Int J Coal Sci Technol (2019) 6(1):84–94

https://doi.org/10.1007/s40789-018-0234-1

A system-wide approach to minimize the operational cost of bench

production in open-cast mining operations
Burak Ozdemir1 • Mustafa Kumral1

Received: 13 August 2018 / Revised: 2 October 2018 / Accepted: 4 December 2018 / Published online: 12 December 2018
 The Author(s) 2018

Abstract The production cycle of open-cast coal mines generally includes drilling, blasting, loading, hauling and coal
preparation activities. Individual optimization of these activities does not mean that the whole system is optimized. This
paper proposes a cost model considering all activities in mining cycle and system-wide approach to minimize the total
mining cost of bench production. Since the fragmentation size and blast-hole diameter are linked to all activities of mining
system, they are considered as decision variables in the problem formulation. The operation costs are then minimized by
using the evolutionary algorithm. Moreover, the impact of the change in the explosive price, and the hourly unit cost of
equipment on total mining cost is quantified by sensitivity analysis. A case study is implemented to demonstrate the
developed model.

Keywords Mine production cycle  Rock fragmentation  Cost optimization  Evolutionary algorithm  Sensitivity analysis

List of symbols L Bench length

A Rock factor Mic Crushing work index
B Burden n Blast geometry constant
CB Blasting cost N Number of blast hole
CC Crushing cost P50 Mean fragmentation size
CD Drilling cost P80 80% passing size
cD Unit cost of the drilling machine pe Price of ANFO
CH Hauling cost Pmax Maximum particle size
CL Loading cost Qe Required amount of ANFO
cL Unit cost of the loader R Maximum allowable feed size of crusher
CM Total mining cost of a bench rD Drill rate
cT Unit cost of the truck S Spacing
D Hole diameter Se Relative weight strength of explosive to ANFO
F Fill factor tc Truck cycle time
H Bench height THT Total hauling time
J Sub-drill tl Loading time
K2 Crusher circuit constant TLT Total loading time
TS Shovel bucket capacity (tonne)
TT Truck capacity (tonne)
& Mustafa Kumral u Energy price
[email protected] V0 Volume of the rock
1
Department of Mining and Materials Engineering, McGill
VS Shovel bucket capacity (m3)
University, 3450 University Street, Montreal, QC H3A 0E8, VT Truck capacity (m3)
Canada

123
A system-wide approach to minimize the operational cost of bench production in open-cast… 85

W Bench width achieved by minimizing the cost of these activities (Kozan

Wc Specific energy and Liu 2017). However, optimizing a single subsystem
x1 Crusher feed size does not mean that the whole system is optimized. To
x2 Crusher product size optimize entire system for more cost-efficient operations, a
qe Bulk density of explosive system engineering approach is required. The systematic
qr Bulk density of rock approach should be used to identify and evaluate the
designs of engineering systems.
Although each of these activities includes various cost
elements, rock fragmentation size affects the cost of all
activities (Fisonga et al. 2017). The relationship between rock
1 Introduction fragmentation and the cost of the mining subsystems is given
in Fig. 1 (Nielsen 1983). The particle size obtained through
Mining companies are usually price-takers. Therefore, they rock fragmentation affects loading, hauling and crushing
do not have much power to set the price which is highly performance. Higher rock fragmentation means higher dril-
volatile in an unpredictable manner. Unfortunately, the ling and blasting costs. However, increasing fragmentation
mining industry has no universally accepted tools to predict provides more cost-effective loading and hauling cycles.
the price direction and its fluctuations. The past evolution Similarly, the crushing cost is likely to decrease with
of price and transactions in the future markets provide very decreasing feed size. More information about the production
limited control over the prices. As a result, the mining activities can be found at Chugh and Behum (2014).
companies concentrate on finding efficient cost reduction
approaches as well as the price prediction/control. One way 1.1 Drilling and blasting
to reduce the costs is to invest in new technologies.
However, the investment in new technologies may not The performance of drilling and blasting (rock fragmen-
always be secure, at least in short-term. For example, tation) is measured by fragmented space and particle size
sufficient financial resources or skills may not be available. distribution of fragmented material. The particle size dis-
Hence, the focus inevitable becomes how to improve tribution is a function of the blast design and the specific
operations in such a way as to take cost advantages. explosive charge of each hole (Ouchterlony et al. 2017).
In an open-cast mining operation, the production cycle Blasting design parameters (such as bench height, burden,
comprises drilling, blasting, loading, hauling and coal spacing, explosive charge, etc.) affect the rock fragmen-
preparation subsystems. Significant cost savings can be tation size. Particle size of blasted material (1) increases if

