18th IFAC Symposium on Control, Optimization and Automation in

18th IFAC Symposium on Control, Optimization and Automation in

Mining,
18th Mineral
IFAC and Metal
Symposium Processing
on Control, Optimization and Automation in
Mining, Mineral and Metal Processing Available online at www.sciencedirect.com
Stellenbosch,
Mining,
18th South
Mineral
IFAC Africa,
and Metal
Symposium AugustOptimization
28-30, 2019 and Automation in
Processing
on Control,
Stellenbosch, South Africa, August 28-30, 2019
Stellenbosch,
Mining, MineralSouth Africa,
and Metal August 28-30, 2019
Processing
Stellenbosch, South Africa, August 28-30, 2019
ScienceDirect
IFAC PapersOnLine 52-14 (2019) 207–212
An
An Optimization Approach for Shovel
An Optimization
Optimization Approach
Approach for for Shovel
Shovel
Allocation
Allocation to
to Minimize
An Optimization
Minimize Fuel
Approach
Fuel Consumption
for Shovel
Consumption
Allocation to Minimize Fuel Consumption
in Open-pittoMines:
Allocation
in CaseFuel
Minimize of Heterogeneous
Consumption
in Open-pit
Open-pit Mines:
Mines: Case
Case of of Heterogeneous
Heterogeneous
Fleet
in Open-pit Mines:
Fleet of Shovels.
Case of Heterogeneous
Fleet of
of Shovels.
Shovels.
Fleet of Shovels.
D.M. Bajany L. Zhang X. Xia
D.M. Bajany L. Zhang X. Xia
D.M. Bajany L. Zhang X. Xia
∗
Department D.M. Bajany L. Zhang Computer
X. Xia
∗
∗
∗ Department of Electrical,
of Electrical, Electronic
Electronic and
and Computer Engineering,
Engineering,
Department of Electrical,
University Electronic
of Pretoria, Pretoriaand Computer
0002, Engineering,
South Africa
∗ University
Department of Pretoria,
of Electrical,
University Pretoria
Electronic
of Pretoria, 0002,
Pretoriaand South
Computer
0002, Africa
Engineering,
South Africa
University of Pretoria, Pretoria 0002, South Africa
Abstract:
Abstract:
Abstract:
This paper presents a realistic model that can be used to minimize fuel consumption in
This
Abstract:
This paper
paper presents
presents a
a realistic
realistic model
model that that can be
be used
candispatches to
to minimize
used trucks minimize fuel
fuel consumption
consumption in
in
surface
surface mining
mining operations.
operations. The
The developed
developed model
model dispatches trucks to
to destinations
destinations by
by optimally
optimally
This
surface
determinepaper thepresents
mining operations.
paths a realistic
that The
can lead model
to the that
developed model
minimum candispatches
be
fuelused to minimize
trucks
consumption to per fuel and
destinations
truck consumption
by optimally
shovel cyclein
determine
surface the
the paths
mining that
that can
operations. The lead to
developedthe
the minimum
model fuel consumption to per truck and shovel cycle
determine
while the paths
handling demands can oflead
each todumping minimumsitedispatches
fuel
are met. trucks
consumption
This model destinations
per
is truck andby
applicable tooptimally
shovel cycle
all types
while
determine
while the
the handling
the paths
handling demands
that canof
demands of each
lead
each todumping
the
dumpingminimumsite
site are met.
fuel
are This
consumption
met. This model
model is
per applicable
is truck andinto
applicable to
to all
shovel
all types
cycle
types
of open-pit mines, including over
over and
and under-trucked
under-trucked mines.
mines. The
The formulation
formulation takes
takes into account
account
of
while
of open-pit
open-pit mines,
the handling
mines, including
demands
including of
overeachanddumping site aremines.
under-trucked met. This
The model is applicable
formulation takes to all
into types
account
the waiting
theopen-pit time
waiting time of of trucks
trucks at at dumping
dumping and loading
and loadingmines. sites and
sites and the
the idle times
idle times of shovels.
of shovels. The
The
of
the waiting mines,
effectiveness time
of theof including
trucks at
proposed over
modelandinunder-trucked
dumping anddetermination
the loading sites of The
and formulation
thethe
best idleshovel
times takes into account
of shovels.
allocation in The
the
effectiveness
the of the proposed model in the determination of the best shovel allocation in the
case waiting
effectiveness time
of theof proposed
of a heterogeneous trucksfleetatofmodel
dumping
shovels and
in isthe loading
determination
also sites of
demonstrated and
inthethe
best
this idleshovel
study. timesallocation
of shovels.in The the
case
case of of aa heterogeneous
effectiveness of the proposed
heterogeneous fleet
fleet of shovels
ofmodel
shovels in is
isthealso demonstrated
demonstratedofin
determination
also this
inthe study.
thisbest shovel allocation in the
study.
©
case 2019,
of aIFAC (International
heterogeneous Federation
fleet of shovelsof Automatic
is also Control)
demonstrated Hostingin by
thisElsevier
study. Ltd. All rights reserved.
Keywords:
Keywords:
Keywords:
Truck-shovel dispatch, fuel minimisation, shovel allocation
Truck-shovel
Truck-shovel dispatch,
Keywords: dispatch, fuel fuel minimisation,
minimisation, shovel shovel allocation
allocation
Truck-shovel dispatch, fuel minimisation, shovel allocation
1. INTRODUCTION dedicated to under-trucked mines as it does not consider
1.
1. INTRODUCTION
INTRODUCTION dedicated
dedicated to
to under-trucked
under-trucked mines as as it
it does
does notnot consider
consider
in its formulation the waitingmines time of trucks at loading and
in its formulation the waiting time of trucks at loading and
Shovels and dump1. trucks INTRODUCTION
are commonly used in open- in dedicated to under-trucked
its formulation
dumping
dumping
sites,
sites,
the waiting
which situation mines
time
canof as it doesat
trucks
happen innot consider
loading
practice. and
Shovels
Shovels
pit minesand
and dump
fordump
trucks
trucks and
ore loading
are commonly
commonly used
are transportation in
in open-
usedbecause open- of dumping sites, which
in its formulation which situation
the waiting
situation can
canofhappen
time trucks at
happen in practice.
in loading
practice.and
pit mines for ore loading and transportation because of The contribution
dumping sites, whichof this study can
situation is the development
happen in of an
practice.
Shovels and dump trucks are commonly used in open- The contribution of this study is the development of
pit
their
mines for
flexibility
ore
and
loading
ability
and
to
transportation
