CIM Journal Vol. 10 No. 4

CIM JOURNAL
Editors International Advisory Board

Agus Sasmito McGill University, Maintenance Engineering Members
and Reliability Society Dr. Saiied Aminossadati, Senior Lecturer, University of
Martin Grenon Université Laval, Rock Engineering Society Queensland, Australia
Ali Madiseh The University of British Columbia, Underground Dr. Juliette Biao Koudenoukpo, Director and Regional
Mining Society Representative, Africa Office, UN Environment, Kenya
Steve McCutcheon McCutcheon GeoConsulting, Geological Dr. Mark Board, Vice President – Technical Services, Hecla
Society Limited, USA
Mr. Vitor Correia, President, European Federation of
FOUNDING EDITOR Chuck Edwards, Chuck Edwards
Geologists, Portugal
Extractive Metallurgy Consulting
Dr. Éva Hartai, Professor, University of Miskolc; Editor,
ASSOCIATE EDITORS European Geologist Journal, Hungary
Donna Beneteau University of Saskatchewan, Underground Dr. Pierre-Yves Le Meur, Professor, Université Paul Valéry,
Mining Society Site Saint-Charles, France and Pacific New Studies Centre,
Erin Bobicki University of Toronto, Canadian Mineral New Caledonia
Processors Society Dr. Leon Lorenzen, Lorenzen Consultants Pty Ltd, Mintrex,
James Budac Sherritt Metals, Metallurgy and Materials University of Stellenbosch, Australia
Society Dr. Baisheng Nie, Professor, China University of Mining and
Gabriel Castillo The University of British Columbia, Technology, China
Environmental & Social Responsibility Society Mr. Bjorn Nielsen, VP Dry Comminution & Sorting, Outotec,
Loy Chunpongtong Gamontip Consultants, Surface Mining Australia
Society Dr. Bob Ring, Principal Consultant, ANSTO Minerals,
Hendrik Falck NWT Geological Survey, Geological Society Australian Nuclear Science and Technology Organisation,
Australia
Édith Garneau Independent Consultant, Environmental &
Social Responsibility Society Dr. A. J. S. (Sam) Spearing, Director, Western Australian
School of Mines, Curtin University, Australia
EDITORIAL COORDINATOR Janice Burke, jburke@cim.org Mr. Barry Vayler, Analysis & Improvement Specialist, BHP,
Australia
PUBLISHER CIM, cim@cim.org
© 2019 Canadian Institute of Mining, Metallurgy and Petroleum

It is a condition of publication that manuscripts submitted to this publication may be reproduced, stored in a retrieval system or
journal have not been published and will not be simultaneously transmitted in any form or by any means, electronic, electrostatic,
submitted or published elsewhere. By submitting a manuscript, the magnetic tape, mechanical, photocopying, recording or otherwise,
authors agree that the copyright for their article is transferred to the without permission in writing from the copyright holder.
Canadian Institute of Mining, Metallurgy and Petroleum if and when
the article is accepted for publication. However, assignment of Permission for other use
copyright is not required from authors who work for organizations The copyright owner’s consent does not extend to copying for
which do not permit such an assignment. The copyright covers the general distribution, for promotion, for creating new works or for
reprints, photographic reproductions, microfilm, or any other resale. Specific written permission must be obtained from the
reproductions of similar nature and translations. No part of this publisher for such copying.
The Institute, as a body, is not responsible for statements made or opinions advanced either in articles or in any discussion appearing in
its publications.
ISSN 1923-6026
Instructions for authors are available on the CIM website
Published quarterly by the Canadian Institute of Mining, Subscription rates

Metallurgy and Petroleum Online version included in CIM Membership ($197/yr).
1250 – 3500 de Maisonneuve Blvd. West Print version for institutions or agencies – Canada: $275/yr (AB,
Westmount, QC H3Z 3C1 BC, MB, NT, NU, SK, YT add 5% GST; ON add 13% HST; QC
Tel: 514.939.2710; Fax: 514.939.2714 add 5% GST + 9.975% PST; NB, NL, NS, PE add 15% HST).
www.cim.org Print version for institutions or agencies – USA/International:
US$325/yr. Online access to single copy: $50.
Layout & design: www.clocommunications.com
CIM JOURNAL
Volume 10, Number 4, 2019
Contents
157 Breaking down energy consumption in industrial grinding mills

J. Bouchard, G. LeBlanc, M. Levesque, P. Radziszewski, and D. Georges-Filteau
165 Hydraulic excavator versus electric rope shovel performance

R. Andreev, T. G. Joseph, J. Sammut, and M. Curley
173 Performance prediction of a full-size rapid dampening suspension in an ultra-class hauler

application
R. K. Soni, T. G. Joseph, and M. Curley
180 Is bigger still better? Considerations in increasing the size of haulage equipment
M. Dotto, T. G. Joseph, and M. Curley
187 Impact of operating conditions on emissions from mine haul trucks operating in the oil
sands region of Alberta, Canada
R. J. Schafer-Frentz, T. G. Joseph, and M. Curley
200 Material handling and flowsheet selection considerations

for particle-based ore sorting
B. Foggiatto, J. T. Ross, T. Donkin, B. Patterson, M. I. Guthrie, and G. Lane
MINERAL PROCESSING 157
Breaking down energy consumption in industrial grinding mills
J. Bouchard
Université Laval, Quebec City, Quebec, Canada
G. LeBlanc
CanmetMINES / CanmetMINING, Natural Resources Canada, Ottawa, Ontario, Canada
M. Levesque
CanmetMINES / CanmetMINING, Natural Resources Canada, Sudbury, Ontario, Canada
P. Radziszewski
Metso Minerals Canada, Lachine, Quebec, Canada
D. Georges-Filteau
Université Laval, Quebec City, Quebec, Canada
https://doi.org/10.15834/cimj.2019.18
ABSTRACT Grinding mills are infamous for their extremely low energy efficiency. This paper reports
the results of an investigation that characterized energy outputs in industrial grinding mill circuits. It
identifies and discusses the potential for recovering energy from grinding circuits. Agnico-Eagle Goldex
Division, Mine Canadian Malartic, and New Gold New Afton Mine participated in the study by provid-
ing operating data for three semi-autogenous grinding mills and four ball mills. Results showed that on
average, 79% of the supplied electrical energy was converted to heat absorbed by the slurry, 8% was lost
through the drive system, and approximately 2% was transmitted to the ambient air. Only 9% of the input
energy was actually used for grinding. The remaining 91%—wasted as heat—could potentially be recov-
ered using suitable technologies or integrated energy management systems. This topic remains to be fur-
ther investigated.
■ KEYWORDS Comminution processes, Energy recovery potential, Grinding mills
RÉSUMÉ Les broyeurs sont tristement célèbres pour leur rendement énergétique médiocre. Cet article
présente les résultats d’une étude qui caractérise les rendements énergétiques des circuits industriels de
broyage et examine les possibilités de récupération d’énergie dans les circuits de broyage. La mine Gol-
dex d’Agnico Eagle, Mine Canadian Malartic et la mine New Afton de New Gold ont participé à l’étude
en fournissant des données de fonctionnement pour trois broyeurs semi-autogènes et quatre broyeurs à
boulets. Les résultats montrent qu’en moyenne 79 % de l’énergie électrique fournie sont convertis en
chaleur absorbée par la pulpe, 8 % sont perdus dans le système d’entraînement et environ 2 % sont dis-
persés dans l’air ambiant. Seuls 9 % de l’énergie primaire sont, de fait, utilisés pour le broyage ; les 91 %
restants (perdus sous forme de chaleur) peuvent potentiellement être récupérés à l’aide de technologies
adaptées ou intégrées dans des systèmes de gestion de l’énergie. Cette question demeure toutefois
ouverte.
■ MOTS CLÉS broyeurs, potentiel de récupération de l’énergie, procédés de comminution
INTRODUCTION
Grinding is largely recognized as a very inefficient pro- 2004) when comparing the input energy to that required for
cess; energy efficiency estimates range from < 1 to 2% generating new mineral surfaces. Criticizing an “ill defini-
(Fuerstenau & Abouzeid, 2002; Tromans & Meech, 2002, tion of the reference for the output energy,” Fuerstenau and
CIM Journal | Vol. 10, No. 4, 2019

158 J. Bouchard, G. LeBlanc, M. Levesque, P. Radziszewski, and D. Georges-Filteau
Abouzeid (2002) proposed a “more meaningful baseline” west of Val-d’Or, Quebec), extracting 55,000 t/day averag-
would be “the energy for producing new surface area by the ing 1.1 g/t to produce 700,000 ounces of gold and
compression loading or impact loading of single speci- 950,000 ounces of silver per year. After a two-stage crush-
mens.” Based on this reference, they concluded that “the ing, the ore is processed in a closed circuit SAG mill
ball mill is reasonably efficient energetically,” exhibiting an (11.6 × 6.4 m EGL; 19.4 MW; gearless drive). Oversize
efficiency of approximately 15% for quartz. material then feeds two pebble crushers, while the passing
Schellinger (1951) and Schellinger and Lalkela (1951) product goes to two identical (treated as one in this study)
defined thermodynamic efficiency as the ratio of effective parallel ball mills (7.3 × 11 m EGL; 11.9 MW each). The
work (in comminution, the difference between the energy product goes to a third grinding stage with one identical
input and that lost as heat) to energy input. They determined ball mill (7.3 × 11 m EGL; 11.9 MW).
that the thermodynamic efficiency of comminution ranges The New Afton operations are located 10 km south of
from 10 to 20%. Tromans (2008) introduced the “relative Kamloops, British Columbia, Canada. New Gold Inc. oper-
efficiency ratio,” involving the concept of “maximum ideal ates this underground blockcaving mine with an output of
limiting efficiency” against which the conventional energy 15,000 t/d, producing over 85,000 ounces of gold and
efficiency is compared. Even using this definition, effi- 75 million pounds of copper per year. The grinding circuit
ciency figures remain very low, ranging from 3 to 26%. comprises a SAG mill (8.5 × 4 m EGL; 5,220 kW), a pebble
Although it is commonly known that industrial grinding crusher processing the oversize material, and a secondary
circuits lose—or generate—significant amounts of heat, lit- ball mill (5.5 × 10 m EGL; 5,220 kW) processing the
tle information is known about the actual energy balance. undersize fraction.
Radziszewski (2013) used a thermodynamic analysis to
estimate that 43% of the energy input in a typical mill is Quantifying energy flows
transferred to the slurry, raising the temperature of the dis- To characterize energy flows in a comminution circuit, it
charge product. It was only recently that a more precise pic- is necessary to i) define a control volume around the rele-
ture began to emerge with a pilot project for developing a vant pieces of equipment and ii) determine the input and
methodology to map energy flows (Bouchard et al., 2016). output energy streams within this control volume, as illus-
Following the idea of recovering energy that is currently trated in Figure 1. Here, ṁ is the mass flowrate; h is specific
wasted (Radziszewski, 2013; Radziszewski & Hewitt, enthalpy; the subscripts sl, air, in, and out are the slurry, air,
2015), the objective was to propose a tool to characterize inlet, and outlet streams, respectively; Ẇelec is electrical
the potential for energy recovery from industrial grinding power input; Q˙ lost is all heat losses (evaporation, convec-
circuits. tion/radiation, and dissipated in mechanical and electrical
This paper reviews the results of a pilot project con- components); and Ẇfrag is work output (creation of new
ducted at Agnico Eagle Mines Goldex Division and supple- surface, liners, and grinding media wear, plastic deforma-
ments them with new survey data from two other tion and mechanical losses).
operations: Canadian Malartic (Agnico Eagle Mines and The energy balance can be written around the control
Yamaska Gold) and New Afton (New Gold). A discussion volume as:
follows that analyzes the potential for energy recovery and
examines avenues to increase energy efficiency. Ẇelec−Ẇfrag− Q ˙ lost=
(1)
ṁsl out hsl out −ṁsl in hsl in+ṁair out hair out −ṁair in hair in
METHODS
Study sites
Agnico Eagle Mines Goldex Division is located in Val-
d’Or in northwestern Quebec, Canada. This underground
mine extracts 5,100 t/d grading 1.5 g/t to produce
100,000 ounces of gold per year. Run-of-mine (ROM) ore
feeds a two-stage crushing circuit before entering the pro-
cessing plant. An open-circuit semi-autogenous grinding
(SAG) mill (7.32 × 3.73 m effective grinding length [EGL];
3,357 kW) processes the product from the crushing stage.
The discharged slurry is further ground in a preclassifica-
tion closed-circuit ball mill (5.03 × 8.23 m EGL; 3,357 kW)
reducing the P80 (sieve dimension greater than 80% of ore
Figure 1. Control volume around a grinding circuit; ṁ: mass flowrate; h:
particles in the product, weight basis) to ~100 μm. specific enthalpy; sl, air, in, and out: slurry, air, inlet, and outlet streams,
Canadian Malartic, a Yamana Gold/Agnico Eagle Mines ˙ : all heat losses; Ẇ : electrical power input; Ẇ :
respectively; Qlost elec frag
joint venture, operates an open-pit mine in Malartic (25 km work output

Breaking down energy consumption in industrial grinding mills 159
The term corresponding to Ẇelec and those on the right the creation of surface energy within the tumbling cham-
hand side can all be characterized from operation data, tember.” In other words, it is the “energy absorption from the
perature measurements, and slurry composition. tumbling system,” calculated as a difference using equa-
˙ , and the members on the
tion 1: after quantifying Ẇelec, Qlost
Quantifying heat losses right hand side, the only remaining unknown is Ẇfrag.
The power losses corresponding to Q˙ lost can be broken As discussed by Bouchard et al. (2016), grinding media
down into three subcomponents: and steel liners typically do not undergo significant plastic
1. heat dissipated in the electrical and mechanical compo- deformation. This suggests that very little of the mill power
nents, draw is involved in mechanical deformation work. Never-
2. heat dissipated at the mill shell through convection/radi- theless, the resulting distribution is included in Ẇfrag as a
ation, and percentage of electrical input power.
3. latent energy absorbed by water during evaporation. Unlike other sources of mechanical work, the energy
The main mechanical and electrical components typi- dissipated as vibrations and noise was independently quan-
cally installed in a grinding mill are the transformer, vari- tified in the pilot phase of this project at Goldex Division.
able frequency drive (for mill speed modulation), electric Results revealed that sound and vibrations exhibited negli-
motor, gearbox, mill trunnions, and oil cooling system. gible power levels (~1.1 kW) (Bouchard et al., 2016). They
Power losses from a given component are dissipated as heat were therefore considered as part of Ẇfrag for the remainder
and are proportional to the power transferred to the equip- of the project.
ment and its efficiency. The input power used to perform mechanical work can-
The mechanical and electrical components used to not be recovered. Thus the potential for energy recovery
power a grinding mill are commonly used in industrial corresponds to that within Q˙ lost as well as that in the mass
applications, thus technical data for these are available flows of the air and slurry streams.
from manufacturers. Heat losses were estimated using these
data. The remaining heat losses corresponding to convec- RESULTS
tion, radiation, and evaporation can be estimated from All results and operation information presented here rep-
equipment dimensions and operation data, provided a few resent data collected during the second half of 2015 and
simplifying assumptions are made, as demonstrated by early 2016 from the three processing plants that partici-
Radziszewski and Hewitt (2015). pated in the study: Goldex, Canadian Malartic, and New
The input variables of the model comprise ROM ore min- Afton.
eral composition, flowrates (ROM ore, water, and lubricating The energy flow model for each mill used in this case
oil), and temperatures (ROM ore, water, lubricating oil, bear- study is depicted in Figure 2, which shows the following
ing, and discharged slurry). Solids fractions (mill fresh feed, eight output heat flows:
mill discharge, and hydrocyclone streams) and mill para- • power loss dissipated as heat in the transformer (Q˙ 1),
meters (dimensions, power draw, and speed) are also variable frequency drive (Q ˙ ), electric motor (Q
˙ ), gear-
2 3
required to map energy flows in a grinding circuit. Whenever ˙
Q ˙
Q
box ( 4), trunnion cooling system ( 5), and convection
possible, the values were retrieved from the data historian and radiation around the mill shell (Q ˙ ); and
6
but in some instances, off-line measurements were taken • enthalpy flows with air ( 7) and slurry (Q˙ 8) streams at
˙
Q
using portable devices such as temperature data loggers. the mill discharge.
Quantifying work output

Estimating the power output result-
ing from mechanical work performed
inside the control volume is more chal-
lenging. There are essentially four
main sources of mechanical work:
1. ore comminution,
2. wear (grinding media and liners),
3. plastic deformation (grinding media
and liners), and
4. vibration and noise.
The interpretation of the concept of Figure 2. Energy flow model common to the three mills used in the case study; Q ˙ : power loss
1
dissipated as heat in the transformer; Q˙ : variable frequency drive; Q ˙ : electric motor; Q˙ : gearbox;
mechanical work in a grinding system 2 3 4
˙ : trunnion cooling system; Q
Q ˙ : convection and radiation around the mill shell; Q ˙ : enthalpy flows
used in this paper follows the one pos- 5 6
˙
7
with air streams at the mill discharge; Q8: enthalpy flows with slurry streams at the mill discharge;
tulated by Schellinger (1952): “the dis- ṁ: mass flowrate; h: specific enthalpy; sl, air, and in: slurry, air, and inlet streams, respectively;
appearance of energy […] caused by Ẇelec: electrical power input; Ẇfrag: work output