Fig. 1 The relationship between rock fragmentation and the cost of the mining activities

123
86 B. Ozdemir, M. Kumral

the burden to hole diameter ratio increases, (2) decreases if the blasted material influence the efficiency of loading and
the spacing to burden ratio increases, (3) increases if hauling activities (Taherkhani and Doostmohammadi
stemming length to burden ratio increases, (4) decreases 2015).
with the increase in explosive charge, and (5) decreases if Hauling is another mining activity needed to be opti-
bench height to burden ratio increases (Singh et al. 2016). mized for mining cost minimization. The production rate
The effect of drilling and blasting geometry on particle size and cost of hauling are substantially influenced by the
distribution can be estimated by using mathematical magnitude of the payloads of the trucks (Dickerson et al.
modelling (Kecojevic and Komljenovic 2007). 1986) which highly depends on a fill factor decided by the
Since the cost of all mining activities consists of the particle size of the blasted material. Lower fragmentation
fragmentation size, the mining cost can be minimized by size results in higher fill factor and higher production rate
optimizing the mean particle size after blasting. Moreover, for the trucks. However, when weight capacity of the
the blast design parameters change depending on the drill trucks is reached, production rate remains constant even if
hole diameter, and the blast design parameters determine fragmentation size decreases. In this case, unnecessary
the drilling and blasting cost. Therefore, drill hole diameter fragmentation causes an increase in unit cost of drilling and
is the other key factor in the calculation of drilling and blasting. If the bulk density of the loaded material is small,
blasting cost. These two factors can be used as the decision the lowering the fragmentation size can be considered to
variable in the optimization problem as they can be mod- increase the truck’s production. Otherwise, the weight limit
ified during the mining operation. The blast pattern can be of the trucks does not allow to increase production even
altered for different blasts in the same operation. Moreover, though the fragmentation size decreases.
the hole diameter can also be changed by changing the size
of the bit as soon as the bit is compatible with the drilling 1.3 Crushing
machine.
The impact of blasting on crushing and grinding was also
1.2 Loading and hauling investigated in mining literature (Workman and Eloranta
2003). Drilling and blasting operation determined the
The optimization of loading subsystems is crucial to min- particle size of the crusher feed. Since more energy
imize the total mining cost. The productivity of loaders is required to break coarser feed size (Pothina et al. 2007),
affected by the characteristics of muck piles such as angle smaller fragmentation size can provide substantial cost
of repose, looseness, particle size distribution and moisture savings for the crushing process by reducing the energy
content (Singh and Narendrula 2006). Inefficient blasting consumption. However, decreasing particle size causes an
causes lower productivity for shovels. The looseness of the increase in the production as more material is loaded and
blasted material has a direct relationship with the mean hauled in a cycle. In this case, the production capacity of
particle size and the index of uniformity; hence, the bucket crusher may be exceeded which leads to a queue at the
fill factor is affected by the looseness. Higher bucket fill crusher and the opportunity costs for the hauling and
factor value is favorable for the production and is obtained loading equipment.
by the decreasing particle size of the material. The loose- Because of the abovementioned impacts of the blasting
ness and angle of repose of the muck pile affect swing on the other mining subsystems, its optimization is crucial
time. The optimal muck pile design provides shorter swing in an open-cast mining operation. This optimization can be
time which leads to more cost-efficient loading. Further- achieved by using different techniques such as genetic
more, the moisture content of the muck pile has an impact algorithm (Monjezi et al. 2011), manual simulation model
on the angle of repose and stickiness of the material. (Neale 2010) or particle swarm algorithm (AminShokravi
Osanloo and Hekmat (2005) investigated the relation- et al. 2017), etc. Testing different blast patterns and eval-
ship between particle size of the blasted material and uating their results on drilling and blasting costs can be
shovel productivity in a surface mine based on Rosin– another alternative for the drill and blast optimization
Rammler–Bennett expression. It is concluded that 50% (Kahriman and Ceylanoalu 1996; Martin 2006).
increase in particle size causes 20% loss in shovel pro- In mining literature, most of the research focused on
ductivity. Moreover, Bogunovic and Kecojevic (2011) optimizing mining subsystems individually. In such an
quantified the impact of bucket fill factor on the shovel approach, the cost can convey to the subsequent subsys-
production and energy consumption. Furthermore, the tems. In other words, while the cost of one subsystem is
efficiency and cost of mining activities are also related to minimized, the cost of following subsystems may increase.
the rock mass mechanical parameters such as RQD, com- In fact, focusing on an individual subsystem may cause
pression strength, joints, discontinuities, etc. The quality, missing the relationship between the subsystems. This
particle size distribution, swell factor and the geometry of paper proposes a novel system-wide approach to optimize

