimated using Inverse Distance Weighting (IDW) The anisotropic ratios were almost equal to 1; this
method, expressed in Equation 1, which is was an indication that the mineralisation is almost
available as a module in Surpac. isotropic. Omni directional semi variograms were
therefore drawn from which the search distance, D s=
1 40 m was deduced. The results of the analysis are
Z
d in
i shown in Table 6.
ZB  (1)
1 Table 6 Variogram Report
 dn Variogram Parameters Values
i
Major / Semi-Major Ratio 0.98
where: ZB is the estimated grade of a block; Major / Minor Ratio 1.05
Max Search Distance of Major Axis (m) 40
Zi is the grade of sample i; Sill (g) 0.94
Nugget (g) 0.86
di is the distance from the centre of the
block of sample i located within a
specified search distance, Ds, around 2.2 Pit Optimisation
the block; and n is a positive integer
power index (Al-Hassan, 2012). Pit optimisation was carried out, using Whittle
software. The resource block model and the
In this study, a power index of +2 was considered economic and technical parameters were used to
acceptable based on a comparative cross-validation produce a set of nested pits. Fig. 4 depicts a
exercise carried out during the pre-feasibility study summarised flow chart for the pit optimisation
of the mine (Mireku-Gyimah, 2005). In order to process.
determine an appropriate search distance, Ds,

30 GMJ Vol. 15, No. 2, December, 2015

 The output then becomes the input for the pit
Importation of block model into Whittle
optimisation. In all, a total of 655 786 arcs were
generated as an output.

Preparation of parameter file 2.2.4 Generation of Pit Outlines

The optimisation carried out produced a set of

nested pit outlines. This was based on the range of
Generation of structure arcs
revenue factor values defined in the parameter file.
A total of 54 nested optimal pits were generated
covering the range of revenue factor values from
Generation of pit outlines 0.3 to 2. This was achieved through the 3-D
Lerchs-Grossman algorithm employed by the
Whittle Optimiser. The results were kept in a
Analysis of pit outlines
results file containing all blocks that must be mined
to obtain the maximum value for a particular pit.
The results file then became the input for analysis.
Exportation of pit outline into Surpac
for detailed pit design 2.2.5 Analysis of Pit Outlines

Fig. 4 Flow Chart for Optimisation in Whittle A number of analyses were carried out by the
optimiser. Microsoft excel was used to prepare a
2.2.1 Importation of Block Model into Whittle graph from the output data to enhance easy
interpretation. The optimal pit was selected based
The parameter file and the contents of the block on worst case and best case scenarios as shown in
model were exported from Surpac to Whittle using Fig. 5. It is a graph of NPV at a minimum rate of
the “block model to Whittle” interface in the 10% against the pits. The best case scenario
Surpac software for pitoptimisation. involves mining out the first pit (the smallest pit)
and then mining out each subsequent pit shell from
2.2.2 Preparation of Parameter File the top down, before starting the next pit shell.
However, the worst case scenario consists of
The parameter file containing essential economic mining each bench completely before starting on
and technical parameters as well as the appropriate the next bench. The advantage of the best case
range of revenue factors needed for optimisation scenario lies in setting an upper limit to the
was set up using FXED in Whittle software. The achievable Net Present Value although its schedule
factors considered include milling and mining is seldom feasible because the push-backs are
recovery factors, cost figures, overall pit slopes, usually too narrow(Anon, 1998).
mining dilution factors and categories of ore to be
processed. Table 7 is a summary of all the cost 2.2.6 Selection of Optimal Pit
figures and parameters used in the optimisation.
Net Present Value (NPV) was used as the criterion
Table 7 Cost Figures and Parameters used for for the selection of the optimal pit in this work.
Optimisation From the result (Table 8), Pit 36 was selected as the
Mining cost per tonne $4.68 optimal pit with a best case NPV value of $105 113
Processing cost per tonne $19.44 678. Table 8 presents the incremental pit value
analysis that shows that the NPV increases
Price of gold $40.6/g ($1134.00/oz)
gradually to Pit 36 and then starts reducing at Pit
Selling cost $0.90/g ($25.70/oz)
37 up to Pit 52.
Capital cost $64 000 000
Discount rate 10 % Table 8 Incremental Pit Values
Mining recovery 95 %
Mining dilution 5% Pit No. NPV($) Increase in NPV
Revenue factor range 0.3 to 2 at 0.02 steps 31 105 105 377 0
Overall pit slope angle 45º 32 105 108 039 2662
33 105 112 230 4191
2.2.3 Generation of Structure Arcs 34 105 112 597 367
35 105 113 528 931
The structure arcs necessary for the optimisation 36 105 113 678 150
were generated using the FXST programme in 37 105 112 847 -831
Whittle software. The purpose of the arcs are to 38 105 111 820 -1027
ensure that the slope angle requirements are
converted into a form suitable for pit optimisation.