The distribution of the power output as a percentage of mills. This suggests that the energy distribution established
the electrical power input is presented in Figures 3 and 4 for here could be used as an initial approximation of energy
the SAG and ball mill circuits, respectively. losses simply upon knowing the total power supplied to a
Results from the case studies reveal that energy trans- grinding system.
ferred to the slurry was by far the most significant loss Table 1 provides some key temperature values in the cir-
within grinding processes. In fact, it represented on average cuits for the process water and mill discharges. It shows
79.4% of the total electric energy supplied to the grinding that even though most energy was captured in the slurry, it
mills. All the other types of heat losses only accounted for consisted of low-grade sources as indicated by the rela-
approximately 11.7%. tively low temperatures reached (< 50°C).
The remaining energy was therefore converted into
mechanical work. As mentioned previously, ore grinding Table 1. Temperatures in grinding circuits for three mills
and internal wear characterize the major part of this expen- Temperature Goldex Canadian New
diture. This fraction of the energy is also unrecoverable. On (°C) Division Malartic Afton
average, it corresponded to 8.2 and 9.6%, respectively, for Process water 20 26 6
the SAG and ball mills and 8.9% for the overall circuits. Ore 18.2 5 13.7
These values are consistent with the fraction representing SAG mill discharge 35 26 19
mechanical work (10–20%) reported by Schellinger (1951)
Ball mill discharge 38 36 16
and Schellinger and Lalkela (1951), which were deter-
mined using a similar approach in a laboratory environ-
ment. DISCUSSION
It is also interesting to note that results were relatively
consistent from one mill to another, especially for ball Energy recovery potential
Thermal energy losses accounted for an average of
approximately 91% of the total supplied energy identified
in this study: 79% was converted to heat absorbed by the
slurry, 8% was lost through the drive system, and approx-
imately 4% was transmitted to the ambient air and cooling
system. These losses could theoretically be recovered,
transferred to other heat consumers, or used to produce
electricity. The most obvious application would be a heat
exchanger between the slurry and a cold fluid (air or solu-
tion) requiring preheating. However, one could also con-
template extracting the heat from electrical rooms
(hosting motor control centers and drives) with heat
pumps also using cold fluids requiring to be preheated as
heat sinks.
Heat exchanger applications generally exhibit efficien-
cies of approximately 75%. However, considering the low
Figure 3. Distribution of the power output in SAG mill circuits temperature range in this case (16–38°C; Table 1), recover-
ing thermal energy could be difficult. Moreover, even if a
substantial amount of heat is potentially available for
recovery, it can only be useful if a demand is identified.
Various scenarios could be further investigated in this
regard. For instance, the low-grade stored energy could be
transferred in the first preheating stages for several applica-
tions, ranging from elution solution in a cyanidation plant,
to underground ventilation air, to reagent preparation.
On the other hand, it must be emphasized that thermal
losses through convection and radiation, as well as those
from the electric motor, gearbox, evaporation, and trunnion
cooling system, are de facto recovered during the cold
weather period of the year as they contribute to heating the
building. This is also true for a portion of the enthalpy con-
veyed by the slurry in pump boxes and in downstream pro-
Figure 4. Distribution of the power output in ball mill circuits cessing tanks. Energy management systems must therefore

integrate both passive and active heat recovery strategies very little effect on the power draw. The issue of reducing
and exploit their full potential, for example, using forced the specific energy consumption (power drawn per ore
convection as a means to facilitate distribution inside a throughput, e.g., kWh/t) can be tackled by operating closer
large volume or between levels. to design capacity (Levesque & Millar, 2015) and using
Conversion to electricity is another avenue for energy process control capabilities (Nunez, MacPherson, Graffi, &
recovery. As explained by Radziszewski (2013), “[h]eat Tuzun, 2009).
engines convert heat to a mechanical form of energy that Properly designed process control strategies facilitate
can be used directly or converted into electricity where the reducing the variability of key process variables, which in
maximum theoretical efficiency is defined by Carnot’s effi- turn can increase the comminution efficiency and decrease
ciency relationship.” Table 2 illustrates that Carnot efficien- grinding requirements. As discussed in depth by Bouchard,
cies—calculated assuming water at 10°C as the cold Sbarbaro, and Desbiens (2018), “automation, process con-
sink—were very low within the temperature ranges trol and real-time optimization allow not only to determine
observed in the survey data. Thus, trying to generate elec- the proper operating point, but also provide means to reach
trical energy is not currently worthwhile with Carnot ther- and maintain it over time, regardless of fluctuating input
mal efficiency averaging 6.7%. material attributes, and process disturbances.”
Solutions are therefore needed to improve the Carnot For new applications, taking advantage of more energy-
efficiency. Among those proposed by Radziszewski (2013), efficient processing equipment, for example, high pressure
dry milling is promising since it involves temperatures of grinding rolls and crushers (Morrell, 2009; van der Meer &
approximately 85°C, which are more favourable for energy Gruendken, 2010; Nordell, Porter, & Potapor, 2016), is also
recovery. Technologies for low-grade heat recovery (e.g., an avenue that must be considered to reduce the specific
heat pumps, heat exchangers, and power generators) must energy consumption of comminution circuits.
also be examined.
Beyond energy recovery considerations, the impact of Reducing comminution requirements
modifying operating temperatures should also be analyzed Perhaps a more holistic approach is to contemplate the
carefully before any action is taken in this respect overall comminution chain as put forward in the so-called
(Radziszewski, 2013). In fact, decreasing the grinding mill “mine-to-mill” concept (i.e., consistent and fine blasting
slurry temperature could have an impact on ball and lifter product; Kanchibotla, Valery, & Morrell, 1999). Torrealba-
plate wear. It could be positive if it increases the life of the Vargas, Esteban, Roy, and Runnels (2016) presented a sim-
consumables or negative if it accelerates corrosion degrada- ulation case study for an open-pit operation showing that a
tion. more “aggressive” blasting strategy can significantly
A similar investigation should also be undertaken reduce grinding requirements. Results demonstrated that
regarding flotation performance. Flotation chemicals react increasing the powder factor from 0.28 to 0.93 kg/t would
differently depending on the temperature. For gold leaching increase the mass fraction of material < 3.175 cm (1.25 in)
plants, higher temperatures tend to accelerate mass transfer at the SAG feed from 34 to 47%, hence a reduction of the
and chemical reactions but reduce the availability of oxy- F80 (sieve dimension greater than 80% of ore particles in
gen, a reactant in the dissolution of gold with cyanides. The the feed, weight basis) from 89 to 79 mm. The total esti-
optimum temperature is found when the oxygen supply mated economic benefits were reductions of operating
becomes insufficient to fuel the reaction. costs from 7.0 to 24.5%, increasing with the powder factor.
There is, however, a strong belief in the industry that
Reducing the specific energy consumption incremental improvements simply would not be enough in
Results of the survey showed that a substantial share of the face of ever-increasing costs of energy, dwindling
the energy footprint of conventional SAG/ball mill circuits grades of remaining orebodies, and higher social expecta-
was being lost as heat. Inadequate or suboptimal operation tions (Kellogg Innovation Network, 2014; Howes, 2016).
explains some of the inefficiency, and part of the solution Additionally, the mining industry is severely lagging
could be found in processing more ore, which should have behind all the other sectors in terms of innovation, with one
Table 2. Carnot efficiency data

Heat losses Energy stored Maximum slurry Carnot
(Q˙ 1−Q˙ 8) in the slurry (Q˙ 8) temperature efficiency (%)†
(kW) (kW) (°C)
Goldex 5,072 4,116 38 9.00
Canadian Malartic 34,623 30,739 36 8.41
New Afton 9,917 8,865 18 2.75

† using 10°C water

of the lowest relative R&D investment allowances, averag- Fluidized-bed flotation has also been shown to be appli-
ing 0.4% of revenues (Gill, 2015). cable for sulfide ores. For example, Mankosa, Kohmuench,
Industry think-tanks, influencers, and executives con- Luttrell, Herbst, and Noble (2016) presented a simulation
sider mining to be at a crossroads and that it needs disrup- case in which a HydroFloat™ separator was retrofitted to a
tive innovation. This means changing the current mining porphyry copper plant. The proposed idea was to use this
and processing paradigm. In mineral processing, the first equipment as a scavenger “designed to reject large ton-
step would be to transform the archaic, widespread, and nages of well-liberated siliceous gangue that often con-
rigid plant design practice. Ball mills may appear hard to sumes a large volume of the circulating load in primary
replace at first glance: they have low capital cost, low main- grind circuits.” In fact, the HydroFloat™ separator was
tenance, and are reliable and proven. However, considering capable of floating particles as large as 850 µm that had
that most (> 99%) of the energy injected for comminution only 1% sulfide mineral exposure (Miller et al., 2016).
is used on worthless material (Lessard, de Bakker, & Therefore, when used as a scavenger, the separator was
McHugh, 2014), one could question the absolute need for capable of producing tails that contained nearly no valuable
large sized and multiple unit installations. Reducing the minerals. In their simulation, Mankosa et al. (2016) showed
fraction of ROM that is being processed could be worth that with this retrofit, it was possible to increase the grind
examining as a practical means for reducing comminution size from 200 to 300 µm, while virtually not affecting the
requirements. overall recovery. Indeed, tails only increased from 0.07 to
In this regard, ore sorting is now a proven technology 0.08% Cu compared to the base case. In the end, the coarser
and should at least be considered in all greenfield projects. grind size combined with the mass rejection resulted in a
Lessard et al. (2014) presented a detailed case study for two 25% increase of the primary mill capacity.
molybdenum mines and showed that ore sorting could be
used after secondary crushing to reject 66.5% of mass, CONCLUSIONS
while recovering 93.4% of Mo. Upstream removal of waste This paper provided an overview of how energy con-
material in the circuit showed a 45% specific energy reduc- sumption is used in grinding circuits. Three SAG/ball mill
tion, from 33,000 kWh (base case) to 17,800 kWh per circuits were surveyed, revealing that on average, 91% of
tonne Mo produced. In a more aggressive sorting scenario, the supplied energy resulted in heat losses, leaving only 9%
90% mass rejection led to 86.4% Mo recovery and 60% for ore breakage. The slurry absorbed most of the energy
energy cost reduction. Moreover, Lessard, Sweetser, Bar- (79%), raising the temperature between the mill inlet and
tram, Figueroa, and McHugh (2016) showed that ore sort- outlet; 8% was lost through the drive system; and approxi-
ing also exhibits very interesting payback periods and mately 4% was transmitted to ambient air and cooling sys-
internal rate of return. In their conservative estimate for a tem. Results from all three circuits were consistent and
US copper mine, they showed that by retrofitting an ore similar to those presented by Schellinger (1951) and
sorter on the pebble circuit of an existing SAG mill, 30% of Schellinger and Lalkela (1951), thus indicating that
the mass could be rejected while recovering 93.3% of Cu. improvements in comminution efficiency have been negli-
This scenario represented a payback period of only gible during the past seven decades.
19.3 months and an internal rate of return of 61%. All of these losses add up to a significant amount of poten-
The whole concept of ore sorting deserves more atten- tially recoverable energy. However, this energy is stored in
tion to be expanded and generalized. Beyond the single the slurry and air at relatively low temperature (< 50°C), mak-
crush-sort-grind-process layout, other configurations such ing conversion into a useable supply challenging.
as multiple sequential crush-sort stages could unveil addi- Future work could examine various strategies and tech-
tional benefits. This is probably where coarse ore separa- nologies for low-grade heat recovery such as heat pumps,
tion technologies could excel. For instance, coarse particle heat exchangers, and power generators, among others.
flotation has been used for many years in potash and phos- Energy recovery solutions would require retrofits in exist-
phate concentrators. Since these minerals usually exhibit a ing plants with the purchase and installation of heat trans-
relatively coarse liberation size, fluidized-bed flotation fer/conversion equipment, which may affect operations,
cells (e.g., HydroFloat™) are used in conjunction with con- metal recovery, and maintenance. Thus a techno-economic
ventional flotation to recover the valuable minerals on a assessment would be required to determine whether the
wider range of particle size, therefore reducing unnecessary options are financially viable. Efforts should also aim to
grinding. In fact, while conventional flotation showed very determine how design practices could be adapted for future
high efficiency in a narrow particle size range (10–150 µm) mine sites to enable an integrated approach for energy man-
(Awatey, Skinner, & Zanin, 2015), the HydroFloat™ tech- agement.
nology was shown to vastly outperform conventional flota- For existing plants, process control and real-time opti-
tion in the recovery of 425 to 1,180 µm particles (Awatey, mization offer great potential to reduce the specific energy
Thanasekaran, Kohmuench, Skinner, & Zanin, 2013; requirements of comminution circuits with minimum capi-
Kohmuench, Mankosa, Thanasekaran, & Hobert, 2018). tal investment. Even though the general principles are

Jocelyn Bouchard is currently an assistant professor in the Department of

known, work is still required to benchmark energy effi- Chemical Engineering at Université Laval, specializing in mineral
ciency gains, develop advanced applications (e.g., processing and process modelling, simulation, control, and optimization.
observers for unmeasured process variables, economic jocelyn.bouchard@gch.ulaval.ca
model-based predictive control), and determine the optimal Gilles LeBlanc graduated from mechanical engineering in 1986 from
load of every stage of the comminution chain. Université Laval. He is R&D team leader at Natural Resources Canada,
CanmetMINING, Ottawa, with research interests in technological
innovations for the mining industry.
ACKNOWLEDGMENTS
The authors would like to thank Agnico Eagle Mines, Michelle Levesque is the engineering technical lead for the Data-Driven
Mine Canadian Malartic, New Gold New Afton Mine, and Technologies Team at CanmetMINING. She obtained an MASc and PhD in
Natural Resources Engineering at Laurentian University, with a focus on
CanmetMINING for granting permission to publish this energy management in mining. Her research interests and current work
work. Further acknowledgments have to be given to Yan include application of digital technologies to improve sustainability in
Germain (CanmetMINING), who was instrumental in mining.
retrieving data, as well as Sam Marcuson and Carl Weath- Peter Radziszewski has a PhD in mechanical engineering from Université
erell (CMIC) and Nabil Bouzoubaâ (CanmetMINING) for Laval. His research work in mineral processing (comminution) has led him
championing this R&D initiative. to McGill University’s Department of Mechanical Engineering and the
Canadian Space Agency. His research efforts have contributed to the
development of discrete element methods in modelling tumbling mill charge
An earlier draft of this paper was published in the dynamics and to development of a steel media wear prediction methodology.
Proceedings of the 49th Annual Canadian Mineral Peter is the author and co-author of more than 100 papers in journals and
conference proceedings. Since joining Metso in 2012, Peter has contributed
Processors Operators Conference prior to undergoing the to the continuous development of an innovation culture and to the
CIM Journal peer-review process. exploration of innovative equipment and process concepts in comminution,
flotation, sensor development, and tailings processing.
Paper reviewed and approved for publication by the David Georges-Filteau graduated from Université Laval in 2017 with a
Canadian Mineral Processors Society of CIM. bachelor’s degree in mining and mineral processing engineering. He then
completed a master’s degree at Université Laval in chemical engineering
with specialization in lithium flotation and process control. David is
currently junior metallurgist at Nemaska Lithium, where he is responsible for
process control at the Whabouchi processing plant in northern Quebec.
REFERENCES
Awatey, B., Skinner, W., & Zanin, M. (2015). Incorporating fluidised- Kanchibotla, S. S., Valery, W., & Morrell, S. (1999). Modelling fines in
bed flotation into a conventional flotation flowsheet: A focus on energy blast fragmentation and its impact on crushing and grinding. Proceedings
implications of coarse particle recovery. Powder Technology, 275(May), of Explo ‘99 – A conference on rock breaking, Kalgoorlie, Australia.
85–93. https://doi.org/10.1016/j.powtec.2015.01.065
Kellogg Innovation Network (2014). Reinventing mining: Creating sustain-
Awatey, B., Thanasekaran, H., Kohmuench, J. N., Skinner, W., & Zanin, able value. Evanston, IL: Author. Retrieved from http://www.kinglobal.
M. (2013). Optimization of operating parameters for coarse sphalerite org/uploads/5/2/1/6/52161657/pb_kin_dpf_final_12_4_5mb.pdf
flotation in the HydroFloat fluidised-bed separator. Minerals Engineering,
50–51(September), 99–105. https://doi.org/10.1016/j.mineng.2013.06.015 Kohmuench, J. N., Mankosa, M. J., Thanasekaran, H., & Hobert, A.
Bouchard, J., LeBlanc, G., Germain, Y., Levesque, M., Tremblay, N., (2018). Improving coarse particle flotation using the HydroFloatTM
Légaré, B., Dallaire, B., & Radziszewski, P. (2016). The CMIC / Can- (raising the trunk of the elephant curve). Minerals Engineering,
metMINES Comminution Energy Recovery Potential Initiative – The 121(June), 137–145. https://doi.org/10.1016/j.mineng.2018.03.004
Agnico Eagle Goldex Division case. Proceedings of the 28th Interna- Lessard, J., de Bakker, J., & McHugh, L. (2014). Development of ore sort-
tional Mineral Processing Congress (IMPC 2016), Quebec City, QC. ing and its impact on mineral processing economics. Minerals Engineer-
Bouchard, J., Sbarbaro, D., & Desbiens, A. (2018). Plant automation for ing, 65(October), 88–97. https://doi.org/10.1016/j.mineng.2014.05.019
energy-efficient mineral processing. In K. Awuah-Offei (Ed.), Energy
Lessard, J., Sweetser, W., Bartram, K., Figueroa, J., & McHugh, L.
Efficiency in the Minerals Industry: Green Energy and Technology.
Cham, Switzerland: Springer International. https://doi.org/10.1007/978- (2016). Bridging the gap: Understanding the economic impact of ore
3-319-54199-0_13 sorting on a mineral processing circuit. Minerals Engineering, 91(May),
92–99. https://doi.org/10.1016/j.mineng.2015.08.019
Fuerstenau, D. W., & Abouzeid, A.-Z. M. (2002). The energy efficiency
of ball milling in comminution. International Journal of Mineral Process- Levesque, M. Y., & Millar, D. L. (2015). The link between operational
ing, 67(1–4), 161–185. https://doi.org/10.1016/S0301-7516(02)00039-X practices and specific energy consumption in metal ore milling plants –
Gill, K. (2015). Innovation in mining – a call to action. Montreal QC: Ontario experiences. Minerals Engineering, 71(February), 146–158.
Canadian Institute of Mining, Metallurgy and Petroleum. https://doi.org/10.1016/j.mineng.2014.11.010
Howes, R. (2016, May). We need disruptive innovation in the mining Mankosa, M. J., Kohmuench, J. N., Luttrell, G. H., Herbst, J. A., &
sector, as well. The Globe and Mail. Retrieved from https://www.the- Noble, A. (2016). Split-feed circuit design for primary sulfide recovery.
globeandmail.com/report-on-business/rob-commentary/we-need-dis- Proceedings of the 28th International Mineral Processing Congress
ruptive-innovation-in-the-mining-sector-as-well/article28817988/ (IMPC 2016), Quebec City, QC.