123
A system-wide approach to minimize the operational cost of bench production in open-cast… 87

the mining cost by considering the relationship among bulk density of the rock (tonne/m3). When the burden is
interrelated subsystems. Since the costs of activities known, the spacing is found by Eq. (3) (Konya 1983).
forming mining cycle are expressed regarding fragmenta- 
0:125  H þ 0:875  B if H\4B
tion size to be obtained and blast-hole diameter, these two S¼ ð3Þ
1:4  B if H [ 4B
parameters are used as common ground in the calculations.
Once the optimum fragmentation size is obtained, the other where S is the spacing (m) and H is the bench height (m).
sections of a specific mining subsystem (such as the size of Moreover, sub-drilling J is equal to (Konya 1983):
the mining equipment, number of trucks, flow rate, etc.)
J ¼ 0:3  B ð4Þ
can be optimized individually. Moreover, the approach
includes a sensitivity analysis to quantify the relative effect Drilling cost is directly related to the total length of the
of the cost components. drilled hole. The total length of a hole is the bench height
The sum of optimal sub-systems does not create an (H) and sub-drilling (J). The total length of a hole is
optimal for the system. In this paper, the costs of activities multiplied with the number of drill holes (N) to find the
forming mining cycle are expressed regarding fragmenta- total drilled length. Given average drill rate (rD), (m/h) and
tion size to be obtained and blast-hole diameter. the hourly operation cost (cD) of drilling machine, the
drilling cost of the bench is calculated by the following
equation.
2 Methodology ðH þ J Þ  N
CD ¼  cD ð5Þ
rD
2.1 Cost analysis
where N is the number of drill holes in the blasting bench
Mining production system consists of drilling, blasting, which is based on the burden, spacing, bench width and
loading, hauling and crushing subsystems. The mean bench length distances.
fragmentation size is related to the cost of all subsystems.
To minimize the mining cost, the relation between the 2.1.2 Blasting cost
mean fragmentation size and the cost of each activity
should be analyzed. Explosive expenditure covers the main part of the blasting
The total mining cost of a bench (CM), consists of cost. Since the fixed costs of blasting (such as detonating
drilling, blasting, loading, hauling and crushing costs, and cords, boosters, detonators, labor, etc.) do not affect rock
is calculated by Eq. (1). fragmentation, they are ignored. The relation between the
CM ¼ CD þ CB þ CL þ CH þ CC ð1Þ rock fragmentation and the required explosive amount is
expressed by Eq. (6) (Kuznetsov 1973).
where CD is the drilling cost ($), CB is blasting cost ($), CL  0:8
V0
is the loading cost ($), CH is hauling cost ($), and CC is P50 ¼ A  Q1=6
e ð6Þ
crushing cost for crusher ($). Qe
where P50 is the mean fragment size (cm), A is the rock
2.1.1 Drilling cost factor, V0 is the volume of the rock per blast hole (m3), Qe
is the mass of TNT containing the energy equivalent of the
The drilling cost depends on bench configuration, the explosive charge in each blast hole (kg). A is 7 for medium
performance of drilling equipment, rock characteristics and hard rocks, 10 for hard but highly fissured rocks and 13 for
blasting pattern. Typical mining bench is illustrated in very hard, weakly fissured rocks. Alternatively, the mean
Fig. 2. In a bench blasting activity, the length, width, and fragmentation size can be estimated by using neural net-
height of the bench are known while designing a blast. works (Kulatilake et al. 2010), Monte Carlo simulation
Besides, other parameters such as burden, spacing, and modeling (Morin and Ficarazzo 2006) or multiple regres-
sub-drilling are decided based on the hole diameter. sion modeling (Faramarzi et al. 2013). The volume of the
The burden is calculated by the following equation rock is equal to:
(Konya 1983).
 0:333 V0 ¼ B  S  H ð7Þ
qe
B ¼ 0:03780  D  ð2Þ The explosive type used in the Eq. (6) is TNT; however, it
qr
can be reformulated depending on the explosive type as
where B is the burden (m), D is the hole diameter (mm), qe given in Eq. (8) (Gheibie et al. 2009).
is the bulk density of explosive (tonne/m3) and qr is the