their flexibility and ability to transport material over long The
transport material
because
over
of
long optimalcontribution
truck-shovel of this studystrategy
dispatch is the development
for existing mines of an
an
pit
theirmines
distances for ore
flexibility
(Bajany andetloading
ability
al. and transportation
to transport
(2017b)). It has material
been because
shown overin of optimal
long
the The truck-shovel
contribution dispatch
of this studystrategy
isboth for existing
the development mines
of an
distances (Bajany et al. (2017b)). It has been shown in the optimal
to minimize
to minimize truck-shovel
fuel dispatch
consumption
fuel consumption strategy
of
of both trucksfor
trucks existing
and mines
shovels
and shovels
their flexibility
distances
literature that and
(Bajany etability
excavating to
thetransport
al. (2017b)). It has material
material been
and shown overinlong
transporting the to optimal truck-shovel dispatch strategy for Moreover,
existing mines
literature that excavating the material and transporting minimize
considering fuel consumption
their technical of both trucks
specifications. and shovelsthis
distances
literature
excavated that (Bajany
material et al.
excavating (2017b)).
represent It
the amaterial has
large part been shown
andoftransporting in the
the opera- to considering
minimize
considering their
fuel
their technical
consumption
technical specifications.
of both
specifications. trucksMoreover,
and
Moreover, this
shovels
this
excavated material represent a large part of the opera- model can also be used by decision-makers as a planning
literature
tional coststhat excavating the(Daamaterial and oftransportinget al. model can also be used by decision-makers as
excavated
tional costs
material
in surface
in surface
represent
mining
mining (Da
large
Cunha
Cunha
partRodovalho
Rodovalho
the opera-
et al. considering
model
tool
tool to
to can their
also
identify
identify be technical
the
the usedbest
best
specifications.
byallocation
decision-makers
allocation of
of as a
Moreover,
shovels
shovels
planning
a that
that
this
planning
will
will
excavated
tional
(2016)). costs
Thematerial
in surface
haulage represent
mining
cost can a be
(Da large
Cunha part
approximately of the 50%
Rodovalho opera-
et al.
of model can also be
the usedbest byallocation
decision-makersof shovelsasthea that
planning
will
(2016)). The haulage cost can be approximately 50% of tool
result
result toin
in identify
the
the minimum
minimum fuel
fuel consumption
consumption for
for the haulage
haulage
tional
(2016)). costs
the entireThe in surface
haulage cost
operational mining
cost or can(Da
60% Cunha Rodovalho
be ofapproximately
the total operational et al.
50% of result tool toinidentify
theduring thethe
minimum bestfuelallocation
consumption of shovels that will
the entireThe operational cost operations planning phase offora the
new haulage
open-
(2016)).
the
costentire
in large haulage
operational
surface cost or
cost
minings can
or 60%
60% be of
(Bajany
the
the
et al.
total
ofapproximately
total operational
50%
operational
(2017a); Alarie of operations
result in
operations the during
minimum
during the
the planning
fuel phase of
consumption
planning phase offoraa the
newhaulage
new open-
open-
cost in large surface minings pit mine or during the re-allocation planning of shovels
the
and entire operational cost or(Bajany
60%
and of
et
Xiathe
al. (2017a);
total Alarie pit
operational mine
cost
and
in large
Gamache
Gamache
surface
(2002);
(2002);
minings
Zhang
Zhang
(Bajany
and Xia
et(2015)).
al. (2017a);
(2015)).
TheAlarie
The
large pit
large in
in mine or
operations
existing
existing
during
during
ormines.
duringIn
mines. In
the
the
the re-allocation
planning
re-allocation
addition,
addition, the
the
phase planning
of a new
planning
developed
developed
of
of shovels
model
model
open-
shovels
can
can
cost
and in
contributorlarge to
Gamache surface
(2002);
the minings
Zhang
transportation (Bajany
and Xia
cost etin al.open-pit
(2017a);
(2015)). The Alarie
large
mining pit mine ormines.
during the re-allocation planning of shovels
contributor to the transportation cost in open-pit mining in existing
assist
assist project
project In
managers
managers addition,
to
to the
estimate
estimate developed
the
the model
expected
expected can
fuel
fuel
and Gamache
contributor
operations istothe (2002);
thefuel Zhang
transportation and
consumptioncost Xia
of the(2015)).
in open-pit The
haul trucks large
mining and assist
in existing mines. In addition, the production
developed model can
operations is project
consumption managers
per tonne fortoa estimate
given the expected fuel
target and
contributor
operations
shovels. Sahoo istothe
the
et
fuel
theal.
fuel consumption
transportation
pointedcost
consumption
(2014)
of
of the
out the
that
haul
in open-pit
haul trucks
thistrucks mining
represents
and
and consumption
assist project
consumption per tonne
managers
per tonne for
fortoaa estimate
given
given production
the
production
justify the budget allocated to the haulage operations.
target
expected
target and
fuel
and
shovels.
operations Sahoo
is et
the al.
fuel(2014) pointed
consumption out
of thethat
thathaulthis represents
thistrucks and justify the
shovels.
32%
32% of
of Sahoo
the
the total
totalet energy
al. (2014)
energy input
inputpointed
in
in out
mines.
mines. represents consumption
justify
The
The model the budget
model
per tonne
budget
presented
presented
allocated
for this
allocated
in
in this
to
a given the
to work
work
haulage
theproduction
haulage
is
is build
build
operations.
target and
operations.
based
based on on
shovels.
32% of theSahoototalet energy
al. (2014)inputpointed out that this represents justify
in mines. the budget allocated to the haulage operations.
In most of the surface mining, diesel The model
m-trucks-for-npresented in
strategy this
is the main fuel used the m-trucks-for-n strategy and is capable of solving the
the and work
is is
capable build
of based
solving on
the
32%
In of
most the
of total
the energy
surface input
mining, in mines.
In
by
by most
trucks
trucks ofand
theshovels.
and mining, diesel
surface However,
shovels. However, diesel
diesel
diesel
is the
the main
isengines
engines mainused
used
fuel
fuel
in
in
used
used
aa lot
lot
The
the model problem
m-trucks-for-n
dispatching
dispatching
presented
problem