Miller, J. D., Lin, C. L., Wang, Y., Mankosa, M. J., Kohmuench, J. N., Schellinger, A. K. (1952). Solid surface energy and calorimetric deter-
& Luttrell, G. H. (2016). Significance of exposed grain surface area in minations of surface energy relationships for some common minerals.
coarse particle flotation of low-grade gold ore with the HydrofloatTM Mining Engineering, 4(4), 369–374.
technology. Proceedings of the 28th International Mineral Processing
Congress (IMPC 2016), Quebec City, QC. Schellinger, A. K., & Lalkela, R. D. (1951). Approximation of the
energy efficiency of commercial ball mills by the energy balance
Morrell, S. (2009). Predicting the overall specific energy requirement method. Transactions AIME, 190, 523–524.
of crushing, high pressure grinding roll and tumbling mill circuits.
Minerals Engineering, 22(6), 544–549. https://doi.org/10.1016/j.mineng Torrealba-Vargas, J., Esteban, R., Roy, D., & Runnels, D. (2016). Mine
.2009.01.005 to mill approach to optimize power consumption in a process plant oper-
ation by modelling and simulation. Proceedings of the 28th Interna-
Nordell, L. K., Porter, B., & Potapor, A. (2016). Comminution energy tional Mineral Processing Congress (IMPC 2016), Quebec City, QC.
efficiency – Understanding the next steps. Proceedings of the 28th Inter-
national Mineral Processing Congress (IMPC 2016), Quebec City, QC. Tromans, D. (2008). Mineral comminution: Energy efficiency consider-
ations. Minerals Engineering, 21(8), 613–620. https://doi.org/10.1016
Nunez, E., MacPherson, G., Graffi, D., & Tuzun, A. (2009). Self-opti-
/j.mineng.2007.12.003
mizing grinding control for maximising throughput while maintaining
cyclone overflow specifications. Proceedings of the 41st Annual Meet- Tromans, D., & Meech, J. A. (2002). Fracture toughness and surface
ing of the Canadian Mineral Processors, Ottawa, ON. energies of minerals: Theoretical estimates for oxides, sulphides, sili-
Radziszewski, P. (2013). Energy recovery potential in comminution cates and halides. Minerals Engineering, 15(12), 1027–1041.
processes. Minerals Engineering, 46–47(June), 83–88. https://doi.org https://doi.org/10.1016/S0892-6875(02)00213-3
/10.1016/j.mineng.2012.12.002 Tromans, D., & Meech, J. A. (2004). Fracture toughness and surface
Radziszewski, P., & Hewitt, D. (2015). Exploring the effect of energy energies of covalent minerals: Theoretical estimates. Minerals Engi-
recovery potential on comminution efficiency: the Glencore Raglan neering, 17(1), 1–15. https://doi.org/10.1016/j.mineng.2003.09.006
Mine case. Proceedings of SAG Conference 2015, Vancouver, BC. van der Meer, F. P., & Gruendken, A. (2010). Flowsheet considerations
Schellinger, A. K. (1951). A calorimetric method for studying grinding for optimal use of high pressure grinding rolls. Minerals Engineering,
in a tumbling medium. Transactions AIME, 190, 518–522. 23(9), 663–669. https://doi.org/10.1016/j.mineng.2009.09.012
©2019 Canadian Institute of Mining, Metallurgy and Petroleum. All rights reserved.
MAINTENANCE, ENGINEERING AND RELIABILITY 165
Hydraulic excavator versus electric rope shovel performance
R. Andreev, T. G. Joseph, J. Sammut, and M. Curley

University of Alberta, Edmonton, Alberta, Canada
ABSTRACT Decisions regarding mining method and subsequent choice of excavator must take into
account the quantitative factors of productivity, availability, and capital, operating, and maintenance
costs. Dig performance was compared between a hydraulic excavator (HE) and electric rope shovel (RS)
of similar size class (i.e., use the same amount of energy to excavate a unit quantity of the same material
from a mining face), focusing on the quantitative factors above. Machine performance, normalized to
bucket capacity, was assessed as a function of hours of service (60,000 hours per machine). Results
showed higher production rates for the HE of similar capacity and age. In general, the RS exhibited lower
service cost per volume capacity, though it had higher capital cost. The RS became more cost effective
than the HE of similar capacity within approximately 30,000 hours of service.
■ KEYWORDS Cost, Electric rope shovel (RS), Excavation, Hydraulic excavator (HE), Surface mining
RÉSUMÉ Les décisions concernant la méthode minière à adopter et le choix ultérieur d’un engin d’ex-
cavation doivent tenir compte des facteurs quantitatifs de la productivité, de la disponibilité ainsi que des
coûts d’investissement, d’exploitation et d’entretien. Cette étude compare les performances en termes
d’excavation d’une excavatrice hydraulique et d’une pelle électrique à câbles de même taille (qui utili-
sent par exemple la même quantité d’énergie pour creuser une quantité unitaire du même matériau dans
un front de taille), en se concentrant sur les facteurs quantitatifs susmentionnés. La performance de la
machine, normalisée à la contenance du godet de volume unitaire, est évaluée en fonction des heures de
service (60 000 heures d’activité par machine). Les résultats montrent des taux de production plus élevés
pour l’excavatrice hydraulique d’une capacité et d’une ancienneté similaires. De manière générale, la
pelle électrique à câbles affiche un coût d’entretien par capacité volumique inférieur, mais son coût d’in-
vestissement est plus élevé. La pelle électrique à câbles devient plus rentable que l’excavatrice hydrau-
lique d’une même capacité au bout d’environ 30 000 heures de fonctionnement.
■ MOTS CLÉS coût, excavation, excavatrice hydraulique, exploitation à ciel ouvert, pelle électrique à câbles
INTRODUCTION
The most common excavating tools in surface mining row excavator size range (e.g., Mattis, Cheskidov, &
operations are the hydraulic excavator (HE) and electric Labutin, 2012). Production per cubic metre of bucket
rope shovel (RS). In general, electric RSs are more reliable capacity has been shown to vary significantly and depend
and cheaper to maintain (P&H Mining Equipment, 2007), on operating conditions, which in turn are a function of the
but where complicated geological conditions exist, HEs are instability of mining conditions and operational activity.
preferred because they are generally considered to be more This paper evaluates the excavation energy derived from
mobile. the area under a dominating hoist power versus time curve
Differences in design and kinematics make HEs and RSs throughout the dig cycle—similar to Joseph and Shi
application-specific and difficult to compare in terms of (2011)—in order to compare HE and RS performance,
performance. Excavator performance can be assessed using availability, productivity, and maintenance costs. The ana-
availability (proportion of time the equipment is available lytic approach was tested against a global set of field data.
to work) and cost per unit of production or bucket capacity Overfill concerns that could possibly skew the data were not
(Burt & Caccetta, 2014). Previous studies of performance considered in this paper. The curl breakout fill mechanism in
and maintenance requirements over relatively short time a HE makes overfilling difficult (Figure 1, left), and the dip-
periods have often focused on single operations and a nar- pers or buckets on a RS (Figure 1, right) are restricted in

166 R. Andreev, T. G. Joseph, J. Sammut, and M. Curley
Figure 1. Idealized balanced hydraulic excavator breakout and crowd forces (left) and predominant electric rope shovel hoist and crowd forces (right).
Adapted from Komatsu Germany GmbH (2016) and Komatsu Mining Corp. (2018).
their ability to achieve fill factors greater than 100% (Joseph 4100XPC electric RS (Table 1, Figure 3). Many costs were
& Shi, 2011; Rasimarzabadi & Joseph, 2016). clearly much higher for the HE, with an overall differential
of C$394.80 per operating hour (approximately C$6.58 per
RESEARCH METHODS AND BASIS FOR dig cycle or C$0.087 per tonne excavated). These costs dif-
ANALYSES fered when Joseph and Sammut (2013) considered the initial
Comparison of dig performance

Joseph and Sammut (2013) compared the dig cycle power
between a Komatsu PC8000 HE (Komatsu Germany GmbH,
2016) and a P&H 4100XPC electric RS (Komatsu Mining
Corp., 2018) operating in a well-broken mining face. The
power plot for the HE was derived from the hydraulic crowd
and breakout power delivered to the paired stick and curl
cylinders (Figure 1, left). The RS power plot was derived
from field armature voltage and current data from the two
motors driving the three sets of reduction gears and the hoist
drum system, given an established efficiency of 86.5%
(Joseph & Shi, 2011). The efficiency relates to the total
power from the motors over the main sheaves at the top of Figure 2. Specific energy versus cycle time for a Komatsu PC8000
the boom to the dipper (bucket) and ground engaging tools. hydraulic excavator and P&H 4100XPC electric rope shovel
The HE excavated 76 t of payload for 14.7 MJ of energy
expended compared to 119 t for 23 MJ of energy expended Table 1. Capital, operating, and maintenance base costs in Canadian
by the RS (Figure 2). However, the overall performance of dollars for a Komatsu PC8000 hydraulic excavator and P&H 4100XPC
electric rope shovel (after Joseph and Sammut, 2013)
both machines was identical (193 kJ/t specific energy). In
Capacity and cost Hydraulic Electric rope
other words, when excavating rock of similar geology and excavator shovel
fragmentation, both the HE and RS expended the same spe-
Rated capacity (m3) 42.0 46.0
cific energy when performing the dig function. This should
Capital cost ($M) 14.3 41.7
not be surprising because dig energy should be a function
Operating parts ($/h) 104.2 49.8
of the degree of difficulty of excavation, which is primarily
Operating labour ($/h) 98.4 53.3
related to the resistance of the material being excavated. In
Maintenance parts ($/h) 162.1 74.1
a well-fragmented mining face, the cohesive resistance of
Maintenance labour ($/h) 150.5 79.9
the material is negligible, leaving friction at the fragmented
Fuel/power ($/h) 202.6 89.1
face to dominate the shear resistance to excavation.
Lubricants ($/h) 69.5 46.3
Wear parts ($/h) 22.0 22.0
Review of a previous cost analysis Total capital ($M) 14.3 41.7
Joseph and Sammut (2013) evaluated the operating and
Total operating/maintenance ($/h) 809.3 414.5
maintenance costs of the Komatsu PC8000 HE and P&H

Hydraulic excavator versus electric rope shovel performance 167
The HE excavating approximately 76 t per 60 s pass and

the RS excavating approximately 119 t per 40 s pass trans-
lated to a 2.35:1 productive availability leveraging ratio in
favour of the RS (Joseph & Sammut, 2013). This suggests
that a 1% decrease in availability for a 76 t capacity HE
could be matched by a 2.35% decrease in availability for a
119 t capacity RS, with no loss of matched productive capa-
bility. With this in mind, the 12% difference in availability
in Figure 4 translates to a 5.1% (i.e., 12/2.35) availability
equivalent difference when considering the productive
capacity of the respective machines at the 25 year point.
Further, it reflects an effective availability advantage for
the HE over the RS for the last 6.5 years—well within the
lifespan of the third commissioned HE. This in turn trans-
lates to an approximate C$31M difference in cumulative
remaining sustaining capital plus higher operating costs for
the HE than the RS. The caveat here is that a small change
in the availability estimate could either create a balance
Figure 3. Maintenance and operating costs per hour for the Komatsu between the two machines within the 25-year timeframe or
PC8000 hydraulic excavator (HE) and P&H 4100XPC electric rope
shovel (RS). After Joseph and Sammut (2013).
show a more substantial advantage for the RS over the HE.
The analysis by Joseph and Sammut (2013) contained a
number of assumptions, the most significant being the
and sustaining capital investment, with an estimated 6% availability decline over a period of time by the RS or HE.
inflation in a total cost of ownership over a 25-year period It was therefore considered appropriate to compile and ana-
(Figure 4). This represented the useful available life of an lyze a more extensive dataset to provide greater clarity on
ultra-class RS (greater than 35 m3 bucket capacity), during availability, as well as productivity and maintenance and
which it continued to exhibit declining availability, reaching operating costs (where available).
70–80% as year 25 approached. During the same period, the
HE would be replaced twice—in 2022 and 2031—prospec- The “global dataset” and analysis reduction
tively maintaining an availability in the 85–95% range. At criteria adopted
the 25 year point, the cost difference between the two The HE and RS performance parameters targeted in this
machines was estimated to be C$75M, representing a 12% global data study are reported as regional means per ore
difference in shovel availability, and the third HE would be type, calculated from either raw data collected by the
at year 7 of its expected 9-year service lifespan (Joseph & authors at mining operations or through access to data pro-
Sammut, 2013). What then is the higher availability value vided courtesy of Komatsu Mining Germany GmbH (2015).
that would generate sufficient higher productive capability The authors respected the requests of the mining companies
to offset an increasing cost of ownership? reporting the data that their identities remain anonymous. To
manage the high volume of data from the 126 operating
units (10 HEs and 116 RSs) sourced from 15 mines across
Europe, Asia, Africa, Australia, the USA, and Latin Amer-
ica, productivity and maintenance cost data were further
normalized to bucket capacity. Means are listed for avail-
ability (Table 2), monthly productivity (Table 3), and main-
tenance cost (Table 4). The majority of the data was sourced
from coal mining operations. Of the 126 datasets, 32 were
sourced from six metal or industrial mineral mining opera-
tions out of the 15 mining operations surveyed.
All operations employed blasting of overburden, waste
rock, and ore prior to excavation. The blasted material was
sized for ease of excavation. Given that diggability is a
function of the size of blast fragments, which are consid-
ered frictional materials, the angularity of the fragments
Figure 4. Cumulative cost of ownership (C) and availability (A) for the
adds to the existing frictional resistance to excavation
Komatsu PC8000 hydraulic excavator (HE) and P&H 4100XPC rope (Rasimarzabadi & Joseph, 2018). It was assumed that dif-
shovel (RS). After Joseph and Sammut (2013). ferences in blast design among mining operations would