123
88 B. Ozdemir, M. Kumral

Fig. 2 Illustration of a mining bench

 0:8  19=30 increases and the fill factor decreases. Decreasing amount
V0 Se
P50 ¼A Q1=6
e ð8Þ of the material in each load means a decrease in the loading
Qe 115
efficiency. On the other hand, if the rock is well frag-
where Se is the relative weight strength of the explosive mented, the fill factor and the loading efficiency will be
according to ANFO. The Eq. (8) can be rewritten for Qe as high. The relation between the fill factor and the mean
shown below. fragmentation size can be found by the regression analysis.
24=19 Given fill factor (F), the total loading time of the
115  A30=19  V0 material in the blasted bench is calculated by Eq. (11).
Qe ¼ 30=19
ð9Þ
P50  Se V0  N
TLT ¼  tl ð11Þ
Once the required amount of the explosive for each hole is VS  F
obtained, the blasting cost of the bench is calculated as where VS is the bucket capacity of the shovel (m3) and tl is
follows. the swing time for one load. Once total loading time is
CB ¼ Qe  pe  N ð10Þ obtained, the loading cost (CL) is found by Eq. (12).

where pe is the price of the explosive ($/kg). V0  N

CL ¼  tl  cL ð12Þ
VS  F
2.1.3 Loading cost where cL is the operating cost of the loader ($/h).

The cost of loading is related to the material movement in 2.1.4 Hauling cost
each load. Although the loader bucket has a certain volume
capacity, it is not possible to utilize all the capacity in a real Similar to the loading cost, the fill factor is one of the major
operation because of the space between the particles. There components of hauling cost. The relation between rock
is a fill factor of the bucket volume depending on particle fragmentation and the hauling cost is quantified by the fill
size distribution. The fill factor is simply the ratio of the factor. To maintain efficient loading operation, higher fill
volume of loaded material to the capacity of the bucket. If factor is preferable. The total time required for hauling the
the particle size increases, the space between the particles

123
A system-wide approach to minimize the operational cost of bench production in open-cast… 89

blasted material from the bench to its destination is cal- K2 is a crusher system constant which is taken 1.0 if
culated by Eq. (13). crusher is operated in a closed circuit with a classifying
V0  N screen and 1.19 if crusher is working in an open circuit.
THT ¼  tc ð13Þ Also, f is a function of particle size and calculated by the
VT  F
following equation (Morrell 2009).
where VT is the heaped capacity of the truck (m3) and tc is    
the cycle time of the trucks. When the total hauling time is f xj ¼  295 þ xj =1000000 ð16Þ
known, the hauling cost of material within a bench is In a bench blasting, the fragmentation size is not con-
obtained by Eq. (14). stant for all particles because of the heterogeneity of rock
CH ¼ THT  cT ð14Þ formation. The parameters of the distribution are obtained
by sieving analysis or image analysis. Typically, it shows a
where cT is the operating cost of the truck ($/h). cumulative distribution as given in Fig. 3. Mean frag-
mentation size (P50) is the sieve interval in which 50% of
2.1.5 Crushing cost the material passes. Similarly, P80 is the sieve interval
which 80% of the material passes and derived by the
As crushing is the last chain of interrelated subsystems Rosin–Rammler particle size distribution function which is
affected by mean fragmentation size, it should be included given below (Vesilind 1980).
in the cost analysis. Crusher cost varies with the particle n
size of the feed. The crushing cost is calculated based on Pð xÞ ¼ 1  elnð2Þðx=P50 Þ ð17Þ
specific energy. The energy requirement increases when where n is the material uniformity constant depending on
the particle size of the material increases. Specific energy blast geometry and rock mass. More information about the
for conventional crushers is obtained by Eq. (15) (Morrell fragment size distribution in a bench blasting can be found
2009). at (Cho et al. 2003).
 
f ðx Þ f ðx Þ The crushing cost is found by:
Wc ¼ K2  Mic  4  x2 2  x1 1 ð15Þ
CC ¼ Wc  u  V0  N  qr ð18Þ
where Wc is the specific energy (kWh/tonne), x1 is the 80%
passing size for the feed (lm), x2 is the 80% passing size where u is the energy price ($/kWh).
(P80) for the product of the crusher (lm), Mic is crushing
work index which is 7.2 kWh/t for conventional crushers.

Fig. 3 Typical cumulative particle size distribution curve

123
90 B. Ozdemir, M. Kumral

2.2 Optimization model The formulation of the optimization problem is defined

below.
The problem is formulated as the sum of sub-activity costs Objective function
per bench to be removed. The main controllable variable of Minimize CM ¼ CD þ CB þ CL þ CH þ CC ð19Þ
the optimization model is the mean fragmentation size
which can change from a blast to another by drill and blast where
design. The relationships among the interrelated mining    0:333 
activities lead to a bathtub curve relation between the total H þ 0:3  0:03780  D  qqe N
r
mining cost and rock fragmentation as seen in Fig. 4. The CD ¼  cD
rD
shaded area in this figure represents the minimum total
ð20Þ
mining cost and corresponding optimum mean fragmenta-
tion size. Finding this area is formulated as an optimization
problem.