for in
strategy
for
this
and work
heterogeneous
heterogeneous
is build
is capable
fleets
fleets
of
of
based the
solving
of equipment,
equipment,
on
In
by
of most
trucksofand theshovels.
surface However,
mining,impact diesel
dieselison the
engines mainused fuel
in used the
a lot includingm-trucks-for-n
dispatching problem strategy
for and
heterogeneous is capable
including trucks and shovels. This model is applicable to fleets of
of solving
equipment, the
of equipment
equipment have
have aa noxious
noxious impact on the
the environment
environment dispatching trucks
problemand
andfor shovels. This
heterogeneous model
fleetsis applicable
is of equipment, to
by
of trucksofand
equipment
because theirshovels.
have However,
a noxious
greenhouse gas diesel
impact
emissions engines
on the used inet
environment
(Rahman al. including
a lot under
under and
and trucks
over-trucked
over-trucked shovels.
mines
mines This
as
as the
the model
waiting
waiting applicable
time
time at
at both
bothto
because
of equipment
because of
of their
theirhavegreenhouse
a noxious
greenhouse gas
gas emissions
impact
emissions on (Rahman
the environment
(Rahman et
et al.
al. including
under
loading andandtrucks and
over-trucked
unloading shovels.
mines
sites This
areas the
takenmodel
waiting
into is applicable
time
account at both
in to
its
(2013); Bagheri et al. (2015)). Lashgari et al. (2014) has loading and unloading sites are taken into account in its
(2013);
because
(2013); Bagheri
ofintheir
Bagheri et al. (2015)).
greenhouse
et al. (2015)). gasLashgari
emissions
Lashgari et
et al.
(Rahman
al. (2014)
(2014) ethas
has under
al. loading and
formulation.andover-trucked
unloading sites minesare as taken
the waiting time at in
into account both
its
indicated
indicated their research
in their research conducted
conducted on a
on a coalcoal open-mine
open-mine formulation.
(2013);
indicated
in West Bagheri
in theiretthat
Virgina al. 44.7%
(2015)).
research conducted
of Lashgari
the on aetnitrogen
total al. (2014)
coal open-mine
oxides loading and unloading sites are taken into account in its
has formulation.
in West Virgina that 44.7% of the total nitrogen oxides In this study, we use empirical values of waiting time in
indicated
in in
Westemission
(NOx) their
Virginaisthat research conducted
44.7%inofthe
produced the on a coal
total nitrogen
haulage oxides formulation.
open-mine
transportation. In
In this study,
thethis study, we
simulations we use empirical
to use
show empirical values
values of
the effectiveness of ofwaiting
waiting time
time in
the developed in
(NOx)
(NOx) emission
in Westemission is
Virginaisthat produced
produced in
44.7%inofthe thethe haulage
haulage transportation.
total nitrogen oxides the
transportation. simulations to show the effectiveness of the developed
Therefore, reducing the fuelinconsumptions In
the this
dispatch study,
simulations
model weto
in use
show empirical
the
optimizing
of the haulage dispatch model in optimizing the route choices of dump values
effectiveness
the route of waiting
of the
choices time
developed
of dump in
(NOx)
Therefore, emission
reducingis produced
the fuel the haulage transportation.
consumptions of the haulage the simulations to show the effectiveness
the route of the
choices developed
of dump
Therefore, reducing athe
operation provides goodfuelopportunity
consumptions to of the haulage
reduce dispatch
trucks and model in
minimizing optimizing
the fuel consumption
opera- trucks and minimizing the fuel consumption of both trucks of both trucks
operation provides a goodfuelopportunity to reduce opera- dispatch andmodel in optimizing the
timeroute
fuel consumption choices
of of trucks
both dump
Therefore,
operation reducing
tional costsprovides athe
good
and environmental consumptions
opportunity
impacts of of
toopen
reduce mines trucks
thepithaulage
opera- and shovels.
and shovels.
minimizing
However,
However,
the
waiting
waiting time
values
values
can
can
be
be
deter-
deter-
tional
operation
tional costs
costs and
provides environmental
a good opportunity
and environmental impacts of open
toInopen
reduce pit mines
opera- trucks
and and
shovels. minimizing
However, the fuel
waiting consumption
time
mined approximately from the database history of the values of
can both
be trucks
deter-
(Jochens
(Jochens
tional
(1980);
(1980);
costs
Abdelaziz
Abdelaziz et
and environmental al.impacts
et al. (2011)).of
(2011)). the pit
the
Inopen mines mined
literature
literature
mines mined and approximately
shovels.
haulage However,
approximately
operations using
from
waiting
from the
thetime
queuing
database
values(May
database
theories
history
can be
history of the
deter-
of
(2013)).the
(Jochens
optimum (1980);
truck Abdelazizmodels
dispatch et al.impacts
(2011)).
do exist, ofIn onlypit
the literature
Bajany haulage operations using queuing theories (May (2013)).
optimum
(Jochens
optimum truck
(1980);
truckhas dispatch
Abdelaziz
dispatch models
et
models al. do
(2011)). exist,
do exist, In only
the Bajany
literature
only Bajany mined
haulage approximately
operations usingfrom the
queuing database
theories history
(May of
(2013)).the
et al. (2017b) developed an optimal dispatch model
et al.
optimum
et al. (2017b)
(2017b)truck has developed
dispatch
has developed models an
an optimal
do
optimal exist, dispatch
only
dispatch model
Bajany
model haulage operations using queuing theories (May (2013)).
that minimizes simultaneously the fuel consumption of
that
et
thatal. minimizes
(2017b) simultaneously
has developed anthe optimalfuel
fuel consumption
dispatch model of
trucksminimizes
trucks
that
and
and shovels.
minimizes
simultaneously
shovels. However,
However, the
simultaneously thethe
the
model
model
fuel
consumption
was
was specifically
specifically
consumption
of
of
trucks and shovels. However, the model was specifically
trucks and shovels. However, the model was specifically
2405-8963 © 2019, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved.
Copyright © 2019 IFAC 217 217
Copyright
Peer review© under
2019 IFAC 217 Control.
responsibility of International Federation of Automatic 217
Copyright © 2019 IFAC 217 217
10.1016/j.ifacol.2019.09.196
Copyright © 2019 IFAC 217 217
2019 IFAC MMM
208
Stellenbosch, D.M. Bajany et al. / IFAC PapersOnLine 52-14 (2019) 207–212
South Africa, August 28-30, 2019