31 GMJ Vol. 15, No. 2, December, 2015

 Discounted Open Pit Value for Best Case
Discounted Open Pit Value for Worst Case

Fig. 5NPV vs Pit Number

Fig. 6 Sensitivity Analysis Graph

2.2.7 Sensitivity Analysis Second, based on the selected optimal nested pit,
the effect of varying the mining cost while keeping
First, based on the selected optimal nested pit, the all other economic parameters constant including
effect of varying the gold price while keeping all the gold price was also examined; the base case
other economic parameters constant was examined; figure for the mining cost was also varied from -
the base case figure for gold price was varied from 25% to 25% so as to check the effect of the
-25% to 25% so as to check the effect of the changes in mining cost on the NPV. Fig.6shows the
changes in gold price on the NPV. results of these analyses. The graph indicates that
the NPV is very sensitive to changes in gold price

32 GMJ Vol. 15, No. 2, December, 2015

but marginally sensitive to mining cost assuming (i) Design process;
all other parameters remained constant. (ii) Pit-topography intersection; and
(iii) Addition of haul roads.
2.3 Detailed Pit Design
2.3.1 The Design Process
The optimal pit outline of Pit 36 generated by
Whittle (Fig. 7) was exported into Surpac for the The expanding pit design approach was used in this
detailed pit design. The designing process involved work. Thus the design began from the bottom of
the addition of berms and a haul road to the pit the pit and expanded upwards till the surface
outline generated by the Whittle optimiser. The ‘Pit topography was intersected.
Design’ menu of Surpac software was used to carry
out the design. Prior to the designing of the optimal The design slope method was chosen to define the
pit, a Digital Terrain Model (DTM) of the surface slope angles. This ensured that each lift of the pit
topography and the string files consisting of bench edge was inclined at an angle corresponding to the
outlines (contours) of the Whittle optimal pit were design slope angle. Safety berms were inserted at
generated. Table 9 contains the parameters used to intervals of 10 m. This corresponded to every two-
undertake the final pit design. bench lifts.

Table 9 Pit Design Parameters 2.3.2 Pit-Topography Intersection

Parameter Value A DTM of the designed pit was built and

Bench Slope Angle (Degrees) (°) 70 intersected with the DTM of the surface
Bench Height (m) 5 topography. The results were extracted to form a
Ramp and Haul Road Width (m) 20 pit-topography string file and a DTM.
Ramp Gradient (%) 10
Berm Width (m) 3.2 2.3.3 Addition of Haul Road
Final Slope Angle (Degrees) 45
Bench Width (m) 10 The nature of the pit required the creation of a
ramp to be used for haulge of ore and waste. It was
created on the high elevations (Fig. 8).
The optimal pit was designed as follows:

Fig. 7 Exported Whittle Optimal Pit Outline

33 GMJ Vol. 15, No. 2, December, 2015

Fig. 8 Optimal Pit Design with Switchback Road

3 Results and Discussion Table 11 Results From Final Pit Design In

Surpac
The volumes of ore and waste as well as the
Gold
respective tonnages in the Whittle optimal pit and Volume (m3) Tonnage (t)
(g/t)
in the final pit designed in Surpac were calculated.
Ore 1844000 5034120 1.7063
The results are shown in Tables 10 and 11. The
designed pit contains 5 034 120 t of ore at an Waste 4097000 11184810 0.08
average grade of 1.7063 g/t. However, it can be Total 5941000 16218930
deduced from the two tables that the tonnage of ore Stripping Ratio: 2.22:1
in the designed pit decreased by 16.05 % as Expected Revenue: $331 305 460.10
compared to the tonnage of ore in the Whittle
optimal pit; while the tonnage of waste in the 4 Conclusions
designed pit increased by 19.58 % when compared
to the tonnage of waste in the Whittle optimal pit. The following conclusions were drawn at the end
Also the average grade of ore in the designed pit of the study:
decreased from 1.7121 g/t in the Whittle optimal
pit to 1.7063 g/t (representing 0.34% decrease). (i) Given any exploration data such as that of
The differences in the tonnages of ore and waste NML, Surpac and Whittle software can be
and the average grade of ore in the designed pit and used to design an optimal open pit;
optimal pit are due to the fact that during the pit
design, the pit bottom was widened to give (ii) In this study, the optimal pit outline
adequate room for equipment maneuverability, a generated by Whittle software contained 5
ramp was introduced to facilitate haulage of ore 996 445 tonnes of ore at an average grade of
and waste and berms were added for safety 1.7121 g/t of gold; the NPV of this pit is
purposes, all of which led to the addition of some sensitive to changes in metal price but
waste, some ore loss and dilution. marginally sensitive to cost of mining; and