Table 2. Hydraulic excavator and electric rope shovel availability by bucket capacity, mined ore, and global region (Komatsu Mining Germany GmbH, 2015)
Hours of service (×103)
Bucket capacity Ore Region No. 6 12 18 24 30 36 42 48 54 60
(m3)
Hydraulic excavator Availability (%)
10 Coal Europe 4 94.5 89.7 88.5 87.6 – – – – – –
11 Coal Europe 4 92.2 91.7 86.9 77.3 82.6 82.5 80.3 76.1 – –
15 Coal Asia 8 97.3 96.2 95.1 92.6 92.8 88.1 – – – –
22 Coal Latin America 5 94.3 89.6 88.8 89.0 90.2 90.8 91.0 93.2 – –
22 Coal Europe 1 95.7 91.6 81.5 73.7 66.1 86.7 – – – –
22 Coal Asia 14 94.5 91.4 91.0 89.4 87.8 86.4 90.6 – – –
22 Coal Africa 5 95.6 94.4 94.0 89.2 – – – – – –
19 Iron ore Australia 5 81.2 78.6 – – – – – – – –
19 Iron ore Africa 5 75.0 84.8 91.6 – – – – – – –
19 Copper Latin America 1 90.9 81.0 87.2 – – – – – – –
28 Coal Europe 1 94.5 92.7 78.0 76.1 – – – – – –
28 Coal Australia 1 – 93.4 – – – – – – – –
28 Coal USA 2 95.8 93.4 90.4 94.9 – 88.9 90.8 91.0 92.6 –
25 Iron ore Australia 5 79.6 77.4 81.5 – 76.8 77.9 – – – –
25 Iron ore Africa 2 – 81.6 77.6 – – – – – – –
25 Iron ore Europe 2 91.3 89.2 84.1 83.1 88.9 – – – – –
28 Copper Latin America 17 88.2 85.8 81.8 82.6 73.8 82.2 82.8 84.6 87.9 85.5
28 Phosphate Africa 1 – – – – 95.0 95.0 – – – –
25 Uranium Africa 4 – – – – – 92.4 – 96.0 93.3 93.3
42 Coal Latin America 19 95.0 91.4 91.1 90.9 89.3 89.3 89.7 89.2 88.6 91.0
42 Coal Africa 2 – 91.5 89.4 84.7 – – – – – –
42 Copper Latin America 8 86.5 83.2 72.6 62.7 – – – – – –
Electric rope shovel
10–18 Coal Europe 6 86.6 86.6 83.9 87.3 81.5 80.4 76.6 74.4 78.0 69.2
20 Coal Europe 2 91.8 88.4 77.9 67.7 65.3 59.3 – – – –
33 Coal Europe 2 91.8 88.1 86.6 84.3 83.9 86.5 82.0 80.8 82.4 –
Dashes indicate no data
Table 3. Hydraulic excavator and electric rope shovel monthly productivity normalized to bucket capacity (Komatsu Mining Germany GmbH, 2015)
(m3)
10 Coal Europe 4 2,295 2,104 2,093 1,993 – – – – – –
11 Coal Europe 4 1,558 1,356 1,474 1,367 1,455 1,372 1,442 1,233 – –
22 Coal Latin America 5 819 784 728 746 677 709 615 654 – –
22 Coal Europe 1 536 537 403 476 333 183 – – – –
22 Coal Asia 3 1,090 – 1,134 – – – – – – –
22 Coal Australia 1 – – 507 – – – – – – –
28 Coal Europe 1 476 539 343 224 – – – – – –
28 Coal Australia 1 – 552 – – – – – – – –
25 Iron ore Europe 2 6,238 6,024 2,610 5,233 4,360 – – – – –
42 Coal Latin America 19 507 489 464 443 438 409 405 421 429 392
12 Coal Europe 6 1,121 1,386 1,372 1,451 1,276 1,393 1,186 1,222 1,094 1,087
19 Coal Europe 4 711 715 – – – – – – 676 662
20 Coal Europe 2 1,034 1,037 600 373 410 437 – – – –
33 Coal Europe 2 594 637 527 514 576 588 568 547 357 –

Table 4. Hydraulic excavator and electric rope shovel maintenance cost (C$) normalized to bucket capacity (Komatsu Mining Germany GmbH, 2015)
(m3)
10 Coal Europe 4 0.13 0.18 0.26 0.20 – – – – – –
11 Coal Europe 4 0.21 0.20 0.21 0.43 0.42 0.38 0.45 0.25 – –
22 Coal Europe 1 0.07 0.17 0.45 0.34 0.21 – – – – –
28 Coal Europe 1 0.11 0.12 0.23 – – – – – – –
12 Coal Europe 5 – – – – 0.10 0.09 0.17 0.19 0.19 0.21
20 Coal Europe 2 0.11 0.14 – – – – – – – –
33 Coal Europe 2 – 0.08 0.12 0.13 0.16 0.13 0.14 0.14 0.18 –
account for differences in diggability reported in Table 2 for Applying these revised availability values to the HE and
a given HE or RS. RS data in Figure 4 generates Figure 6, which illustrates a
Availability and production data were reduced to difference of 16.5% availability for the same cumulative
monthly means normalized to bucket capacity, allowing the operating plus capital cost differential. Following the appli-
wide range of excavator sizes to be included in the evalua- cation of the 2.35 productive capability factor differential
tion. Over 4,700 months of records (equivalent to 3.4 mil- defined by Joseph and Sammut (2013), an effective avail-
lion calendar hours of operation) were analyzed for the ability difference of approximately 7.0% occurs at year 25.
126 operating units. For consistency in interpretation,
availability (A) is defined as “a percentage measure of the
degree to which machinery and equipment is in an operable
state” (Katukoori, 2007). This definition is expanded in
equation 1:
Available
hours
A = –––––––– =
Scheduled
hours
{ Scheduled − Maintenance
hours delays
}
–––––––––––––––––––––– × 100%
Calendar − Scheduled
hours delays
(1)
In the case of availability, there were no discernable dif-

ferences between coal mines and base metal mines, thus all
datasets were pooled. Given that coal mines were the sole
operations with productivity and maintenance cost data for
Figure 5. Hydraulic excavator (HE) and electric rope shovel (RS)
both HEs and RSs, data from coal mines are the focus of availability versus hours of service
this paper. Additional data in Tables 2–4 are presented here
because they may be useful to future parallel studies.
RESULTS AND ANALYSIS
Availability
Figure 5 shows the mean availability versus hours of
service from Table 2, regardless of commodity mined and
regional location. Both the HE and RS availability data
show well-defined negative logarithmic relationships with
hours of service. If we assume for a given operation that the
shift change and rest/inspection breaks amount to approxi-
mately 1.5 hours over a 12-hour shift, this would translate
to 87.5% utilization of the equipment, equivalent to
7,665 h/yr. If we extrapolate the lines in Figure 5 to
25 years (191,625 h), HE and RS availabilities are 83.5 and Figure 6. Revised Figure 4 using availability (A) relationships from
67.0%, respectively. Figure 5

Productivity be needed to match the 25 year life point capability of a

The following scenario at a well-blasted rock mining P&H 4100XPC RS. However, it is unreasonable to consider
face was used to demonstrate application of the analysis: the HE operating effectively at the 25 year point. A more
• bank density of the rock = 2.65 t/m3 reasonable approach would be consider HE replacement at
• swell factor = 1.25 on blasting 9 years, such that productivity (i.e., 9 × 7,665 = 68,985
• additional swell factor on filling excavator bucket = 0.05 hours of service) would be 732.3 t/m3 per 12-h shift or
• Broken rock density (in bucket) = 2.04 t/m3 2,380 t/h. Projecting out to a comparable asset life at
2.65
(i.e., ––––––––)
(1.25+0.05)
27 years (replacing the HE twice), a comparable perfor-
mance for the RS (i.e., 27 × 7,665 = 206,955 hours of ser-
Using the example HE versus RS comparison of Joseph vice) would be 1,005 t/m3 per 12-h shift or 3,853 t/h. The
and Sammut (2013), a 76 t payload HE bucket capacity RS/HE performance ratio is thus reduced from 1.96 to 1.62.
would be 37.3 m3 (loose cubic metres), within the rated Application of the relationships from Figure 7 to the HE
Komatsu PC8000 HE bucket volumetric capacity of 42 m3. versus RS example of Joseph and Sammut (2013) reveals
Similarly, for a 119 t payload RS, dipper capacity would be that two HE units would be required per RS (Figure 8). Fur-
58.3 m3. However, the volumetric capacity limit for the ther, the availability and perhaps field utilization impact on
example P&H 4100XPC dipper is 46 m3, thus a volumetric productivity would be much more significant for the HE
capacity restriction must be enforced in a comparative anal- than the RS. In operations, the high-mobility advantage of
ysis. The work of Joseph and Shi (2011) and Rasimarzabadi HEs is often poorly managed, resulting in lower productive
and Joseph (2018) was used as a guideline in justifying that utilization of HE assets. Including the two HE replacement
the likelihood of the dataset being skewed due to possible points illustrates the overall declining performance (cyclic
bucket overloading. with HE asset replacement) with time (Figure 8).
Normalizing by bucket capacity permitted the majority
of the available (i.e., global coal mine) data to be included Maintenance costs
in this analysis. Figure 7 again shows very clear trends for Analysis of the data in Table 4 reveals exponential rela-
both HE and RS performance, which affords the ability to tionships between maintenance costs and hours of service
account for productivity restrictions due to the loose rock for the HE and RS (Figure 9). These relationships were
density, as seen above with the calculated (58.3 m3) versus applied to the HE and RS scenario data from Joseph and
actual (46 m3) RS dipper capacities. The trends in Figure 7 Sammut (2013) over the projected 25-year span. The same
extrapolated to the 25-year point (191,625 hours of service) labour and major component costs were used because the
suggest the HE maximum productivity per 12-hour shift cumulative maintenance costs in the global data set did not
would be 604.6 t/m3 capacity and the RS maximum produc- consistently include either labour or major components
tivity would be 1,007 t/m3. (Figure 10).
In this scenario, at the projected 25year asset age mark These relationships were then extended to re-evaluate
for a Komatsu PC8000 39 m3 rated HE bucket and a P&H the adaptation of the Joseph and Sammut (2013) scenario
4100XPC 46 m3 rated RS dipper, the predicted productivity (Figure 6) to add maintenance and operating costs to the
values would translate to 1,965 and 3,860 t/h, respectively. overall cost of ownership (Figure 11). The availability mar-
The RS/HE performance ratio (1.96) indicates that—solely gin remained the same as that reflected in Figure 6, but the
on a productivity basis—two Komatsu PC8000 HEs would difference in cost of ownership narrowed by C$15M by
Figure 7. Hydraulic excavator (HE) and electric rope shovel (RS) Figure 8. Relationships from Figure 7 applied to the Komatsu PC8000
productivity per 12-hour shift hydraulic excavator (HE) and P&H 4100XPC electric rope shovel (RS)

Figure 10. Relationships from Figure 9 applied to operating plus

Figure 9. Hydraulic excavator (HE) and electric rope shovel (RS) maintenance costs for the Komatsu PC8000 hydraulic excavator (HE) and
maintenance costs per 12-hour shift in Canadian dollars P&H 4100XPC electric rope shovel (RS) over a 25-year period
using the refined global dataset trends. With an effective status (Figure 12). The negative slopes of the productivity
availability change at 25 years of 3.5% relative to the orig- relationships are functions of the non-availability (1–A);
inal analysis, this translates to C$4.3M for each 1% incre- the RS productivity slope is four times more steep than the
ment in availability for the HE, assuming the dataset is a HE slope. This is partly due to the RS experiencing a more
representative sample of reported operational field data. rapid loss of productivity—with a duty cycle effectively
50% faster than that of a HE mode of operation and bucket
Cost and productivity versus availability capacity more than 25% larger—even with comparable
A final set of relationships was assembled from the class machines. These two considerations alone would
above relationships and the full global dataset to predict account for half of the slope difference.
performance for HEs and RSs as a function of bucket The maintenance cost relationships also have a common
capacity. Figure 12 illustrates the cross-relationships intercept of C$0.00 at 100% availability status: no mainte-
between maintenance cost and productivity as a function of nance would be needed. The positive slopes of these rela-
availability (reported here as a fraction). Given that the tionships are also proportional to the non-availability
availability (Figure 5), productivity (Figure 7), and mainte- (1–A): maintenance costs are proportional to the proportion
nance costs (Figure 9) were logarithmic or exponential of time that a piece of equipment is unavailable to do work.
functions of hours of service, Figure 12 shows simpler lin- The steeper slope for the HE line is partly a function of the
ear cross-relationships as functions of availability. more complex and time-consuming maintenance required
The productivities of both HEs and RSs intercept the for the hydraulic system in a HE.
axis at a common maximum (1,165 t/m3 bucket capacity
per 12-h shift), which is the theoretical 100% availability
Figure 12. Productivity and maintenance costs (Canadian dollars) versus

Figure 11. Figure 6 revised using additional relationships from Figures 9 availability for a hydraulic excavator (HE) and electric rope shovel (RS)
and 10 using the global dataset

Tim G. Joseph, PhD, P.Eng., FCIM, is associate dean of engineering and

CONCLUSIONS professor of mining engineering at the University of Alberta, and director of
Although RSs generally have a higher initial capital cost, the Alberta Equipment–Ground Interactions Syndicate (AEGIS) research
they exhibited overall lower service cost per volume capac- program. He is vice-president of the Association of Professional Engineers
and Geoscientists of Alberta (APEGA), an active ambassador and former
ity. They became more cost effective than HEs of similar vice-president to the CIM, the president and principal engineer of JPI Mine
capacity within approximately 30,000 h of operation. Equipment, and recognized as an expert in mine equipment performance
The equations presented in Figure 12 could facilitate related to ground–equipment interactions. He delivers certification industry
short courses and mining equipment performance evaluations to the global
decision-making regarding selection of excavation equip- resource industry, mining equipment manufacturers, and service providers.
ment where the geology is uniform or fragmentation distri- He is a former recipient of the CIM Distinguished Service Medal for his
bution characteristics on blasting are predictable, thus contributions to the global mining industry and mining engineering
education and a recipient of the CIM Past President’s Memorial Medal for
allowing for use of either HEs or RSs. The more focused early achievements in education to the global mining industry. Dr. Joseph
relationships presented in Figures 5, 7, and 9 provide will retire from the University of Alberta in 2020 but will continue to
greater confidence in comparing RSs and HEs. practice through JPI Mine Equipment for the foreseeable future.
tjoseph@ualberta.ca
An earlier draft of this paper was published in the John Sammut’s career has spanned over 35 years since he graduated with a
Proceedings of the Maintenance, Engineering and BSc (Hons) in mining engineering in 1980 from Cambourne School of
Mines, UK, including 23 years with UK mining operations and equipment
Reliability/Mine Operators Conference (MEMO 2017) suppliers. From 2004 to 2013, he was general manager (UK) and area
prior to undergoing the CIM Journal peer-review process. manager (Canada, USA, and Mongolia) for Komatsu Mining Germany
GmbH, specializing in large hydraulic excavators. After an additional two
years with Panafrican Group in West Africa, John joined the University of
Paper reviewed and approved for publication by the Alberta, School of Mining & Petroleum Engineering in 2016, where he is
Maintenance, Engineering and Reliability Society of CIM. Industrial Mining Professor in Residence; he was an adjunct professor in
mining engineering at the University of Alberta from 2011 to 2015. John
Rodion Andreev resides in the Novosibirsk Region of the Russian Federation continues to work closely with Dr. Joseph on a variety of industry and
and is currently assistant head of the Mine Ventilation and Safety Department, student-focused projects.
Uvalnaya Coalmine, Sibirskaya Coal Company, Russia. In 2015, he graduated
with a MSc in mining engineering under the direction of Dr. Joseph and Mark Curley, MSc, EIT, is a graduate of the University of Alberta mining
Professor Sammut, specializing in performance criteria for hydraulic engineering BSc (2013) and MSc (2016) programs. As a mining engineer in
excavators and rope shovels. In addition to his mining engineering Alberta, he works for BGC Engineering in Edmonton on international
background, Rodion specializes in strategic planning and contract negotiation. mining projects, with considerable time spent in the Dominican Republic
expanding exploration targets. He has also worked for Dr. Joseph at the
University of Alberta as a research assistant.
REFERENCES
Burt, C. N., & Caccetta, L. (2014). Equipment selection for surface min- from https://www.komatsu-mining.de/fileadmin/komatsu/documents
ing: A review. Interfaces, 44(2), 143–162. https://doi.org/10.1287 /PC8000-6_A4_internet.pdf
/inte.2013.0732
Komatsu Mining Corp. (2018). P&H 4100XPC (Product specification).
Joseph, T. G., & Sammut, J. (2013, May). Battle of the titans – hydraulic Milwaukee, WI: Author. Retrieved from https://mining.komatsu/docs
excavators versus rope shovels. Paper presented at Haulage and Load- /default-source/product-documents/surface/electric-rope-shovels/4100
ing 2013, Phoenix, AZ. xpc-specification-sheet.pdf?sfvrsn=6c470a6b_40
Joseph, T. G., & Shi, N. (2011). Scaling analysis via geometric and per- Mattis, A. R., Cheskidov, V. I., & Labutin, V. N. (2012). Choice of the
formance data for shovel dipper design. CIM Journal, 2(4), 171–177. hard rock surface mining machinery in Russia. Journal of Mining Sci-
ence, 48(2), 329–338. https://doi.org/10.1134/S1062739148020141
Katukoori, V. K. (2007). Standardizing availability definition. Unpub-
lished manuscript, Department of Engineering, University of New P&H Mining Equipment (2007). Electric versus hydraulic. Mining
Orleans, New Orleans, LA. Retrieved from https://www.plant-mainte- Magazine, 42–51.
nance.com/articles/Availability_Definition.pdf
Rasimarzabadi R., & Joseph T. G. (2016). Particle flow mechanism into
Komatsu Mining Germany GmbH (2015). Production data for global cable shovel dippers. Journal of Terramechanics, 64(April), 10–22.
surface mining operations (Internal report). Düsseldorf, Germany: https://doi.org/10.1016/j.jterra.2015.12.003
Author.
Shemetov, P., Rubtsov, C., & Shlylov, A. (2005). Exploitation of rope
Komatsu Mining Germany GmbH (2016). PC8000-6 hydraulic excava- shovels and hydraulic excavators at Muruntau open pit. Mining Indus-
tor (Product specification). Düsseldorf, Germany: Author. Retrieved try, 5 (Published online).
Performance prediction of a full-size rapid dampening suspension