2. FUEL CONSUMPTION OF TRUCKS AND Unloading points i

SHOVELS i=1 i=2

Fuel consumption and emissions of diesel trucks and

shovels depend on a variety of factors such as the effi-
ciency of the diesel engine, the operating conditions of the
trucks/shovels, and drivers’ behavior, ignition timing, to J=1 J=2 J=3
name a few (Suzdaleva and Nagy (2014)). For estimation 1600 t/h 1600 t/h 2000 t/h
purposes as suggested in Peralta et al. (2016), the rate of Loading points j
(a)
fuel consumption by trucks can be determined by equation Unloading points i
(1). i=1 i=2

f y = 0.3 × P y × LF, (1)

where P y , 0.3 and LF represent the maximal engine power
(kW), the unit conversion factor (L/kW/h) and the engine
load factor respectively. The load factor is the portion of
the full power required by the truck. Common load factors J=1 J=2 J=3
for dump trucks are given in the manufacturer’s catalogue. 1600 t/h 2000 t/h 1600 t/h
In this study, to determine the fuel consumption of dump Loading points j
(b)
trucks, the following values of load factors are considered: Unloading points i
45% for a loaded truck (high load), 20% for an empty truck i=1 i=2

and 10% for an idling truck (Kecojevic and Komljenovic

(2010)).
Similarly, the fuel consumption of shovels can be estimated
knowing their operating conditions. Indeed, the engine fuel
is controlled by the engine load factor, which depends on J=1 J=2 J=3
the application of the shovel. In this study, the hourly fuel 2000 t/h 1600 t/h 1600 t/h
consumption rate fj of shovels during working times will Loading points j
(c)
be estimated by assuming that shovels are working with
high load factors. The hourly fuel consumption fj,idle of
Fig. 1. (a) j3 = shovel 2000 t/h, (b) j2 = shovel 2000 t/h;
shovels during idle times will be assumed to be 10% of the
(c) j1 = shovel 2000 t/h
hourly fuel consumption of those shovels when they are
working with low load factors. where S is the number of shovels, SC1 , SC2 ...SCj=k are the
numbers of shovels of capacity Cj and k is the number of
types of shovels.
3. TRUCK-SHOVEL DISPATCHING PROBLEM
4. MATHEMATICAL FORMULATION
In the case of a heterogeneous fleet of shovels, during the
planning phase of a new open-pit or during a short- and In the haulage operations, the fuel consumption of each
medium-term planning of existing mines where it is needed equipment (truck and shovel) depends on its size, type, op-
to reallocate the shovels, one question arises: “what is the erating time, operating conditions and technical specifica-
best allocation of shovels that can result in the minimum tions, etc. Trucks consume fuel during loading, unloading
liters of fuel consumed per tonne moved for the entire and travel periods, whereas shovels consume fuel during
haulage operations?” An illustration of this problem is their working and idling periods.
shown on Fig. 1. This figure displays, for a mixed fleet
of shovel made up of shovels of two different capacities, The travel time from the i-th unloading point to the j-th
all the possible allocations of shovels in different pits of shovel, tye,ij , and the travel time from the j-th shovel to the
a generic open-mine constituted of two unloading points β-th unloading point, tylo,jβ , are calculated by equations (3)
and three loading points. and (4) respectively:
To answer this question, given a fleet of trucks, the model dij
tye,ij = y (i = 1, 2, ..., U ; j = 1, 2, ..., S), (3)
developed in this study is used to determine the expected ve,ij
liters of fuel consumed per tonne of material moved for djβ
all the possible shovel allocations in different pits. The tylo,jβ = y (β = 1, 2, ..., U ; j = 1, 2, ..., S), (4)
obtained results are then compared to find the allocation vlo,jβ
that results in the minimum liters of fuel consumed per where the subscripts e denotes an empty truck, lo denotes
tonne of material moved. a loaded truck; y is the truck index, dij is the distance
between the unloading point i to the shovel j, djβ is the
For this purpose, the total number of possible shovel distance between the shovel j to the unloading point β,
allocations in different pits for a mixed fleet of shovels, y
ve,ij is the speed at which the empty truck y travels from
Pj , is determined by the following equation: y
an unloading point i to a shovel j, vlo,jβ is the speed at
S! which the loaded truck y travels from the shovel j to an
Pj = , (2)
SC1 ! × SC2 ! × ... × SCj=k ! unloading point β, U is the number of unloading points

218