  0:333 
24=19
30=19 qe
115  A  0:03780  D  q  ð0:125  H þ 0:875  BÞ  H
r
CB ¼  pe  N ð21Þ
ðP50 Þ30=19  Se
  0:333 

qe
0:03780  D  qr  ð0:125  H þ 0:875  BÞ  H  N
CL ¼  t l  cL ð22Þ
VS  F ðP50 Þ
  0:333 

qe
0:03780  D  q  ð0:125  H þ 0:875  BÞ  H  N
r
CH ¼  tc  cT ð23Þ
VT  F ðP50 Þ
0 0  n 1 1
 lnð2Þ 80
P50
B  n !@ 295þ1e 1;000;000 AC
B 80 C
B  lnð2Þ C
CC ¼ K2  Mic  4  B75; 000295:075Þ 
P50
1e C  u  V 0  N  qr ð24Þ
B C
@ A

Subject to
2.2.1 Problem formulation
(1) The hole diameter should be equal to one of the drill
bit size specified in the industrial standards (Sandvik
The total mining cost is minimized by changing two con-
2015) as given in Eq. (25).
trollable parameters which are the blasthole diameter and
mean fragmentation size. The problem is a non-linear D 2 f159; 171; 187; 200; 216; 229; 251; 270; 311;
problem and hard to solve through an exact method. 349; 381; 406g
Therefore, a meta-heuristics, evolutionary algorithm (EA), ð25Þ
is selected to solve the model. (Sauvageau and Kumral
2016). A metaheuristic like the EA has two drawbacks: (1) (2) The weight of the load should not pass the payload
it does not guarantee the optimality, and (2) the EA- capacity for both shovel and load as given in
specific parameters (e.g., crossover and mutation rates) are Eqs. (26) and (27).
selected arbitrarily. However, if the solution space is VS  F ðP50 Þ  qr  TS ð26Þ
searched sufficiently, the solution can approach the opti-
mality. Regarding the parameter selection, employing dif- VT  F ðP50 Þ  qr  TT ð27Þ
ferent parameter configurations can also reduce the (3) The maximum particle size should be smaller than
uncertainty associated with the EA’s parameters. the maximum allowable feed size of crusher as given
in Eq. (28).

123
A system-wide approach to minimize the operational cost of bench production in open-cast… 91

simulation-based costing, particle swarm optimization, an

artificial neural network, etc. Nolan and Kecojevic (2014)
used analytical hierarchy process to improve current min-
ing practices by selecting the optimal mining method based
on cost, production and environmental criteria.

3 Results and discussion

3.1 Mine model

A case study was conducted in an open-cast coal mine.

Bench geometry and rock characteristics parameters of the
Fig. 4 The relationship between total mining cost and the rock mine are given in Table 1.
fragmentation Regression analysis of the field data unveiled the rela-
tion between the fill factor and mean fragmentation size as
PMAX  R ð28Þ given in Eq. (29).
F ðP50 Þ ¼ 0:9833  0:0022  P50 ð29Þ
The model was solved by Excel Solver. The workflow of
the evolutionary algorithm used in this research is given in The R2 and standard error of the regression are 89.2% and
Fig. 5. The input parameters are initially entered into the 0.032, respectively. This shows that the fill factor equation
system. The first population is generated, and the fitness of is representative for this specific mining operation. This
the population is evaluated. If the stopping criterion is met, information is used in the optimization model.
the optimization process is terminated and this solution is Table 2 denotes the characteristics of the mining
reported as the optimal solution. Otherwise, it continues to equipment. Since the effect of fragmentation size on the
search through creating a new population applying cross- costs is investigated, the costs related to the equipment’s
over and mutation operators. Then, the new population is
evaluated. This workflow continues until the stopping cri- Table 1 Bench geometry and rock characteristics
terion met. Parameter Value
The cost calculation can also be done by using other
Rock factor, A 10
methods such as mixed-integer non-linear algorithm,
Blast geometry constant, n 1.3
Explosive price ($/kg) 2.5
The bulk density of the explosive (tonne/m3) 0.82
Length of the bench (m) 150
The width of the bench (m) 50
The height of the bench (m) 10
The bulk density of the rock (tonne/m3) 2.7