in an ultra-class hauler application
R. K. Soni, T. G. Joseph, and M. Curley
ABSTRACT Rapid dampening suspension is a proposed suspension design for ultra-class haulers that
is intended to avoid metal-to-metal contact inside a strut and under shock loading, thus improving ride
quality and ameliorating fatigue loading issues on a hauler structure. Laboratory testing has shown the
ability of the proposed design to reduce the opportunity for bottoming out, that is, metal-to-metal contact
between the colliding ends of oleo-pneumatic, two-piece suspension struts during the compression cycle.
This paper predicts the performance of a full-size rapid dampening suspension strut by applying scaling
rules to test the proposed design based on a set of field data from a Caterpillar 797B ultra-class haul truck
operating in the oil sands region of the province of Alberta, Canada. The predicted performance of the
suspension strut is compared to that of the original equipment manufacturer model, which was monitored
on board the truck during field data collection.
■ KEYWORDS Haul truck, Original equipment manufacturer (OEM), Rapid dampening, Strut, Suspension
RÉSUMÉ La suspension à amortissement rapide est une conception de suspension proposée pour des
camions de transport de catégorie ultra-class (des engins dont la charge utile dépasse les 320 tonnes
métriques), visant à empêcher un contact métal sur métal à l’intérieur d’une jambe de suspension et dans
des conditions de choc au chargement. Elle améliore la qualité du roulement et les problèmes de charge
de fatigue sur la structure d’un camion de transport. Les essais en laboratoire ont confirmé que la concep-
tion proposée évitera au cylindre interne de heurter le cylindre externe, autrement dit empêchera le contact
métal sur métal des jambes de suspension oléopneumatique en deux parties pendant le cycle de compres-
sion. Cet article prévoit la performance d’une jambe de suspension à amortissement rapide à taille réelle
en appliquant les lois d’échelle afin de mettre à l’essai la conception proposée, sur la base d’une série de
données de terrain relatives à un camion de transport 797B de catégorie ultra-class de Caterpillar
(aujourd’hui remplacé par le nouveau modèle 797F) utilisé dans la région des sables bitumineux de la pro-
vince de l’Alberta, Canada. On compare la performance prévue de la jambe de suspension à celle du
modèle du fabricant d’équipement d’origine (FEO), qui a été contrôlée à bord du tombereau pendant la
collecte des données de terrain.
■ MOTS CLÉS amortissement rapide, camion de transport, fabricant d’équipement d’origine (FEO), jambe
de suspension, suspension
INTRODUCTION
Research basis for the paper

This paper is based on the unpublished MSc thesis work scale performance indications highlighted here. Although
of El-Sayed (2003), Santos (2007), and Soni (2009) under this paper uses a specific Caterpillar 797B suspension strut
the direction of Dr. Tim Joseph within the oil sands–equip- for field monitoring and as a basis for scale prediction of a
ment interactions program (OsEIP). OsEIP is part of the modified rapid dampening suspension strut design, the
Alberta equipment–ground interactions syndicate (AEGIS) design scaling principles used here could be applicable to
surface mining research group, School of Mining and any such design change consideration for an oleo-pneu-
Petroleum Engineering, University of Alberta. The afore- matic suspension system.
mentioned work spanned a decade of progressive studies Ultra-class (> 290 tonne capacity) haulers operating on
into rapid dampening suspension systems, leading to the oil sands and other soft ground conditions in northern

174 R. K. Soni, T. G. Joseph, and M. Curley
Alberta, Canada, are routinely subjected to high-load ure 1). The modified oleo-pneumatic, gas spring design
events of short duration, especially during the summer. included a new “poled” conical pin that progressively closed
Existing truck suspension systems cannot absorb high- a static orifice and restricted fluid flow, thereby greatly
intensity shock loads, resulting in a higher probability of enhancing strut stiffness performance and reducing the like-
metal-to-metal contact between the colliding ends of con- lihood of bottoming out events.
ventional oleo-pneumatic, two-piece suspension struts dur-
ing the compression cycle. Frequent occurrence of this METHODS
phenomenon, known as “bottoming out,” can result in early
failure of truck structural components, often attributed to Rapid dampening suspension design overview
rack (frame twisting) motions (Joseph, 2003). Moreover, The proposed rapid dampening strut design permits
the whole body vibrations and jarring effect generated from operational action as a semi-active suspension and pro-
all motions—but exacerbated by bottoming out—can cre- duces a continuously increasing dampening coefficient in
ate severe health and safety issues (e.g., lower back injury) progressive compression via proportionally closing an ori-
for truck operators (Berezan, Joseph, & del Valle, 2004). fice. The design is a simple retrofit of an OEM passive sus-
The bottoming out phenomenon was identified at the Uni- pension system consisting of a conical metering pin and an
versity of Alberta as an industrial problem that led to a pro- orifice plate installed into a two-piece OEM strut (Fig-
posed new suspension design termed “rapid dampening ure 2). The modified strut used in scale testing of the prin-
suspension.” The design is based on the concept of a displace-
ment-dependent variable orifice originally proposed by El-
Sayed (2003), validated by Santos (2007), and extended by
Soni (2009). El-Sayed first defined a mathematical model to
simulate and compare the performance of the proposed design
to that of a conventional suspension design. This research
demonstrated that the variable-orifice shock absorber provided
high dampening forces that depend on the rate of input dis-
placement via piston velocity. Santos (2007) validated the pro-
posed suspension system through laboratory tests of a scaled
version of the original equipment manufacturer (OEM) con-
figuration and a scale-matched prototype rapid dampening
strut. At this stage of concept development, no gas spring was
used, and the system relied solely on fluid transfer to generate
dampening forces. The simple damper—modelled at scale and
based on an existing haul truck suspension design—was mod-
ified to accept a variable orifice in the cylinder. Compared to Figure 1. Dampening forces at 414 kPa and 0.5 Hz for two original
the OEM configuration, the modified strut provided improved equipment manufacturer (OEM) and two modified suspension strut designs
dampening forces towards the
end of the stroke as the effective
orifice area was reduced, which
agreed with independent indica-
tions from Currey (1988). The
increasing dampening force as a
function of displacement
reduced the probability of bot-
toming out events.
Primarily based on the work
of Soni (2009), this paper
expands upon the previous
research through laboratory
tests using scaled struts charged
with and without nitrogen gas.
The combination of a metering
pin and nitrogen gas is postu-
lated to greatly increase the
dampening force as a function
of piston displacement (Fig- Figure 2. Suspension strut used for laboratory scale test

Performance prediction of a full-size rapid dampening suspension in an ultra-class hauler application 175
ciple is a double-acting hydraulic cylinder capable of sus- the stiffness of the scaled rapid dampening suspension was
taining internal pressures up to 20.7 MPa, with a bore diam- determined as the derivative of the strut force-deformation
eter of 102 mm and stroke length of 170 mm. curve. Figure 4 shows the variance of the modified suspen-
The operating principle of the modified strut is as fol- sion strut stiffness throughout a single compression–
lows (Figure 3): rebound duty cycle. The maximum stiffness
• During a complete stroke, the metering pin (10) moves (corresponding to a maximum dampening force) occurred
in (compression) and out (rebound) of a central orifice at the end of the compression stroke. During the transition
(8), varying the orifice available flow area to restrict from compression to rebound, the stiffness theoretically
fluid flow. approached an infinite value because a change in deforma-
• This causes a gradual increase in the dampening force dur- tion is zero when an applied force shifts from compressive
ing the compression cycle and a gradual decrease during to tensile action. The stiffness calculation was thus
rebound retraction, which yields a variable dampening restricted, removing points causing very high values.
coefficient. The effective orifice area is at a
maximum at the beginning of the compres-
sion stroke and decreases to a minimum as
the compression stroke is completed, when
the dampening coefficient would reach a
maximum value.
• The orifice area completely depends on
the metering pin (10) position. After
insertion of the metering pin, the central
orifice effective flow area was designed
to remain less than the sum of the opening
areas associated with the OEM check ball
valve (6) and orifice (8). This relationship
ensures the dampening characteristics of
the strut are principally controlled by the
metering pin effect.
Scale laboratory tests

To quantify the dampening characteristics
of the proposed design and compare perfor-
mance to that achieved by Santos (2007), the
Figure 3. Proposed full-size modified suspension strut design: 1-rod, 2-nitrogen chamber, 3-
scale laboratory tests of Soni (2009) were per-
housing cylinder, 4-cavity, 5-oil chamber, 6-check ball valve, 7-cavity, 8-orifice, 9-plate with
formed on modified and OEM design config- orifice, 10-metering pin
urations charged with and without nitrogen
gas. The setup consisted of a strut, a metering
pin, pressure transducers, a data acquisition
system, and a cyclical loading frame. The
strut tested was modified from that used by
Santos (2007) by using a cylindro-conical
metering pin mounted at the base of the cylin-
der housing. A clearance of 12.7 mm was
maintained between the base of the cylinder
and piston head at maximum closure in order
to avoid metal-to-metal contact at the end of
the compression stroke, permitting a focus on
the oleo-pneumatic reactions.
Rapid dampening strut stiffness

scaling
Predicting the rapid dampening perfor-
mance of a full-scale modified suspension
system required understanding the stiffness
of the scale modified strut. In the laboratory, Figure 4. Stiffness for the modified strut for one cycle charged at 414 kPa and 0.25 Hz

For comparison, Figure 5 shows the stiffness curve of

the scaled OEM strut configuration, effected with no rapid
dampening metering pin installed in the same system.
( KOEM,F
KOEM,S) (
–––––––– =
Closure
KMOD,F
)
––––––––
KMOD,S
Closure
(1)
Predicting the performance of the full-scale rapid where K is the stiffness of the system (kN/m), F and S
dampening suspension denote full and scale size, respectively, and MOD denotes
The performance was tested of three suspensions (OEM, the modified design.
scaled OEM, and scaled modified). The performance of a Equation 2 compares designs while keeping the size the
full-size modified suspension was then predicted by apply- same at the same percentage of closure.
ing the scaled modified test results to field data acquired
from the OEM suspension system of a CAT 797B haul
truck operating in the Alberta oil sands region. The known
performances of the scaled OEM and modified struts were
( KMOD,S
KOEM,S) (
–––––––– =
Closure
KMOD,F
)
––––––––
KOEM,F
Closure
(2)
correlated, and the relative performance ratios were applied

to the CAT 797B OEM suspension data. In terms of dampening force, the stiffness ratio is stated
The domain assumptions used in this prediction are by equation 3:
detailed in Figure 6. Two factors, size and design, vary
FOEM,F
when predicting a full-size modified configuration. These
are essentially two domains: a size domain (scale to full- ( KOEM,F
–––––––– ) = e
–––––––– / =
FOEM,F
–––––––– (3)
FOEM,S FOEM,S
size) and a design domain (no-pin to pin). The fundamental
KOEM,S
Closure e /
assumption is that any performance advancement achieved
for a scale change would follow the same size increase rule where e is an equal system closure strain percentage. The
for either the OEM or scaled modified configuration. stiffness ratio is then equal to the dampening force ratio of
The performance parameter chosen for comparison was the suspension for the same imposed strain.
stiffness because it represents the geometric and internal Field data recorded via the OEM onboard system were
reaction of each system. Hence, a comparison would used to evaluate the pressures inside the four suspension
include a size and charging pressure factor. It would also struts mounted on the CAT 797B haul truck monitored in
reflect the design change from introducing the metering pin the field at a 1 Hz recording frequency. Figure 7 outlines
through comparing the scale performance of the OEM to the pressure response of the right rear suspension as an
modified configurations. The following hypothesis was example. At a stationary tare weight condition and when
tested based on strut stiffness: A scaling factor based on a charged with nitrogen gas, the CAT 797B rear suspension
stiffness ratio for two given sizes will be a constant for any has 0.152 m available stroke, during which the gas under-
common design regardless of relative size. This is mathe- goes isentropic compression during in-motion loading
matically expressed as equation 1, which compares size activity. The static unloaded pressure inside the strut when
while keeping the design the same at the same percentage the truck is stationary with no payload—representing the
of closure. Closure is defined as a percent closure achieved initial isentropic state system—was approximately 2 MPa.
through the stroke of a given suspension design. Using equation 4, the compressed length of the nitrogen
Figure 5. Stiffness of the scaled original equipment manufacturer strut for Figure 6. Approach to predict performance of a full-size modified
one cycle charged at 414 kPa and 0.25 Hz suspension; OEM: original equipment manufacturer

column in the cylinder and the external force

were determined from the recorded data.
PV γ = Constant (4)
where P is pressure (kPa), V is volume (m3) at

any instant of evaluation, and γ is the isentropic
gas constant (1.39 for nitrogen).
For two initial (reference) and at any instant
pressure states, the stroke positions are detailed in
equation 5.
1
P0 /γ
( )
Lx = –––– · L0
Px
(5)
Figure 7. Pressure response for the right rear suspension on a CAT 797B haul truck
where P0 and Px are the initial and instant pres-
sure states (kPa), respectively, and L0 and Lx are
the corresponding stroke positions (m).
RESULTS
Using equation 5, the stroke position was
determined for each data point in Figure 7, per-
mitting the generated force to be evaluated. It is
evident that the force follows a negative power
relation with respect to stroke (Figure 8), which
can be extrapolated to any stroke position.
Figure 9 shows the available remaining
stroke for a full-size OEM CAT 797B strut cal-
culated from the recorded field data. What is
evident is the very small magnitude of remain-
ing stroke, such that metal-to-metal contact is
imminent during loaded motion of the haul Figure 8. Force vs. stroke length for the CAT 797B right rear suspension
truck.
The percentage system closure was evaluated
as an independent variable proportional to the
available strain and as a dimensionless quantity to
be independent of suspension strut scale (Fig-
ure 10). Following from the hypothesis illustrated
in equation 1, the trend of the force ratio to per-
centage closure for the modified strut was
assumed to then follow the same trend as the
OEM design from scale- to full-size. It follows
from the hypothesis that the force ratio of the
modified strut may be written as equation 6:
FMOD,F = 106P –1.39FMOD,S (6)
where P is the percentage closure and FMOD,F and