2019 IFAC MMM
D.M. Bajany et al. / IFAC PapersOnLine 52-14 (2019) 207–212
Stellenbosch, South Africa, August 28-30, 2019 209

and S is the number of shovels. For a complete shift, the idle time Ij of the j-th shovel is
calculated as:
N 
 U
The loading time tyl,j of a truck y with the j-th shovel is Ij = sh − y
Zlo,jβ tyl,j , (11)
determined by: y=1 β=1
Cy
tyl,j = , (5) where sh is the shift duration, and, the fuel Fidle,j con-
Cj sumed by the j-th shovel during its idling period is given
where Cy is the loading capacity of the truck y and Cj is as:
the capacity of the shovel j. Fidle,j = Ij fj,idle , (12)
The fuel consumption of a dump truck y per single cycle with fj,idle , the fuel consumption rate of shovel j during
is given as: idling time. For a complete shift, the fuel Fu,j consumed
by the j-th shovel during periods when it was used to load
y y
trucks is calculated as:
Fijβ = fe,ij tye,ij + flo,jβ
y
tylo,jβ N  U
y
y
+ fidle (tyl,j + tyu + tyw,j + tyw,β ), (6) Fu,j = fj Zlo,jβ tyl,j , (13)
y=1 β=1
y where fj is the fuel consumption of shovel j during the
where fe,ijis the fuel consumption of the empty truck y
moving from the unloading point i to the shovel j, flo,jβ y time it is used to load a truck.
is the fuel consumption of the loaded truck y moving from The total fuel Fj consumed by the j-th shovel during a
the shovel j to the unloading point β, fidle y
is the fuel shift is determined by:
consumption of the truck y during engine idling time, tyu Fj = Fidle,j + Fu,j . (14)
is the unloading time of the truck y, tyw,j is the waiting The total fuel F s consumed by all the shovels for the whole
time of the truck y at the shovel j, tyw,β is the waiting time shift duration is given as:
S
of the truck y at the dumping point β. Note that i = β
implies that the truck returns to the original unloading Fs = Fj . (15)
point where it started its cycle. j=1
The number of times xyj a y truck is loaded by the j-th The total fuel F consumed for whole haulage operations
shovel during a shift and the number of times a y truck during a shift is calculated as:
has dumped his load at the dumping β during a shift are F = F t + F s. (16)
calculated by equations (7) and (8) respectively: There are many variables that can be used to evaluate
U
 U
 the haul truck energy efficiency (Krzyzanowska (2007)).
xyj = y
Ze,ij or xyj = y
Zlo,jβ , (7) In this study, the liters of fuel per tonne moved is adopted
i=1 β=1 as a performance indicator for the truck-shovel dispatching
S S
system. The liters of fuel per tonne moved at the end of a
  shift is evaluated by:
xyβ = y
Ze,ij or xyβ = y
Zlo,jβ , (8)
F
j=1 j=1 LT ts = N U S . (17)
with U , the number of unloading points, and S, the  y
Zlo,jβ Cy
number total of shovels.
y=1 β=1 j=1
For xyj cycles that a y truck has accomplished to the j-
As previously mentioned, the objective of this study is to
th shovel during a shift, the fuel consumption F tyj of that
minimize the fuel consumption of both trucks and shovels.
truck is calculated as: The model is built in such a way that all the requirements
U
 U
 of dump sites are met and the optimal number of trips
F tyj = y
Ze,ij y
fe,ij tye,ij + y
Zlo,jβ y
flo,jβ tylo,jβ of each truck on each route of the pit is determined. The
i=1 β=1 technical specifications, such as payload of trucks, loaded
U
   capacity of shovels, and fuel consumption in function of the
y y
+ Zlo,jβ fidle tyl,j + tyu + tyw,j + tyw,β , (9) operating conditions of equipment are directly considered
β=1 in the mathematical model. The objective function is given
y in the following:
where Ze,ij is the number of journeys that the truck y has
min F. (18)
travelled from the i-th unloading point to the j-th shovel
y The operating constraints of the problem include the fol-
during a shift and Zlo,jβ is the number of journeys that
lowing:
the truck y has travelled from the j-th shovel to the β-th - The material transported from all shovels should be
unloading point during a shift. greater than the handling demand of each unloading point
From the above equation, for a given shift, the total (of ore and waste). This constraint ensures that trucks are
amount of fuel F t consumed by all trucks is calculated dispatched so that the production target at the mine is
as: satisfied. The way this requirement is taken into account
S  N
Ft = F tyj , (10) is shown in equation (19):
j=1 y=1 N 
S
 y
with N , the number of trucks used in the haulage opera- Zlo,jβ Cy ≥ Dβ sh, ∀β ∈ {i, w}, (19)
tions. y=1 j=1