Table 10 Results From Whittle Optimal Pit (iii) The final pit designed using the Surpac
software contains5 034 120 tonnes of ore at
Gold an average grade of 1.7063 g/t and a total
Volume (m3) Tonnage (t)
(g/t) volume of material (waste and ore) of 5 941
Ore 2196500 5996445 1.7121 000 m3. However, the designed pit contains
Waste 3295000 8995350 0.06 19.58% more tonnes of waste, 16.05% less
Total 5491500 14991795 tonnes of ore and 0.34% lower average
Stripping Ratio: 1.5:1 grade of ore than the respective figures in
the Whittle optimal pit. The differences in

34 GMJ Vol. 15, No. 2, December, 2015

 the figures are due to the widening of the pit Authors
bottom and the creation of a ramp and berms
in the designed pit which resulted in the
E. J. A. Appianing holds HND in Building
addition of some waste, ore loss and
Technology from Kumasi Polytechnic. He
dilution. also holds BSc and MPhil degrees in Mining
Engineering from the University of Mines
References and Technology, Tarkwa, Ghana. He is
currently an Assistant Lecturer in the Mining
Engineering Department of the University of
Al-Hassan, S. (2012), “Mineral Resource Mines and Technology, Tarkwa, Ghana. Areas of his research
Evaluation”, Unpublished MSc Lecture Notes, interest are Mine Planning and Design, Operations Research and
University of Mines and Technology, Tarkwa, Mineral Economics.
Ghana, 162 pp.
D. Mireku-Gyimah is a Professor of
Annels, A. E.(1991), Mineral Deposit Evaluation:
Mining Engineering and a Consulting
A Practical Approach, Chapman and Hall, 436 Engineer currently working at the
pp. University of Mines and Technology,
Anon. (1998), “Four-XTM Strategic Planning Tarkwa, Ghana. He holds the degrees of
MSc from the Moscow Mining Institute,
Software for Open Pit Mines”, Reference
Moscow, Russia, and PhD and DIC from
Manual, 446 pp. the Imperial College of Science, Technology and Medicine,
Anon.(2010), “Open Pit Design”, Datamine User London, UK. He is a member of Institute of Materials, Minerals
Guide, Datamine Software Limited, 62 pp. and Mining of UK and New York Academy of Sciences and
also a fellow of Ghana Institution of Engineers and the Ghana
Anon. (2014a), “Maps for all Over the World
Academy of Arts and Science. His research and consultancy
(Latitudes and Longitudes)”, www.getamap.net. works cover Mine Design and Planning, Mine Feasibility Study,
Accessed: May 15, 2014. Operations Research, Environmental Protection and Corporate
Anon.(2014b),“Weather Forecast for 8 Million Social Responsibility Management.
Locations in the World”, www.storm247.com.
Accessed: May 16, 2014.
Aseidu-Asante, S. K.(2012), “Computer-Aided
Open Pit Optimisation and Design,
Unpublished MSc Lecture Notes, University of
Mines and Technology, Tarkwa, Ghana, 72 pp.
Mireku-Gyimah, D. (2005), “Feasibility Study of
the Nkroful Gold Mining Project, Western
Region, Ghana”, Unpublished Feasibility
Consultancy Report, 80 pp.
Mireku-Gyimah, D. (2013), “Mine Economics and
Financial Evaluation”, Unpublished MSc
Lecture Notes, University of Mines and
Technology, Tarkwa, Ghana, 148 pp.
Whittle, J. (1993), “The Use of Optimisation in
Open Pit Design”, Unpublished Short Course
Notes, Whittle Programming Proprietary
Limited, Melbourne, Australia, pp. 40-41.

35 GMJ Vol. 15, No. 2, December, 2015

View publication stats