FMOD,S are dampening forces for the full-size and Figure 9. Available suspension stroke length for a CAT 797B suspension strut
scale modified struts, respectively. It is interesting
to note that the power 1.39 remains unchanged and is the the scaled-modified strut using equation 6. The force
isentropic constant (γ for the nitrogen gas used. response of the two configurations appear similar until 33%
The dampening force response for the full-size modified closure remains, at which point the modified dampening
suspension was estimated from the laboratory test data for force increases exponentially compared to the OEM con-

figuration (Figure 11). The difference between the

two responses increases as the compression peaks.
The predicted peak dampening force of the OEM
strut at the end of compression is approximately
27,000 kN, while the modified design achieves a
much higher force (in the 106 kN range), thus effec-
tively avoiding metal-to-metal contact of the suspen-
sion under severe loading conditions.
Using the data from Figure 11 and the dampening
force from the field dataset, the available remaining
closure of a modified full-scale strut was predicted.
It is evident that for higher suspension forces and
deformations corresponding to high strut pressures,
the modified strut responded to the same forces, pro-
portional to the pressures in Figure 7, with less clo-
sure (Figure 12). This result shows the predicted Figure 10. Force ratio relative to percentage suspension closure for the original
ability of the modified design to resist higher metal- equipment manufacturer (OEM) configuration
metal bottoming out contact shocks by providing
higher dampening force per unit strain.
CONCLUSIONS
Based on the laboratory test results and corre-
sponding comparison to the field data, two principal
conclusions can be made.
1. While the stiffness of the OEM configuration
remained relatively constant throughout the
cycle, the stiffness of the modified configuration
increased during the stroke.
2. The performance of a modified rear suspension of
the CAT 797B haul truck was estimated from field
and laboratory test data. The predicted dampening
force effectively increased at 33% closure and
overcame the limiting dampening force of the
OEM configuration at 0.127 m of closure. Modi-
fied suspension peak values reached dampening Figure 11. Predicted force responses of the full-scale modified and original
forces of 106 kN, roughly 14 times the peak force equipment manufacturer (OEM) struts with respect to stroke
achieved by the OEM configuration, but with no
metal-to-metal contact.
The work performed here creates an opportunity
for OEMs and mine operators to weigh the benefits
of an alternate higher dampening suspension design.
Reducing shock through higher dampening capabil-
ity could effectively reduce suspension rapid closure
shock events that would otherwise translate to higher
jarring and cumulative whole body vibration events
(Berezan et al., 2004) or to unbalanced four-point
suspension events that give rise to rack, which
causes structural frame fatigue (Joseph, 2003).
This work proposes a simple, rapid dampening
suspension design, achieved through low-cost scale
research. More importantly, the scale testing
approach illustrated here provides a mode of low-
cost design investigation for other suspension con-
figurations. Figure 12. Predictive closure comparison between full-scale designs

vice-president to the CIM, the president and principal engineer of JPI Mine
Paper reviewed and approved for publication by the Equipment, and recognized as an expert in mine equipment performance
Maintenance, Engineering and Reliability Society of CIM. related to ground–equipment interactions. He delivers certification industry
short courses and mining equipment performance evaluations to the global
Rakesh Soni graduated in 2006 from the Indian Institute of Technology, resource industry, mining equipment manufacturers, and service providers.
Dhanbad, with a BTech in mining-mechanical engineering. He completed his He is a former recipient of the CIM Distinguished Service Medal for his
MSc in mining engineering in 2009 from the University of Alberta under the contributions to the global mining industry and mining engineering
supervision of Dr. Joseph. His thesis research was on alternative rapid education and a recipient of the CIM Past President’s Memorial Medal for
dampening suspension systems for mining haul trucks. From 2009 to 2013, early achievements in education to the global mining industry. Dr. Joseph
he worked for oil sands mining companies in northern Alberta before he will retire from the University of Alberta in 2020 but will continue to
cofounded LoginRadius, which is based in Vancouver. He is currently the practice through JPI Mine Equipment for the foreseeable future.
CEO of LoginRadius, specializing in assurance of digital identities for tjoseph@ualberta.ca
corporations worldwide.
Mark Curley, MSc, EIT, is a graduate of the University of Alberta mining
Tim G. Joseph, PhD, P.Eng., FCIM, is associate dean of engineering and engineering BSc (2013) and MSc (2016) programs. As a mining engineer in
professor of mining engineering at the University of Alberta, and director of Alberta, he works for BGC Engineering in Edmonton on international
the Alberta Equipment–Ground Interactions Syndicate (AEGIS) research mining projects, with considerable time spent in the Dominican Republic
program. He is vice-president of the Association of Professional Engineers expanding exploration targets. He has also worked directly for Dr. Joseph at
and Geoscientists of Alberta (APEGA), an active ambassador and former the University of Alberta as a research assistant.
REFERENCES
Berezan, J. J., Joseph, T. G., & del Valle, V. D. (2004). Monitoring whole Joseph, T. G. (2003). Large mobile mining equipment operating on soft
body vibration effects on ultra-class haulers. CIM Bulletin, 97(1082), 1–4. ground. Proceedings of the 18th International Mining Congress and
Exhibition of Turkey (IMCET 2003), Antalya, Turkey, 143–147. Retrieved
Currey, N. S. (1988). Aircraft landing gear design: Principles and prac- from http://www.maden.org.tr/resimler/ekler/ad21787ac6487a4_ek.pdf
tices. Washington, DC: American Institute of Aeronautics and Astronau-
Santos, R. F. (2007). A rapid dampening suspension for ultra-class haulers
tics. https://doi.org/10.2514/4.861468 (Unpublished master’s thesis). University of Alberta, Edmonton, AB.
El-Sayed, M. (2003). Suspension system in large mining trucks (Unpub- Soni, R. K. (2009). Scaled rapid dampening suspension evaluation
lished master’s thesis). University of Alberta, Edmonton, AB. (Unpublished master’s thesis). University of Alberta, Edmonton, AB.
180 MAINTENANCE, ENGINEERING AND RELIABILITY
Is bigger still better?

Considerations in increasing the size of haulage equipment
M. Dotto, T. G. Joseph, and M. Curley
ABSTRACT Equipment selection is a key decision in mine planning. Equipment size affects decisions
from pit size to the total operation cost. Particularly when replacing haulage fleets for long-life surface min-
ing operations, mining operations focus solely on haulage productivity. Road width and specifically ramp
width are significantly affected by an increase in hauler size in terms of overall strip ratio through required
highwall set-backs or sterilizing of lower ore access due to required overlying ramp fill structures. In addi-
tion, in light of ever increasing regulatory and reporting emissions requirements, bigger equipment may not
translate into lower emissions per tonne moved. This study analyzes hauler-scale impacts on aspects not
currently incorporated into conventional mine planning, including expansion of roads to accommodate
larger equipment, road layer thickness variation depending on hauler size, and fuel consumption and emis-
sions. Results indicate that such relationships have a large impact on cost. Therefore, an expanded consid-
eration of the capacities of increasing equipment size in mine planning is highly recommended.
■ KEYWORDS Equipment selection, Grade resistance (GR), Haul road, Haul truck, Mine planning, Payload,
Rolling resistance (RR), Total resistance (TR)
RÉSUMÉ La sélection du matériel est une décision importante dans l’aménagement des mines. La taille
de l’équipement a une incidence sur la prise de décisions, depuis la taille de la fosse jusqu’aux coûts totaux
des activités. L’exploitation minière se concentre uniquement sur la productivité des moyens de transport,
particulièrement lorsqu’il s’agit de remplacer les parcs de camions de transport dans les exploitations à ciel
ouvert à longue durée de vie. La largeur des routes, et spécifiquement la largeur des descenderies sont consi-
dérablement affectées par une augmentation de la taille des camions de transport en termes du coefficient
global de recouvrement qui dépend de la taille requise des parois (si les descenderies sont agrandies, les
parois devront être creusées de haut en bas afin d’agrandir la taille du fond de la fosse). Il faut aussi pouvoir
procéder à l’excavation des gradins les plus profonds de la mine et ainsi préserver l’accès à cette zone, ce
qui pourrait devenir difficile si la quantité de stériles utilisés pour la structure des descenderies augmente et
s’ils occupent une trop grande partie du fond de la fosse. En outre, au vu des exigences de plus en plus
strictes en matière de réglementation et de déclaration des émissions, un équipement plus volumineux ne
sera pas nécessairement synonyme d’une réduction des émissions par tonne déplacée. Cette étude analyse
l’incidence de la taille des camions de transport sur des aspects actuellement non intégrés dans l’aménage-
ment classique des mines, notamment la création de routes permettant le passage d’un équipement plus
volumineux, la variation de l’épaisseur de la couche d’asphalte des routes en fonction de la taille des
camions de transport, ainsi que la consommation de carburant et les émissions. Les résultats indiquent que
ces relations ont des répercussions importantes sur les coûts. Ainsi, on recommande d’accorder une atten-
tion accrue aux capacités d’augmentation de la taille de l’équipement dans l’aménagement des mines.
■ MOTS CLÉS aménagement des mines, camion de transport, charge utile, résistance au roulement,
résistance due aux rampes, résistance totale, route de transport, sélection du matériel
INTRODUCTION
Over the past five decades, the mining industry has pro- often bigger equipment iteration, while older mines are re-
gressively attempted to increase production rates through evaluated to keep pace with upgraded mixed fleets.
the use of larger haulage equipment. Each new generation Whereas production gains are often realized with larger
of mine is designed to operate with the most recent and equipment, there are additional aspects to consider when

Is bigger still better? Considerations in increasing the size of haulage equipment 181
evaluating the feasibility of moving to increased equipment This paper utilizes the mechanistic approach described
size. This paper considers the impact of an increase in by Thompson (2011), taking into account the differing mate-
hauler size and the consequences on corresponding haul rial properties of each road construction layer, thereby lim-
road geometry, rolling resistance (RR), and fuel usage. iting load-induced strain in progressively thicker sub-grades
to below an empirically derived critical value. Based on
Haul roads field observations, maximum vertical strain limits have been
One of the greatest impacts of increasing hauler size is established as 1,500–2,000 microstrain (µm/m or µε) for
the impact on the size of haul roads, specifically the width typical haul roads. Values exceeding 2,500 µε are associated
dictated by safe operating practices. This subsequently with unacceptable structural performance, except for light
affects highwall/footwall set-back, which increases strip- traffic and short-term roads (Thompson & Visser, 1997).
ping ratios and thus the volume of material that needs to be
moved to access an orebody in a surface mine. In general, RESEARCH METHODS
the higher the size class of haul truck, the greater the To illustrate the effect of truck size on road requirements,
advanced efficiency of the engine to haul loads. However, an example haul road was evaluated for each of three truck
compounded with a higher hauler weight, any increase in size classes. Relevant equipment parameters used in this
RR has a negative and often overwhelming impact on the evaluation are shown in Table 1. For comparison purposes,
ability of loads to be hauled up ramps to exit pit locales. each of the three roads was evaluated using typical haul road
construction materials to meet equivalent arbitrary produc-
Haul road structural design tion requirements using each of the three truck options. The
At any truck and shovel mining operation, the haul roads effect of truck size on road construction was quantified from
serve as an integral arterial part of the haulage system; it is the difference in layer thickness and overall material require-
imperative to productivity that the roads are designed to ade- ments. The proposed roads comprised the following layers: a
quately support loaded haul trucks. Although larger trucks are crushed rock surface, a pit-run base layer, and a mine
capable of larger payloads per cycle, it is logical that heavier spoils/till sub-base on a firm clay sub-grade. The example
vehicles require a more substantial road structure. Histori- construction material properties are given in Table 2.
cally, off-highway roadways have been designed using the
California Bearing Ratio (CBR) method (Kaufman & Ault, Applying a critical strain methodology
1977). The CBR is the ratio between the bearing capacity of Using the method presented by Thompson (2011), each
a given soil and the bearing capacity of a standard-crushed of three proposed roads was designed to haul 40 Mt/y for
rock. Empirical curves then relate the required fill thickness five years. Truck cycle time was set as 30 minutes and was
and applied wheel load to the CBR (Kaufman & Ault, 1977). assumed to be independent of truck size; truck availability
For road design, the technique is used such that each succes- and utilization were assumed to be 90% each. Table 3
sive layer should be of higher CBR than the preceding layer shows the critical strain road response calculations using
for a given wheel load (Thompson, 2011). The CBR method Thompson’s method, as well as the fleet size requirements
is not optimal for a mine haul road and has largely been super- to meet annual production. Note the critical strain limits
seded by the mechanistic design method presented by presented in Table 2 are reduced by 20% to reflect higher
Thompson (2011), a recognized author in this journal. road stresses induced by payload overloading following the
Table 1. Hauler specifications; GW: gross weight (Caterpillar, 2012)
Truck Width Length Empty Payload Vehicle Payload: Net power Tire size *Surface
(m) (m) weight (t) GW vehicle (hp) ground stress
(t) (t) GW (kPa)
CAT 785C 6.3 11 102 136 249.4 0.57 1,348 33.00R51 523
CAT 793F 8.3 13.7 163 227 390.0 0.58 2,478 40.00R57 626
CAT 797F 9.5 14.8 260 363 623.6 0.58 3,793 59/80R63 754
* At 1 g loading
Table 2. Construction material specification
Layer Typical material California Bearing Ratio Resilient modulus

(%) (MPa)
Surface Crushed rock 95 330
Base Pit-run 60 245
Sub-base Till mine spoil 25 130
Sub-grade Firm clay 4 40

182 M. Dotto, T. G. Joseph, and M. Curley
Table 3. Critical strain estimation for three proposed roads; GW: gross weight
Truck Max 20% GW No. units Cycles/24 h Production, fixed Critical strain
overload (overload) (kt/24 h) (µε)
CAT 785C 163.2 265.4 17 672 175 1,506
CAT 793F 272.2 435.5 10 403 175 1,730
CAT 797F 435.6 696.3 6 252 175 1,964
10:10:20 rule, as well as dynamic hauler

motions. This reduction was suggested by
Thompson (2011) and supported by previ-
ous evidence from Joseph (2002, 2003)
and Sharma (2009).
Using a Boussinesq ground stress anal-
ysis model, the layer thickness and overall
road design required to support each of the
truck types was calculated independent of
the number of cycles, where construction
layer thickness was affected by the weight
of the truck. Comparative results for the
road design suitable for each truck size are
presented in Figure 1. A required road
thickness increase of 20% transitioning
from models 785C to 793F, 18% from
793F to 797F, and 43% from 785 to 797F
Figure 1. Comparative road requirements by equipment type
was determined for each change in truck
size usage. While this increase in required
road thickness is substantial, mines that can justify the tran- which requires larger construction volumes moved to
sition stand to gain payload increases of 74% moving from develop ramp systems for ingress and egress. To study the
the 130 t 785C to the 230 t 793F and 59% from the 793F to effect of truck size on haul road geometry and construction
the 360 t 797F. In extreme cases, mines that move to tran- volume, the design of a haul road for a 100 m deep pit was
sition from the 785C directly to the 797F experience a examined. Initial assumptions were that the pit had an over-
177% payload per cycle return on the volumetric change all slope of 70°, a ramp gradient of 8%, and maximum
investment required for 43% thicker roads, plus equipment anticipated downhill vehicle speed of 30 km/h (based on
capital and maintenance-related costs. recommended tire operational duty). A single switchback
was included in the ramp design for each truck.
Impact of equipment size on road geometry To determine an appropriate road width, switchback radius,
Equipment size has a pronounced effect on the vertical and switchback length, the approach of Thompson (2011)
haul road construction profile through the gross vehicle was used, whereas a method derived by Bozorgebrahimi
weight. The relationship between equipment width and the (2004) was implemented to evaluate the effect of equip-
haul road geometry required to provide a safe and efficient ment width on overall pit slope. Results of the analysis are
travel-way is also of critical importance. It is very probable presented in Table 4, providing a sense of the amount of
that an increase in equipment size requires a three-dimen- additional waste that must be stripped to widen a road to
sional increase in road geometry. This can result in a mine support a change in equipment size. Apart from the
haulage route targeting planned lower overall pit slopes, increased stripping to accommodate an increased road size,
Table 4. Ramp design by equipment type

Truck Road Truck Switchback length Overall pit slope
design width turn (m) (°)
(m) radius
(m)
CAT 785C 26 150 471 70.0
CAT 793F 34 159 500 61.5
CAT 797F 40 166 531 55.5

pit bottoms must maintain ample space to

allow safe operation of trucks and corre-
sponding loading equipment. In practice,
stripping ratios will likely be much greater
than indicated, such that the reported values
should be taken as minimum suggested
requirements.
Presented in Figure 2 are the construc-
tion material requirements for an in-pit haul
road (including ramps) by truck size. Exten-
sions of mine roads to waste dumps and/or
crusher/stockpiles as destinations would
also need to be constructed, likely further
increasing the volumes. The distance
depends on the dump locations, and there-
fore the construction volumes in Figure 2
will likely be vastly higher. The cost of Figure 2. Total construction material by equipment size
preparing (crushing) construction materials
may vary from one operation to another
depending on material quality and availability. When Table 5. Comparative total construction volumes
deciding on truck size, the impact of road construction Construction volumes (%) for three truck models
materials must also be considered in order to determine the Layer 785C to 793F 793F to 797F 785F to 797C
overall cost of increasing truck size at an operation. As sup- Surface 67 48 146
ported by both Figure 2 and Table 5, the economic benefit
Base 58 37 117
of an increase in equipment size must provide the operation
Sub-base 58 39 131
with a substantial economic return to offset the increase in
haulage route construction and later maintenance costs.
Fuel consumption and diesel emissions

In addition to the capital costs associated with hauler size
variation, a mining operation must consider how variation in
hauler size will affect fuel consumption and emissions. One
of the disadvantages of truck and shovel mining systems is
high diesel consumption and combustion emissions, partic-
ularly in a far more focused emissions-conscious industry
where emissions credits have come into play post-2010. In
operations such as those found in the oil sands region of
northern Alberta, a large percentage of NOx emissions is
attributed to mine fleets (Singh, Jain, Onder, & Unrau,
2008). However, existing literature suggests that an increase
in haul truck payload results in lower fuel consumption per
ton moved, as supported by Figure 3 (Leslie, 2000). Figure 3. Relationship between truck fuel consumption per ton and
Although the payload-scaled efficiency is suggested, it is payload derived from original equipment manufacturer specification
important to note that much of this is due to efficiency sheets; after Leslie (2000)
achievements associated with each new “tier” of engine
developed and proven; thus at least some of this effect is due RESULTS AND ANALYSIS
to the general evolution of diesel engine controls.
Table 6 lists fuel consumption and emissions data for The effect of size and total resistance on fuel
various payload classes of haul truck (Leslie, 2000). Singh consumption and emissions
et al. (2008) conducted research to determine diesel emis- From the NOx and CO relationships shown in Figure 4, the
sions from mobile equipment in field operations in the relationship in Figure 5 was discerned between fuel consump-
Athabasca oil sands region of Alberta. The analysis consid- tion and total resistance (TR; sum of grade resistance [GR]
ered equipment size, actual operating conditions, and oper- and RR). A 1% increase in TR results in a fuel consumption
ating modes, from which the relationships in Figure 4 were increase of 0.09 (L/h)/ton for a CAT785C, 0.07 (L/h)/ton for
established. a CAT793F, and 0.04 (L/h)/ton for a CAT797F.