219
2019 IFAC MMM
210
Stellenbosch, D.M. Bajany et al. / IFAC PapersOnLine 52-14 (2019) 207–212
South Africa, August 28-30, 2019

with Dβ , the handling demand per hour of the unloading times that a truck travels empty from that unloading point
point i and sh, the shift duration. to different shovels can not exceed one. This constraint is
- Material transported by trucks during the shift duration written as follows:
from each loading point is less than the shovel capacity S S
y y
allocated to that pit. Ze,ij − Zlo,jβ ≤ 1 f or i = β, ∀j, ∀y. (27)
j=1 j=1
N 
 U
y - The difference between the number of times an empty
Zlo,jβ Cy ≤ Cj sh, ∀j. (20)
truck y travels from an unloading point i to different
y=1 β=1
shovels and the number of time this truck dumps its load at
- The ore transported by trucks during the shift duration that unloading point can not exceed one. This constraint
from a loading point j to the unloading points of ore is is written as follows:
less or equal to their available quantity at the considered S S
loading point. y
Zlo,jβ − y
Ze,ij ≤ 1 f or i = β, ∀j, ∀y. (28)
N
y j=1 j=1
Zlo,jβ C y ≤ Oj , (21)
y=1
- The last constraint (29) ensures that the number of trips
of trucks are positive and integer.
∀j, ∀sh, β ∈ i with i = {i1 , i2 , ..., iq }, with Oj , the y y
available quantity of ore at the loading point j and q, the Ze,ij ∈ N + ; Zlo,jβ ∈ N + ; i = 1, 2, ..., U ; β = 1, 2, ..., U ;
total number of ore unloading points. j = 1, 2, ..., S; y = 1, 2, ..., N. (29)
- The waste transported by trucks during the shift duration
Constraints (26), (27) and (28) ensure that the continuity
from a loading point j to the unloading points of waste is
of loading and transportation is maintained.
less or equal to their available quantity at the considered
loading point.
N 5. SIMULATIONS
y
Zlo,jβ Cy ≤ Wj , (22)
y=1 A case study of a hypothetical downgrade open-pit mine
with two unloading points (one for waste materials and
∀j, ∀sh, β ∈ w with w = {w1 , w2 , ..., wr }, with Wj , the
one for ores) and three shovels located at three pits is
available quantity of waste at the loading point j and r,
considered for optimization and simulation of the model.
the total number of waste unloading points.
Among these shovels, two have a capacity of 1600 t/h and
- For the sh-th shift, the utilization time of each shovel
one, a capacity of 2000 t/h. The length of each road in the
is less or equal to the shift duration. This constraint is
haulage network is given in Table 1. The total resistance
written as follows:
of the haul road is assumed to be equal to 4%. Operating
N
 speeds of loaded trucks are estimated using the Rimpull-
xyj tyl,j ≤ sh, ∀j. (23) Speed-Grade curves given in Caterpillar.Inc (2007) 1 . The
y=1
mine topography and operating parameters of the loading
and haulage equipment used as input parameters of the
From equation (7), the constraint (23) can also be written
model are shown in Table 1. The three possible shovel
as:
allocations in different pits are determined by equation
3!
N 
U (2) as Pj = 2!1! = 3, they are displayed on Fig. 1.
 y
Zlo,jβ tyl,j ≤ sh, ∀j. (24)
y=1 β=1 6. SIMULATION RESULT
- The sum of cycle times of all cycles performed by a truck
y during the shift is less or equal to the shift duration. In order to show the effectiveness of the developed method
and the impact of the waiting time on the fuel consumed
S 
U S 
U
in the haulage operations, three dispatch scenarios with
  different waiting time at loading and unloading points have
F tyj = y
Ze,ij y
fe,ij tye,ij + y
Zlo,jβ y
flo,jβ tylo,jβ been considered in the simulations. The waiting times con-
j=1 i=1 j=1 β=1
sidered at loading and dumping points were as follows: 0.30
S 
 U   min for the first scenario, 0.48 min for the second scenario
y y
+ Zlo,jβ fidle tyl,j + tyu + tyw,j + tyw,β ≤ sh, and 0.78 min for the third scenario. The shift duration of
j=1 β=1 each scenario was 8 hours. Solving the developed model
∀y. (25) with a mixed integer nonlinear programming algorithm
- The number of trips that a truck y makes to a shovel for the three scenarios, the following results were found.
equals to the number of trips that the same truck leaves For the same amount of material transported (37248 t),
that shovel. Equation (26) shows how this requirement is the first scenario, with the shovels allocate following the
considered: allocation (a), results in 130.99 milliliters per tonne and
4879.80 liters of fuel consumed for the whole haulage
U
 U
 operation; with the allocation (b), it results in 136,90
y y milliliters per tonne and 5099.30 liters of fuel consumed for
Ze,ij = Zlo,jβ , ∀j, ∀y. (26)
i=1 β=1 the whole haulage operation and with the allocation (c),
- The difference between the number of times a truck y its results in 132,92 milliliters per tonne and 4851,12 liters
dumps its load at an unloading point i and the number of 1 Caterpilar. Inc, 777D Off-Highway Truck, USA.2007.