Table 2 Characteristics of mining equipment

Characteristic Value

Drill rate (m/h) 25

Shovel bucket capacity (m3) 27.5
Shovel bucket capacity (tonne) 90
Loading time (minutes) 0.5
Truck capacity (m3) 130
Truck capacity (tonne) 320
Truck cycle time (min) 15
Fig. 5 Flow chart of the proposed evolutionary algorithm

123
92 B. Ozdemir, M. Kumral

Table 3 The operation cost of mining equipment (CostMine 2016) Table 5 Optimization results
Equipment Operation cost ($/h) Parameter Optimum value

Drilling machine (cD) 99.69 Mean fragmentation size, P50 (cm) 68

Loader (cL) 360.63 Hole diameter (mm) 200
Haul truck (cT) 498.79 P80 (cm) 130
Amount of required explosive (kg/hole) 60
Total mining cost ($) 131,670
Blasting cost ($) 35,297
Table 4 Parameters of the crushing operation Drilling cost ($) 9391
Parameter Value Loading cost ($) 8649
Hauling cost ($) 75,941
80% passing size of crusher feed, x1 (lm) 1,389,406
Crushing cost ($) 6907
80% passing size of crusher product, x2(lm) 75,000
Crushing work index, Mic 7.2
Crusher constant, K2 1.19
Energy price, u ($/kWh) 0.072 hole diameter should be 68 cm and 200 mm, respectively.
The other results of the optimization are given in Table 5.

3.3 Sensitivity analysis

part wear and other fixed costs were ignored. Shovel bucket
and truck payload capacity were considered as constraints The previous sections describe how to calculate the cost of
in the optimization model. If bulk density of the loaded the activities and optimization of total cost in a mining
material is too high, it is not possible to utilize all volume operation. However, some of these costs may change
of the shovel or truck even though particle size is small. In during the operation depending on equipment reliability,
that case, the particle size could be large. The optimum human performance, and seasonal effect, etc. For example,
fragmentation size and hole diameter are decided by using the explosive and electricity price, the hourly unit cost of
abovementioned methodology. drilling machine, loader and truck may vary in different
The operations costs for the mining equipment were mine or change during the operation. To quantify the effect
obtained from CostMine (2016), and they are given in of variation of these costs on the total mining cost, sensi-
Table 3. These costs are likely to change depending on tivity analysis is conducted. Sensitivity analysis is a tech-
operational conditions. They should be updated during the nique to quantify the relative importance of different
operation for more accurate results. independent parameters on a dependent output variable
The parameters of crushing operation and crushing cost (Saltelli et al. 2000). In the mining industry, it was used to
calculation are given in Table 4. 80% passing size of the understand the effect of road length, diesel price and
crusher feed is calculated from Eq. (17) and energy price is electrical price on the capital and operational expenditures
taken from CostMine (2016). (Dilhuydy et al. 2017). Moreover, de Werk et al. (2017)
applied sensitivity analysis to evaluate the parameters of
3.2 Cost optimization two different material haulage methods as in-pit crushing
and conveying systems and traditional truck-shovel system.
Once the mine model is formed and all the parameters are Furthermore, a sensitivity analysis was used to quantify the
connected to each other, evolutionary algorithm was impact of the purchase price, operating and maintenance
implemented by Excel Solver. Before starting to opti- cost of the drilling machine (Al-Chalabi et al. 2015).
mization, mean fragmentation size and the hole diameter The results of sensitivity analysis are summarized in
are taken as 50 cm and 159 mm, respectively. The total Fig. 6. As can be seen from the figure, the mining cost
mining cost is calculated $158,311 for the bench with the highly depends on the unit cost of trucks. The reason
initial values. The solver engine divided the problem to behind this is that the truck usage time is longer than the
13,341 subproblems and reached to the minimum total other equipment because of the longer cycle time. A 10%
mining cost in 43.9 s by changing these two variables increase in the unit cost of trucks leads to $7382 increase in
while considering the constraints. The minimum total total mining cost. The explosive price also has a consid-
mining cost to extract given bench is obtained as $131,670. erable effect on the total mining cost. The total mining cost
So, the total mining cost is reduced by 16.8%. To achieve increases $2736 if explosive price increases 10%. More-
this cost minimization, the mean fragmentation size and the over, electricity price affects the crushing cost. If electricity