NOx, hydrocarbon, CO, and particulate matter maximum expected RR are enveloped. Considering a GR
emission estimates of 8%, the operating RR varies from 4 to 10%. Above this
Diesel emissions per ton decrease with increasing truck range, the rate of increase in emissions is much faster.
size due to higher haulage capacity and engine fuel effi- Comparing both NOx and HC emissions, as well as CO
ciency. Relationships between NOx and hydrocarbon (HC) emissions, a mean decrease of 21% was seen moving from
emissions and TR verify this (Figure 6). It is evident—and models CAT785C to CAT793, 45% from CAT793F to
not unexpected—that an increase in TR increases fuel con- CAT797Fs, and 57% from CAT785C to CAT797F (Figure
sumption per ton of hauler capacity and that this is 6). Further, CO emissions per ton were higher on smaller
inversely related to truck size. Here, flat ground, ramp, and trucks than larger trucks (Figure 7).
Table 6. Fuel consumption and emissions data (Leslie, 2000 and original equipment manufacturer specification sheets); HC: hydrocarbon, PM: particulate matter
Hauler payload Fuel Fuel Power NOx CO HC PM
(short ton) (L/h)/t (L/h) (hp) (g/hp-h) (g/hp-h) (g/hp-h) (g/hp-h)
240 1.96 463 2,337 4.8 2.6 4.8 0.150
320 1.64 515 2,700 2.3 0.7 2.0 0.100
345 1.42 491 3,188 1.0 0.7 0.5 0.075
400 1.09 436 3,793 0.5 0.5 0.5 0.025
Figure 4. Relationship between NOx and CO emissions and total Figure 5. Relationship between fuel consumption per ton moved and total
resistance, TR; after Singh et al. (2008) resistance (TR) for three truck sizes
Figure 6. Relationship between NOx and hydrocarbon (HC) emissions Figure 7. Relationship between CO emissions per ton payload and total
and total resistance (TR) for three truck models; RR: rolling resistance resistance (TR) for three truck models; RR: rolling resistance

Similar to the relationships presented above, particulate • Increasing TR increased fuel consumption and related
matter (PM) emissions per ton capacity were negatively emissions. The impact of increasing TR was proportion-
related to truck size and positively related to TR (Figure 8). ally higher for smaller truck sizes.
The relationships are effectively linear, with an average The authors are reminded of a statement by Kim Bar-
percentage decrease in PM emissions with increasing truck rowman made almost two decades ago when he introduced
size of 24% moving from models CAT785C to CAT793F, the phrase “roads make loads.” His words are as true today
48% from CAT793F to CAT797F, and 58% from CAT785C as they were then.
to CAT797F. The PM emissions are an order of magnitude
smaller than CO emissions and approach two orders of “Taking care of road conditions and maintenance
magnitude smaller than HC emissions. This emphasizes the to keep rolling resistance low does ensure that fuel
effectiveness of newer engine generations in controlling usage and emissions remain low. There are pros
PM emissions. and cons to increasing hauler size. While gains in
productivity are to be had with potentially lower
fuel usage and emissions; the cost of road con-
struction, maintenance and overall waste material
movement to ensure safety and highwall stability
need to be considered in the balance of decision.”
An earlier draft of this paper was published in the

Proceedings of the Maintenance, Engineering and
Reliability/Mine Operators Conference (MEMO 2017)
prior to undergoing the CIM Journal peer-review process.
Paper reviewed and approved for publication by the

Maintenance, Engineering and Reliability Society of CIM.
Magreth Dotto, M.Sc., P.Eng., is currently a PhD student at the University
of Alberta under the direction of Dr. Joseph and Dr. Yashar Pourrahimian,
researching energy optimization in mine planning. She has previously
Figure 8. Relationship between particulate matter (PM) emissions per ton worked in academia as an assistant lecturer (2014–2017) at the University
payload and total resistance (TR) for three truck models of Dar es Salaam and as a mining engineer (2009–2010) at the Geita Gold
Mine, part of the Anglo Gold Ashanti group. She earned her BSc in mining
engineering from the University of Dar es Salaam in 2009. She is a
CONCLUSIONS professional mining engineer registered with the Engineers Registration
Board (ERB) in Tanzania.
The notion that “bigger is better” regarding equipment
size has been the conventional wisdom for several decades Tim G. Joseph, PhD, P.Eng., FCIM, is associate dean of engineering and
in the mining industry. Mine operators are cognizant of professor of mining engineering at the University of Alberta, and director of
the Alberta Equipment–Ground Interactions Syndicate (AEGIS) research
opportunities to phase out smaller trucks and employ bigger program. He is vice-president of the Association of Professional Engineers
units with new and expanded projects. The analysis pre- and Geoscientists of Alberta (APEGA), an active ambassador and former
sented here shows that, while an increase in hauler size pro- vice-president to the CIM, the president and principal engineer of JPI Mine
Equipment, and recognized as an expert in mine equipment performance
vides an increase in material haulage capacity, the following related to ground–equipment interactions. He delivers certification industry
additional costs and benefits should also be recognized: short courses and mining equipment performance evaluations to the global
• The haul road widths and pit sizes needed to accommo- resource industry, mining equipment manufacturers, and service providers.
He is a former recipient of the CIM Distinguished Service Medal for his
date larger equipment result in larger quantities of waste contributions to the global mining industry and mining engineering
stripping. Additional road bed materials and costs will education and a recipient of the CIM Past President’s Memorial Medal for
be required to construct roads to accommodate larger early achievements in education to the global mining industry. Dr. Joseph
will retire from the University of Alberta in 2020 but will continue to
trucks. practice through JPI Mine Equipment for the foreseeable future.
• The rate of increase of RR was the same regardless of tjoseph@ualberta.ca
hauler size because RR is a material property.
Mark Curley, MSc, EIT, is a graduate of the University of Alberta mining
• In general, all emissions categories evaluated here engineering BSc (2013) and MSc (2016) programs. As a mining engineer
decreased with increasing truck size. Fuel consumption in Alberta, he works for BGC Engineering in Edmonton on international
and emissions were reduced on a per-ton by payload mining projects, with considerable time spent in the Dominican Republic
expanding exploration targets. He has also worked for Dr. Joseph at the
basis as equipment size increased. This was partially due University of Alberta as a research assistant.
to engine improvements, which increase fuel efficiency
and emissions control.

REFERENCES
Bozorgebrahimi, E. (2004). Evaluation of the haulage truck size effects Leslie, D. R. (2000). Diesel emissions evaluation for the surface mining
on open pit mining (PhD thesis). University of British Columbia, Van- industry. In R. K. Singhal & A. K. Mehrotra (Eds.), Proceedings of the
couver, BC. https://open.library.ubc.ca/cIRcle/collections/ubctheses/ 6th International Conference on Environmental Issues and Management
831/items/1.0081125 of Waste in Energy and Mineral Production (SWEMP 2000) (pp. 211–
218). Brookfield, VT: A.A. Balkema Publishers.
Caterpillar (2012). Caterpillar performance handbook (49th ed.). Peo- Sharma, A. (2009). Scale tire – oil sand interactions (Unpublished mas-
ria, IL: Caterpillar Inc. ter’s thesis). University of Alberta, Edmonton, AB.
Joseph, T. G. (2002). OsEIP: The oil sands – equipment interaction pro- Singh, R., Jain, R., Onder, K., & Unrau, G. (2008). Assessment of
gram. CIM Bulletin, 95(1064), 58–61. mobile emissions in the Athabasca oil sands region – CALPUFF mod-
elling case study using high resolution emission estimates. Proceedings
Joseph, T. G. (2003). Large mobile mining equipment operating on soft of the Air and Waste Management Association, 6, 3543–3549.
ground. Proceedings of the 18th International Mining Congress and Exhi-
bition of Turkey (IMCET 2003), Antalya, Turkey, 143–147. Retrieved Thompson, R. J. (2011). Building better haul roads. Engineering and
Mining Journal, 212(5), 49–53.
from http://www.maden.org.tr/resimler/ekler/ad21787ac6487a4_ek.pdf
Thompson, R. J., & Visser, A. T. (1997). An introduction to the inte-
Kaufman, W.W., & Ault J. C. (1977). The design of surface mine haul grated design of surface mine haul roads. International Journal of Sur-
roads (Information Circular No. 8758). Washington DC: US Bureau of face Mining, Reclamation and Environment, 11(3), 115–120.
Mines. https://doi.org/10.1080/09208119708944074
Impact of operating conditions on emissions from mine haul trucks

operating in the oil sands region of Alberta, Canada
R. J. Schafer-Frentz, T. G. Joseph, and M. Curley
ABSTRACT Greenhouse gas emissions—specifically of carbon dioxide—are a concern in Canada,

where industrial sectors such as mining are stringently regulated for carbon release. Cumulative emis-
sions are positively related to fuel consumption, which is controlled by several factors. This study inves-
tigates one of these factors: the product of vehicle speed and running surface total resistance (TR; sum
of rolling resistance and grade resistance). Results show that TR directly affects fuel consumption and
emissions—a relationship that is mathematically predictable. Fuel consumption and emissions, which
are principally associated with the durability or deterioration of wearing course materials, were analyzed
to aid selection of road construction capping material of low rolling resistance. In the oil sands region in
the province of Alberta, ground conditions change seasonally. Therefore, the annual fuel consumption
for the same wearing course material was also compared. Results demonstrated that haul truck size was
a contributing factor, but vehicle speed and TR more strongly influenced fuel consumption and emis-
sions.
■ KEYWORDS Carbon dioxide, Emissions, Fuel consumption, Grade resistance (GR), Greenhouse gas
(GHG), Haul truck, Rolling resistance (RR), Total resistance (TR)
RÉSUMÉ Les émissions de gaz à effet de serre (GES), et particulièrement celles de dioxyde de carbone,
sont un problème au Canada, où les secteurs industriels tels que l’exploitation minière sont rigoureuse-
ment réglementés pour leurs émissions de carbone. Les émissions cumulées sont indéniablement en lien
avec la consommation de carburant, qui est contrôlée par plusieurs facteurs. Cette étude examine l’un de
ces facteurs, à savoir le produit de la vitesse du véhicule et de la résistance totale de la surface de roule-
ment (la résistance totale, ou RT, est la somme de la résistance au roulement et de la résistance due aux
rampes). Les résultats montrent que la RT affecte directement la consommation de carburant et les émis-
sions, un lien qui est mathématiquement prévisible. La consommation de carburant et les émissions, qui
sont principalement associées à la durabilité ou la détérioration des matériaux composant la couche
d’usure, ont été analysées afin de contribuer à la sélection de matériau de recouvrement pour la construc-
tion de routes ayant une faible résistance au roulement. Dans la région des sables bitumineux de l’Al-
berta, les conditions au niveau du sol changent à chaque saison. Ainsi, la consommation annuelle de
carburant pour le même matériau composant la couche d’usure a aussi été comparée. Les résultats ont
montré que, si la taille des camions de transport est un facteur déterminant, la vitesse des véhicules et la
RT influencent davantage la consommation de carburant et les émissions.
■ MOTS CLÉS camion de transport, consommation de carburant, dioxyde de carbone, émissions, gaz à
effet de serre (GES), résistance au roulement, résistance due aux rampes, résistante totale (RT)
INTRODUCTION
In Canada and around the world, global climate change truck-shovel operations, where a large fleet of ultra-class
has become a topic of great concern and discussion. Pres- trucks is loaded by shovels. Ultra-class haul trucks, classi-
sure on world leaders to quickly reduce their carbon foot- fied as those with a capacity of 218–363 tonnes (t), emit
print is prevalent, prompting new standards and regulations high levels of carbon as they travel to processing entry
in various industries, including mining and oil and gas. In points and waste dumps as designated by the mine plan and
2017, the Government of Alberta implemented a carbon mine dispatcher. Mined ore and waste is typically moved
tax. The majority of mining activities in this province are over two 12-hour shifts per day throughout the year.

188 R. J. Schafer-Frentz, T. G. Joseph, and M. Curley
Over the past two decades, the mining industry has been 2020 (Environment Canada, 2007). Existing facilities were
developing novel, tiered-system engines to improve effi- given a reduction target of 6% per year for the first three
ciency and lower emissions. However, engine performance years and 2% per year for the next six years, for an overall
in the field has fallen short of anticipated standards. Thus, in reduction of 26% by 2015. In 2011, Canada withdrew from
the near-term, indirect improvements may be the most viable the Kyoto Protocol because the United States was not par-
alternative to reduce engine workload, fuel usage, and emis- ticipating and the Protocol favoured countries that had
sions. One such adjustment is to control the rolling resistance small contributions to global GHG emissions (Environment
(RR) for running surfaces. RR is an opposing moment to Canada, 2012a). Instead, Canada opted to follow the
motion acting at the vehicle tire–ground interface, hence it is Copenhagen Accord (2009) and Cancun Agreements
a significant factor in fuel consumption (Dotto, Joseph, & (2010), targeting emissions reductions of 17% relative to
Curley, 2019). For any heavily utilized hauler fleet, it is 2005 levels (730 Mt for Canada) by the year 2020. This
apparent that even small adjustments to lower fuel consump- effectively would reduce Canada’s overall emissions from
tion could greatly reduce cumulative emissions. the 2017 value of 716 Mt to 606 Mt CO2 eq. (Environment
This paper focuses on the impact of a change in RR on and Climate Change Canada, 2019).
fuel consumption and emissions by ultra-class haul trucks In March 2016, the Canadian GHG emissions target was
as a function of haul truck size and running surface condi- updated to a 30% reduction by the year 2030 relative to
tions. The datasets on oil sand in-pit and constructed haul 2005 levels (Government of Canada, 2016). The Govern-
road running surfaces were collected by the authors and ment of Canada also set a deadline for provinces and terri-
supplemented by data from Leslie (2000), Singh, Jain, tories to develop a carbon levy system or equivalent
Onder, and Unrau (2008), and Watson, Chow, Wang, cap-and-trade program or face a federal carbon tax of
Zielinska, Kohl, and Gronstal (2013). Carbon dioxide C$10/t in 2018, increasing to C$50/t by 2022 (Government
(CO2) emissions were principally investigated, though of Canada, 2016). In 2017, Alberta households and busi-
nitrous oxide (NOx), carbon monoxide (CO), unburnt nesses were levied a carbon tax of C$20/t, which increased
hydrocarbon (HC), and particulate matter > 2.5 µm (PM) to C$30/t on January 1, 2018 (Government of Alberta,
emissions were also evaluated. 2016). This equated to a C$0.0803/L increase in diesel
prices and a C$0.0673/L increase in gasoline prices in 2018.
Background drivers – a brief regulatory review To address high GHG emissions in the mining and other
Canada contributes 2% of the world’s total greenhouse industrial sectors, the Government of Alberta designated
gas (GHG) emissions, making it one of the highest emitters facilities that exceeded 100,000 t GHG emissions as “large
in the world on a per capita basis (Environment Canada, industrial emitters" (Government of Alberta, 2019a).
2007). The top five GHG-emitting provinces, Alberta, Unlike households and small businesses and until the end
Ontario, Quebec, Saskatchewan, and British Columbia, of 2017, large industrial emitters were required to decrease
emitted 650 million tonnes (Mt) CO2 eq. in 2017, repre- their site-specific annual emission intensity by 12%, as out-
senting 91% of Canada’s total CO2 eq. emissions that year lined in the Specified Gas Emitters Regulation (SGER). If
(Environment and Climate Change Canada, 2019). this goal was not achieved, the industry could either pur-
Alberta’s contribution alone was 38% of the national total. chase carbon offset credits or contribute to the Alberta Cli-
In 2017, 80% of Canada’s GHG emissions were from CO2 mate Change and Emissions Management Fund with a
(716 Mt CO2 eq.), with the majority due to combustion of C$30/t overage fee. In January 2018, the regulations transi-
fossil fuels (Environment Canada, 2019). An additional tioned from the SGER to a new output-based allocation
18.8% of GHG emissions was methane released from live- approach (Government of Alberta, 2019b). This regulation
stock and landfills and fugitive emissions from oil and nat- continues to apply to large industrial emitters, but it now
ural gas industries. NOx emissions contributed 5%, recognizes the effect of the tax on competitiveness for
primarily from the agriculture and transportation sectors. trade-exposed industries. Facilities are permitted to emit a
By sector, the biggest culprit for GHG emissions in specified level of GHGs, which is predetermined by a prod-
Canada in 2017 was mining and upstream oil and gas pro- uct-specific benchmark created to ensure protection from
duction, with total contributions of 104.5 (43% from the competitive impact (Government of Alberta, 2019b).
Alberta oil sands) and 56 Mt CO2 eq., respectively (Envi- The mining industry in Alberta is required to reduce
ronment Canada, 2019). Emissions from the mining sector emissions, with substantial financial repercussions for non-
(oil, including the oil sands, coal, and uranium extraction; compliance. In 2017, the carbon tax was applied to mine
iron ore, gold, copper, diamonds, potash, and aggregate sites in accordance with the SGER; GHG emissions associ-
mining) increased more than six-fold between 1990 and ated with a haul fleet were used to calculate the total emis-
2017, from 6.8 to 43.4 Mt (Environment Canada, 2019). sions and determine if the mine site exceeded the 100,000 t
Following the signing of the Kyoto Protocol in October GHG threshold. In 2018, all mining companies that
2006, the Government of Canada committed to reducing exceeded the threshold or chose to register in adopting the
GHG emissions by 20% relative to 2006 levels by the year 2017 carbon tax regulation immediately became subject to