220
2019 IFAC MMM
D.M. Bajany et al. / IFAC PapersOnLine 52-14 (2019) 207–212
Stellenbosch, South Africa, August 28-30, 2019 211

Chart Title
Table 1. Mine topography and resources 3500
3000

Parameters Specifications 2500

Model 777D 2000

Litres
1500
Rated payload 96 t
1000
Gross vehicle weight 161 030 kg
500
Speed of empty truks 50 km/h
0
Speed of loaded trucks 36 km/h Allocation (a) Allocation (b) Allocation (c)
Scenario 1: 0.30 min 2325,13 2554,64 2530,2
Fuel consumption (idle time) 22.38 L/h
Scenario 2: 0.48 min 2498,4 2741,5 2682,35
Fuel consumption empty truck 44.76 L/h Senario 3: 0.78 min 2525,23 2945,33 2792,16
Fuel consumption loaded truck 78.33 L/h
Shovel capacity 1600 t/h
2000 t/h
Fig. 2. Fuel consumed in the haulage by trucks and shovels
per scenario Chart Title
Fuel consumption of shovel
- during idle time 1600 t/h - 6.6 L/h 5600
2000 t/h -9.5 L/h 5400
- during working time 1600 t/h - 117 L/h 5200

Liters
2000 t/h -130 L/h 5000
Mine topography Downgrade mine 4800
Fuel consumption loaded truck 78.33 L/h 4600
Gradient 1:14 40 4400
Distance between dumping Allocation (a) Allocation (b) Allocation (c)
and loading point i1 − j1 1.5 km Scenario 1: 0.30 min 4879,8 5099,3 4951,12
Scenario 2: 0.48 min 4956,8 5203,37 5052,16
i1 − j2 2.5 km
Scenario 3: 0.78 min 5259,14 5520,5 5338,45
i1 − j3 3 km
i2 − j1 3.5 km
i2 − j2 2.5 km Fig. 3. FuelChart
consumed
Title by trucks Chart Title
i2 − j3 1.5 km
Demand of ore at unloading
point i1 2375 t/h 32% 30%
35%
44%

of fuel consumed for the whole haulage operation. The 24% 35%
second scenario, with the allocation (a), results in 133,08
Shovel j1=1600t/h
milliliters per tonne and 4956 liters of fuel consumed for Shovel j2=1600t/h Chart Title
Shovel j1=1600t/h
Shovel j2=2000t/h
the whole haulage operation; with the allocation (b), it Allocation (a)
Shovel j3=2000t/h Allocation (b) Shovel j3=1600t/h

results in 139,69 milliliters per tonne and 5203,37 liters of

fuel consumed for the whole haulage operation and with
33%
the allocation (c), it results in 135.53 milliliters per tonne 39%

and 5052,16 liters of fuel consumed for the whole haulage.

The last scenario, with the allocation (a), results in 141,19 28%

milliliters per tonne and 5259,14 liters of fuel consumed Shovel j1=2000t/h
for the whole haulage operation; with the allocation (b), Shovel j2=1600t/h
Allocation(c)
it results in 148,21 milliliters per tonne and 5520.50 liters Shovel j3=1600t/h

of fuel consumed for the whole haulage operation and

with the allocation (c), it results in 143.32 milliliters and Fig. 4. Contribution of each shovel to the total handling
5338,45 liters of fuel consumed for the whole haulage demand Chart Title
(Waiting time: 0.30 min))
Chart Title
operation. The fuel consumed in the haulage by trucks
and shovels and by trucks itself per scenario are displayed
in Fig. 2 and Fig. 3 respectively. The contribution of each 34%
33% 33%
42%
shovel to the total handling demand per scenario is shown
in Fig. 4, Fig. 5 and Fig. 6 respectively. Comparing the 24% 34%
obtained results, the following observations were made.
The first scenario with small waiting time at loading and Shovel j1=1600t/h
Shovel j2=2000t/h
dumping sides is the most efficient scenario as it results Chart Title
Shovel j1=1600t/h
Shovel j2=1600t/h Allocation (b) Shovel j3=1600t/h
Allocation (a)
in less fuel consumed in the haulage and small liters per Shovel j3=2000t/h

tonne.
34%
The applicability of the developed model to determine 43%
the best shovel allocation was also demonstrated in this
study. Indeed, from the obtained results it is clear that 23%

the second allocation is the least efficient one compared to Shovel j1=2000t/h
the other two. Indeed, for the first scenario, by allocating Shovel j2=1600t/h
the shovel with the highest capacity at the second loading Shovel j3=1600t/h

point and for the same amount of material transported,

roughly 242.5 liters and 148.18 liters of fuel are addi- Fig. 5. Contribution of each shovel to the total handling
tionally consumed in the haulage operations compared to demand (Waiting time: 0.48 min)
the allocations (a) and (c) respectively . The explanation

221
2019 IFAC MMM
212
Stellenbosch, D.M. Bajany et al. / IFAC PapersOnLine 52-14 (2019) 207–212
South Africa, August 28-30, 2019