123
A system-wide approach to minimize the operational cost of bench production in open-cast… 93

Fig. 6 Sensitivity analysis

price change 15%, the total cost change about $1036. On The optimization is achieved by considering the entire
the other hand, the unit cost of drilling machine and loader mining system instead of focusing on a single subsystem.
does not have a significant impact on the total mining cost. By this way, the relationship among the interrelated sub-
If the unit cost of drilling machine and loader change 10%, systems in open-cast mining operations is considered in the
the total mining cost changes $463 and $841, respectively. optimization model. The methodology was applied in an
open-cast coal mine. The evolutionary algorithm was used
in the optimization of the fragmentation size and hole
4 Conclusions diameter. The results of the case study show that the total
mining cost is reduced $26,641 for the bench which cor-
This paper presents a system-wide cost optimization responds a 16.8% decrease for one mining bench. If the
approach to minimize the operational cost of open-cast usage of this methodology is extended to the other blasting
mining operations by analyzing the relationship among the benches throughout the mine, the total gain increases.
mining subsystems including drilling, blasting, loading, Furthermore, the results of the sensitivity analysis show
hauling and crushing. The particle size of the blasted muck that the total mining cost highly depends on the unit cost of
pile is the common cost parameter for all activities. Thus, haul trucks, explosive cost and electricity price. The unit
the cost calculations are based on the mean fragmentation cost of loaders and drilling machines does not have a sig-
size which is decided by the blast design parameters such nificant effect on the total mining cost.
as burden, spacing, sub-drill, the explosive charge, etc. The proposed approach optimizes the cost of a bench
Lowering the particle size causes an increase in the drilling production based on given geometric parameters which are
and blasting costs. On the other hand, it provides more defined by considering technical feasibility and safety. In
efficient loading, hauling and crushing till a certain size. the future, the research will be extended such that more
Mean particle size affects the loading and hauling by the specific constraints (e.g., dust, noise, geotechnical and
fill factor. Moreover, the energy requirement of the equipment restrictions) are incorporated into the model.
crushers is directly related to the crusher feed size. When Also, the uncertainties associated with the parameter values
these relations between the mean fragmentation size and will be considered to assess the risk on the decision-making
the cost of mining activities are considered, optimizing the process.
mean fragmentation size can provide substantial cost sav-
ings in mining operations. Acknowledgements The authors thank Natural Sciences and Engi-
neering Research Council of Canada (NSERC) for funding this
research (ID: NSERC 236482).