Chart Title Chart Title

Alarie, S. and Gamache, M. (2002). Overview of solution
33% 33% 31%
strategies used in truck dispatching systems for open
42%
pit mines. International Journal of Surface Mining,
25% 36%
Reclamation and Environment, 16(1), 59–76.
Bagheri, F., Fayazbakhsh, M., Thimmaiah, P.C., and
Shovel j1=1600t/h Shovel j1=1600t/h

Chart Title
Shovel j2=1600t/h Shovel j2=2000t/h Bahrami, M. (2015). Theoretical and experimental
Allocation (a) Shovel j3=2000t/h Shovel j3=1600t/h
investigation into anti-idling a/c system for trucks.
Energy Conversion and Management, 98, 173–183.
34%
Bajany, D., Xia, X., and Zhang, L. (2017a). A milp model
44% for truck-shovel scheduling to minimize fuel consump-
tion. Energy Procedia, 105, 2739–2745.
22%
Bajany, M.D. et al. (2017b). A mixed integer linear pro-
Shovel j1=2000t/h gramming model for truck-shovel scheduling to minimize
Allocation (c)
Shovel j2=1600t/h
Shovel j3=1600t/h
fuel consumption. Master’s thesis, University of Preto-
ria.
Fig. 6. Contribution of each shovel to the total handling Caterpillar.Inc (2007). 777D Off-Highway Truck. USA.
demand (Waiting time: 0.78 min) Da Cunha Rodovalho, E., Lima, H.M., and de Tomi, G.
(2016). New approach for reduction of diesel consump-
of this observation is that, with the allocation (b), the tion by comparing different mining haulage configura-
total travel distances of loaded and empty trucks are much tions. Journal of Environmental Management, 172, 177–
longer than those obtained with the other two allocations. 185.
Hence, trucks consume more fuel with the allocation (b). Jochens, P. (1980). The energy requirements of the mining
This is shown in the Fig. 3. Another explanation of this and metallurgical industry in south africa: presidential
observation is that the shovel of capacity 2000 t/h is not address. Journal of the Southern African Institute of
properly utilized with the allocation (b). This can be seen Mining and Metallurgy, 80(9), 331–343.
by comparing the contributions of this shovel to the total Kecojevic, V. and Komljenovic, D. (2010). Haul truck
handling demand displayed in Fig. 4, Fig. 5 and Fig. 6. To fuel consumption and co 2 emission under various engine
be specific, only 35% of materials moved in the haulage load conditions. Mining Engineering, 62(12), 44–48.
are loaded by the shovel of capacity 2000 t/h; whereas Krzyzanowska, J. (2007). The impact of mixed fleet haul-
for the two other allocations, 44% (allocation (a)) and ing on mining operations at venetia mine. Journal of
39% (allocation (c)) of materials are loaded by this shovel. The Southern African Institute of Mining and Metal-
With the 2000 t/h shovel being the most fuel efficient, less lurgy, 107(4), 215–224.
utilization of it resulted in more fuel consumption for the Lashgari, A., Johnson, C., Kecojevic, V., Lusk, B., and
allocation (b). Hoffman, J. (2014). Nox emission of equipment and
blasting agents in surface mining. In Proceedings of
The results of this work are important as the proposed the 143rd Annual Meeting and Exhibition of the Society
method considers the waiting time of trucks at shovels of Mining, Metallurgy and Exploration, Salt Lake City,
and dumping points and the specific characteristic of each UT.
truck and shovel used in the haulage, unlike previously May, M.A. (2013). Applications of queuing theory for open-
reported optimization work. pit truck/shovel haulage systems. Ph.D. thesis, Virginia
Tech.
7. CONCLUSION Peralta, S., Sasmito, A.P., and Kumral, M. (2016). Re-
liability effect on energy consumption and greenhouse
A model for minimization of fuel consumption of dump gas emissions of mining hauling fleet towards sustainable
trucks and shovels in open-pit mine was developed. To mining. Journal of Sustainable Mining, 15(3), 85–94.
evaluate the developed model, a case study of an open- Rahman, S.A., Masjuki, H., Kalam, M., Abedin, M.,
pit mine with two unloading points and three shovels was Sanjid, A., and Sajjad, H. (2013). Impact of idling on
considered. In the case of a mixed fleet of shovels, for each fuel consumption and exhaust emissions and available
possible shovel allocation, the optimal number of trips idle-reduction technologies for diesel vehicles–a review.
that each truck should make on each route of the mine Energy Conversion and Management, 74, 171–182.
during a shift was realized. It was possible to determine Sahoo, L.K., Bandyopadhyay, S., and Banerjee, R. (2014).
the best shovel allocation which is the allocation with Benchmarking energy consumption for dump trucks in
the lowest fuel consumption in the haulage operation and mines. Applied Energy, 113, 1382–1396.
therefore with the lower diesel emission. The results of this Suzdaleva, E. and Nagy, I. (2014). Data-based speed-limit-
work serve the foundation for future research that could respecting eco-driving system. Transportation Research
include the fact that, in practice, the waiting time varies Part C: Emerging Technologies, 44, 253–264.
stochastically. Zhang, L. and Xia, X. (2015). An integer programming
approach for truck-shovel dispatching problem in open-
REFERENCES pit mines. Energy Procedia, 75, 1779–1784.
Abdelaziz, E., Saidur, R., and Mekhilef, S. (2011). A
review on energy saving strategies in industrial sector.
Renewable and Sustainable Energy Reviews, 15(1), 150–
168.

222