123
94 B. Ozdemir, M. Kumral

Open Access This article is distributed under the terms of the drilling, blasting and excavating operations. Int J Min Reclam
Creative Commons Attribution 4.0 International License (http://crea Environ 31(7):457–474. https://doi.org/10.1080/17480930.2016.
tivecommons.org/licenses/by/4.0/), which permits unrestricted use, 1160818
distribution, and reproduction in any medium, provided you give Kulatilake P, Qiong W, Hudaverdi T, Kuzu C (2010) Mean particle
appropriate credit to the original author(s) and the source, provide a size prediction in rock blast fragmentation using neural net-
link to the Creative Commons license, and indicate if changes were works. Eng Geol 114(3):298–311
made. Kuznetsov V (1973) The mean diameter of the fragments formed by
blasting rock. J Min Sci 9(2):144–148
Martin PL (2006) Drill and blast optimization at the Sparkhule
Limestone Quarry. J Explos Eng 23(4):6–12
References
Monjezi M, Khoshalan HA, Varjani AY (2011) Optimization of open
pit blast parameters using genetic algorithm. Int J Rock Mech
Al-Chalabi H, Lundberg J, Ahmadi A, Jonsson A (2015) Case study: Min Sci 48(5):864–869
model for economic lifetime of drilling machines in the Swedish Morin MA, Ficarazzo F (2006) Monte Carlo simulation as a tool to
mining industry. Eng Econ 60(2):138–154 predict blasting fragmentation based on the Kuz–Ram model.
AminShokravi A, Eskandar H, Derakhsh AM, Rad HN, Ghanadi A Comput Geosci 32(3):352–359. https://doi.org/10.1016/j.cageo.
(2017) The potential application of particle swarm optimization 2005.06.022
algorithm for forecasting the air-overpressure induced by mine Morrell S (2009) Predicting the overall specific energy requirement of
blasting. Eng Comput 1:1–9. https://doi.org/10.1007/s00366- crushing, high pressure grinding roll and tumbling mill circuits.
017-0539-5 Miner Eng 22(6):544–549
Bogunovic D, Kecojevic V (2011) Impact of bucket fill factor. Min Neale AM (2010) Blast optimization at Kriel Colliery. J South Afr
Eng 63(8):48–53 Inst Min Metall 110(4):161–168
Cho SH, Nishi M, Yamamoto M, Kaneko K (2003) Fragment size Nielsen K (1983) Optimisation of open pit bench blasting. Paper
distribution in blasting. Mater Trans 44(5):951–956 presented at the 1st international symposium on rock fragmen-
Chugh YP, Behum PT (2014) Coal waste management practices in tation by blasting, Lulea, Sweden
the USA: an overview. Int J Coal Sci Technol 1(2):163–176. Nolan TA, Kecojevic V (2014) Selection of overburden surface
https://doi.org/10.1007/s40789-014-0023-4 mining method in West Virginia by analytical hierarchy process.
CostMine (2016) Mine and mill equipment costs. InfoMine USA, Int J Coal Sci Technol 1(3):306–314. https://doi.org/10.1007/
Inc., Spokane Valley s40789-014-0019-0
de Werk M, Ozdemir B, Ragoub B, Dunbrack T, Kumral M (2017) Osanloo M, Hekmat A (2005) Prediction of shovel productivity in the
Cost analysis of material handling systems in open pit mining: Gol-e-Gohar iron mine. J Min Sci 41(2):177–184
case study on an iron ore prefeasibility study. Eng Econ Ouchterlony F, Sanchidrián JA, Moser P (2017) Percentile fragment
62(4):1–18 size predictions for blasted rock and the fragmentation-energy
Dickerson AW, Alexander RB, Hollis AJ (1986) Measurement of fan. Rock Mech Rock Eng 50(4):751–779
payloads carried by mine haul trucks and the influence of Pothina R, Kecojevic V, Klima MS, Komljenovic D (2007) Gyratory
payloads on production rates and material movement costs. crusher model and impact parameters related to energy con-
Paper presented at the Symposia Series—Australasian Institute sumption. Miner Metall Process 24(3):170–180
of Mining and Metallurgy Saltelli A, Chan K, Scott EM (2000) Sensitivity analysis, vol 1.
Dilhuydy E, Ozdemir B, Kumral M (2017) Economic analysis of Wiley, New York
waste crushing in semi-mobile in-pit crushing and conveying Sandvik (2015) Rotary drilling bits and drill string tools
systems. CIM J. https://doi.org/10.15834/cimj.2017.14 Sauvageau M, Kumral M (2016) Genetic algorithms for the optimi-
Faramarzi F, Mansouri H, Ebrahimi Farsangi MA (2013) A rock sation of the Schwartz-Smith two-factor model: a case study on a
engineering systems based model to predict rock fragmentation copper deposit. Int J Min Reclam Environ. https://doi.org/10.
by blasting. Int J Rock Mech Min Sci 60(Supplement C):82–94. 1080/17480930.2016.1260858
https://doi.org/10.1016/j.ijrmms.2012.12.045 Singh S, Narendrula R (2006) Factors affecting the productivity of
Fisonga M, Garcı́a YD, Besa B (2017) Burden estimation using loaders in surface mines. Int J Min Reclam Environ
relative bulk strength of explosive substances. Appl Earth Sci 20(01):20–32
126(1):31–37. https://doi.org/10.1080/03717453.2017.1296673 Singh P, Roy M, Paswan R, Sarim M, Kumar S, Jha RR (2016) Rock
Gheibie S, Aghababaei H, Hoseinie S, Pourrahimian Y (2009) fragmentation control in opencast blasting. J Rock Mech
Modified Kuz—Ram fragmentation model and its use at the Geotech Eng 8(2):225–237
Sungun Copper Mine. Int J Rock Mech Min Sci 46(6):967–973 Taherkhani H, Doostmohammadi R (2015) Transportation costs: a
Kahriman A, Ceylanoalu A (1996) Blast design and optimization tool for evaluating the effect of rock mass mechanical param-
studies for a celestite open-pit mine in Turkey. Mineral Resour eters on blasting results in open pit mining. J Min Sci
Eng 5(2):93–106. https://doi.org/10.1142/S095060989600008X 51(4):730–742
Kecojevic V, Komljenovic D (2007) Impact of burden and spacing on Vesilind PA (1980) The Rosin-Rammler particle size distribution.
fragment size distribution and total cost in quarry mining. Trans Resour Recov Conserv 5(3):275–277
Soc Min Metall Explor 320:133 Workman L, Eloranta J (2003) The effects of blasting on crushing and
Konya CJ (1983) Blast design. In: Sendlein LVA, Yazicigil H, grinding efficiency and energy consumption. In: Proc 29th Con
Carlson CL (eds) Surface mining environmental monitoring and Explosives and Blasting Techniques, Int Society of Explosive
reclamation handbook. Elsevier, New York Engineers, Cleveland OH, pp 1–5
Kozan E, Liu SQ (2017) An operational-level multi-stage mine
production timetabling model for optimally synchronising

123