Impact of operating conditions on emissions from mine haul trucks operating in the oil sands region of Alberta, Canada 189
an overall output-based assessment approach. Additionally, Joseph et al. (2017) re-evaluated RR as a reaction torque
the Government of Alberta primarily targeted mine opera- or moment, assessing the ground reaction to tire loading at
tors in the oil sands region in 2018 by enacting a maximum any instant in motion as a Boussinesq stress pressure bulb
emissions level of 100 Mt GHG across the entire mining resisting tire penetration and ground deformation. The
industry, which was 30 Mt higher than the former allowable authors were able to determine field RR knowing the
sum emissions level to allow for industry growth (Govern- ground stiffness, tire stiffness, and loads imparted from the
ment of Alberta, 2019c). tires to the ground reflected by the haul truck strut pressures
The entire Canadian mining industry will soon be sub- recorded on board.
ject to Canada-wide carbon pricing. With the continued Financial and environmental implications make RR a
reliance on haul trucks—powered by diesel engines for the significant consideration for mining operations. Poor RR
foreseeable future—the industry will have no choice but to control can result in high vehicle operating costs and road
adhere to the carbon policy. This is perceived as a disadvan- maintenance costs. For example, consider a surface mining
tage to the Canadian industry, which must now compete operation with a fleet of 290 t payload rear-dump haul
with countries selling the same mined product but with no trucks moving 30 Mt of material per year and operating on
such price on carbon. a 10% GR incline over a road length of 4 km. Reducing the
One strategy to help mining operations meet the carbon RR from 6 to 2% would decrease operating fuel costs per
tax challenge is to decrease daily fuel consumption and thus truck by 16% (Thompson & Visser, 2006). The capital costs
reduce the haul fleet GHG emission rate. This could be of equipment required to move the same amount of material
achieved through the relatively simple action of decreasing would decrease by 18% because the sustaining capital for
running surface RR by improving haul road construction replacing damaged units in a fleet would be lower.
and maintenance.
Emissions, the tiered system, and load factor
RESEARCH APPROACH Haul road RR is a principal driver for fuel consumption
Dotto et al. (2019) evaluated emissions data from haul in surface mining operations. Reducing RR reduces the fuel
trucks operating in the Alberta oil sands on known running consumed by a vehicle, resulting in a reduction of CO,
surfaces and found that diesel fuel consumption and emis- CO2, NOx, and volatile unburnt HC emissions (Soofastaei,
sions were a function of the total resistance (TR) that a vehi- 2016). Leslie (2000) calculated the diesel combustion prod-
cle must overcome. TR is the sum of RR and grade ucts using equation 1:
resistance (GR; Soofastaei, 2016). Whereas Dotto et al.
(2019) focused on the impact of hauler size on emission lev- CxHy(Sz) + O2 + N2 → CO2 + H2O +
(1)
els as part of mine planning, this paper expands the response {NOx + HC + OOC + C + CO + SOx}
variables to include fuel consumption and the influencing
factors to include vehicle speed and road conditions. where CO2, HC, and free carbon (C) in the exhaust stream
are the result of incomplete combustion, whereby a portion
Rolling resistance of the fuel passes through the exhaust unburned or partially
The definition of RR depends on the application. Bar- burned. This typically occurs under fuel-rich or high-RR
rand and Boker (2009) defined RR as “the energy con- combustion conditions that are related to the production of
sumed by a tire per unit distance travelled.” Thompson and excess CO (Flagan & Seinfeld, 1988). Unburnt HCs may
Visser (2006) defined RR as the additional resistance acting further react in the atmosphere to create other pollutants
against a vehicle in motion as a result of a variety of factors. such as ozone and smog (Leslie, 2000). The SOx in acidic
RR has also been described as a function of haul truck tire airborne combustion products is directly related to the sul-
air pressure, speed, type, and weight and ground conditions fur content of the diesel fuel.
(Joseph, Curley, & Anand, 2017). Discussion of diesel engine emissions regulations in
The RR of a haul road is strongly influenced by the North America began in the early 1990s, when noncompli-
unevenness of the rolling surface, the magnitude and fre- ance with United States National Ambient Air Quality Stan-
quency of ruts or potholes, structural deformities, and the dards was identified (Weyant, 2010). This moved the
roughness of the running surface material. On soft surfaces, United States Environmental Protection Agency (US EPA)
a tire penetrates more deeply into a wearing course surface to propose a “tiered” engine system emissions rating, which
and influences the sub-grade material performance. Empir- has been used in the United States and Canada for the past
ical estimates show that an increase of 0.6% RR will occur 20 years. The two main components regulated under the
per centimetre of tire penetration or road surface flexing tiered system are particulates and NOx. In 1999, the Cana-
(Thompson & Visser, 2006). Since penetration or surface dian Environmental Protection Act officially aligned itself
deformation creates an effective slope immediately in front with US EPA standards to incorporate emissions regula-
of a tire in motion, the vehicle constantly travels uphill on tions for off-road engines (DieselNet, 2017). This brought
what appears to be a flat running surface. a compliance requirement to bear on mining, farming,

forestry, and construction diesel-powered equipment to the Kecojevic and Komljenovic (2010) established a linear
US EPA Tier 1 standard, which was effectively met across relationship between fuel consumption and load factor,
industries by 2003. defined as the fraction of available engine power utilized to
To move from Tier 1 to Tier 2, a further reduction in par- do work. However, fuel consumption increased with truck
ticulate emissions was accomplished with the introduction size and was proportional to payload carrying capacity.
of the diesel particulate filter, generating a more easily cap- According to Kecojevic and Komljenovic (2010), this was
tured exhaust ash at temperatures above 315°C (Weyant, an important revelation for mine operators because they
2010). Although considerable research and development, could identify and address the factors that influence fuel
field trials, and modifications have ensued over the past consumption, permitting such strategies as prioritizing
decade, original equipment manufacturers, mining opera- larger haulers to areas of higher road quality to lower the
tions, and researchers have struggled to consistently effective load factor or RR response. Fuel costs were also
improve emission reductions beyond the Tier 2 standard. proportional to payload capacity by reduction in load fac-
The Tier 3 initial and final standards were replaced by tor. Kecojevic and Komljenovic (2010) also calculated CO2
the Tier 4 standard slated to be in place by 2015 (Environ- eq. emissions (t/h) as a function of diesel fuel consumption
ment Canada, 2012b) and finally realized in 2017. Figure 1 for Tier 2 engines (equation 2):
shows that particulate emissions were brought under con-
trol quickly and effectively to the present Tier 4 standard, CO2 eq. = 0.00268 FC (2)
with only trace particulates escaping capture. However,
controlling NOx emissions has been much more difficult. where FC is the diesel fuel consumption (L/h) and 0.00268
Engine system bypasses activated under high engine is a conversion factor based on the product of the carbon
loads—in part because of high TR—result in unburnt HC content of diesel fuel (7.3×10-6 t/L) and a dimensionless
venting along with uncaptured NOx. This in turn resulted in effective oxidation factor constant of 367. CO2 emissions
early Tier 2 and 3 engines performing at an equivalent were directly proportional to load factor in the form of run-
lower Tier state. For mines experiencing poor running sur- ning surface road quality. It was also noted that hauler over-
face conditions with high TR, achieving the Tier 4 standard loading contributed to higher CO2 emissions, as might be
remains an operational challenge. logically expected.
Under-inflated tires increase fuel consumption by creat-
ing a larger tire–ground footprint and increasing RR, Estimating emissions from field data and fuel
whereas maintaining a constant vehicle speed of 30 km/h consumption
while increasing the tire pressure from 50 to 300 kPa Mean fuel consumption data originally assembled by
decreases RR (Djordjevic, Jankovic, & Jeremic, 2009). Leslie (2000) were extrapolated to include haul truck sizes
Djordjevic et al. (2009) also showed that an increase in tire from payloads of 70 t (108 t gross vehicle weight or GVW)
pressure above 300 kPa did not lower fuel consumption. to 363 t (624 t GVW; Table 1) based on supplementary field
For most ultra-class haul trucks in surface mining opera- fuel data collected by the authors and sourced from Singh et
tions, tire inflations are nominally set at 600 kPa, such that al. (2008) and Watson et al. (2013). This permitted the use of
an operational variance in tire pressure (± 100 kPa) should the CO2 emissions estimate of Kecojevic and Komljenovic
not significantly increase fuel consumption and emissions. (2010) from equation 2 to generate smaller Tier 2 performing
hauler class CO2 emissions data
to add to the measured CO2
emissions data reported by Wat-
son et al. (2013) for ultra-class
400 t payload (624 t GVW)
trucks. Until the work of Watson
et al. (2013), there was a defi-
ciency of reported and verifiable
CO2 emissions data for mine
haul trucks.
Figure 2 shows distinct sea-
sonal variation in fuel consump-
tion driven by differences in
running surface operating con-
ditions. In winter, the oil sands
running surfaces freeze, creat-
ing hard rock-like ground con-
Figure 1. Particulate and NOx emissions relative to Tier 1–4 standards. Adapted from O’Neill (2006). ditions, whereas in summer, the

same running surfaces Table 1. Haul truck fuel consumption and CO2 emissions (adapted from Leslie, 2000)
thaw and soften to behave Payload Gross vehicle weight Effective engine power Mean fuel consumption CO2 emissions
like soft soil (Joseph, (t) (t) (kW) (L/h) (t/h)
2003) with high RR. 70.3 108.4 590 212.3 0.57
Spring and autumn gener- 90.7 161.0 700 261.9 0.70
ate weathered lateritic- 108.9 189.9 870 301.5 0.81
like ground conditions 136.1 231.3 960 352.8 0.95
with stiffness behaviour 154.2 249.4 960 389.5 1.04
that falls between the win- 176.9 317.5 1,320 411.8 1.10
ter and summer extremes 217.7 376.5 1,615 429.1 1.15
(Joseph et al., 2017). 290.3 521.6 2,200 410.9 1.10
A density of 326.6 556.0 2,400 355.0 0.95
0.8171 kg/L for diesel fuel 362.9 623.7 2,510 217.3 0.58
at 25°C (Environment
Canada, 2001) was used to
convert the Leslie (2000) data for fuel consumption
from in lb/h to L/h. The fuel consumption and corre-
sponding CO2 emissions increased with truck size up
to the ultra-class category (218 t payload and
377 GVW) (Table 1, Figure 3). To understand why
the trend declined thereafter, one must consider the
significant efforts made by original equipment man-
ufacturers with the advent and development of the
ultra-class truck after 1997, when engine develop-
ment primarily targeted ultra-class haulers. Emis-
sions capture was driven by the advent of the tiered
emission regulations of the US EPA, with an initial
focus on PM and NOx. In addition, technological
advances resulted in more efficient fuel control and
consumption. Consequently the trend observed with
increasing truck sizes since the mid-1990s is under-
standable. It is interesting to note that Leslie (2000)
was already observing lower emissions in larger Figure 2. Fuel consumption and CO2 emissions for a fleet of 50 haul trucks (624 t
class field trials and also in smaller units as devel- gross vehicle weight) operating in an oil sands mine
oped from pre-Tier 1
through Tier 1 and target-
ing Tier 2 engines in field
trials.
NOx, CO, PM, and
unburnt HC Tier 2 engine
equivalent emissions data
reported by Leslie (2000),
Singh et al. (2008), and
Watson et al. (2013) were
collated and regressed
against the GVW of haul
trucks from 108 to 624 t
GVW (70–363 t payload).
All data were gathered by
these researchers during
field monitoring of haul
trucks operating in oil
sands mining operations.
Figure 3. Oil sands mine fuel consumption and CO2 emissions by truck gross vehicle weight (GVW) based on
The collated emissions measured data reported by Singh et al. (2008) and Watson et al. (2013), as well as Leslie (2000) data estimated
data initially appeared to from equation 2

have a linear relationship with GVW (Figure 4). However, The vertical axis in Figure 6 expresses a scale for all
at a hypothetical zero GVW, there can be no emissions, thus emissions in common units of tonnes for the 624 t GVW
the relationships should be asymptotic approaching zero hauler example. For parts of the duty cycle where no GR
emissions for zero GVW. was provided, those sections were essentially flat with no
Given that the trends in Figure 4 are similar among emis- elevation grade change.
sion types, it is possible to express emissions as a function of The relationship between fuel consumption and effec-
CO2 emissions (Figure 5). The impacts of technological tive power used the Watson et al. (2013) fuel consumption
advances after 1995 (advent of the Tier 2 engine) and through data converted to litres using a density of 0.8171 kg/L for
the mid-2000s (increasing engine fuel burn efficiency) are diesel fuel at 25°C (Environment Canada, 2001). Higher
evident in the change from more linear relationships in unreg- RR areas in a mine would require greater power delivery
ulated engines prior to 1995. The authors postulate that as the on a haul truck, which translates to higher fuel consump-
major emission type by volume or tonnage, CO2 is effectively tion (Figure 7). It can then be considered that power and
a carrier within which all other
emissions may be considered as a
fraction of the CO2 stream. Fig-
ures 4 and 5 support this con-
tention: for haulers performing at
Tier 2 and higher emissions con-
trol standards, emissions varied
little with increasing CO2 emis-
sions relative to the quantity of
fuel consumed. This also sug-
gests that a decrease in CO2
emissions, as a principal mea-
surement indicator, would result
in reduction of all other emis-
sions of concern (Figure 5).
The trends in Figures 4 and 5
were applied to a 624 t GVW
(363 t payload) truck duty cycle
dataset from an oil sands opera-
tion (Watson et al., 2013). The
Figure 4. Tier 2 haul truck CO2, NOx, CO, unburnt hydrocarbon (HC), and particulate matter > 2.5 µm (PM)
resulting emission levels were emissions estimates versus gross vehicle weight based on data originally reported by Leslie (2000), Singh et
generated from the emissions per al. (2008), and Watson et al. (2013)
kg fuel consumption reported by
Watson et al. (2013) for the 624 t
GVW hauler example and differ-
ent sections of hauler activity.
The fuel consumption was pro-
portional to the effective power
generated and hence speed to
overcome TR to motion. The
fuel consumption was then eval-
uated using equation 3 (Caterpil-
lar Inc., 2016):
P = TR C g v (3)
where P is the effective power

(kW), TR is total resistance (%),
C is a constant proportional to
TR (10 kg/t GVW per % TR), g
is acceleration due to gravity
Figure 5. Tier 2 haul truck NOx, CO, unburnt hydrocarbon (HC), and particulate matter > 2.5 µm (PM)
(9.81 m/s2), and v is the truck emissions estimates versus CO2 emissions based on data originally reported by Leslie (2000), Singh et al.
speed (m/s). (2008), and Watson et al. (2013)

fuel consumption are functions of RR.

If one considers the impact of higher
RR in pit and dump areas and the
compounding impact of ramp grades
in and out of pit and dump areas in an
oil sands mine, it would be expected
that higher emissions should domi-
nate during those duty cycle activities.
Figure 6 does not indicate such domi-
nance, but suggests that the main haul
roads—where lower RR and GR
would be expected—generated
equally high emissions.
The speed at which the haul truck
is able to travel compounds the impact
of RR and GR. On main haul roads,
faster speeds can be maintained. In
the case of oil sands mining opera-
Figure 6. Emissions from a 624 t (gross vehicle weight) haul truck for a single duty cycle during
tions, speeds of 16–30 km/h and operations in an oil sands mine; RR: rolling resistance, GR: grade resistance, HC: unburnt
greater than 35 km/h are typical for hydrocarbon, PM: particulate matter > 2.5 µm. Data from Watson et al. (2013).
loaded and unloaded haul trucks,
respectively. In the pit and dump areas, truck
speed is vastly reduced due to traffic, as well as
poor underfoot operating conditions; thus,
speeds of 5–10 km/h are common. Similar
speeds would be expected on ramps in the 6–
8% slope range. As an example, Figure 8 illus-
trates the relationship between TR derived
from the Watson et al. (2013) data and vehicle
speed. In light of these considerations, Figure 6
reflects what one would expect for emissions
relative to operating conditions and vehicle
speed: emissions should be expressed as a
function of TR and vehicle speed.
To verify the generality of the patterns in
Figure 6 for a 624 t GVW truck, data from
Singh et al. (2008) were plotted for a much Figure 7. Fuel consumption as a function of effective power generated (from equation 3)
for a 624 t (gross vehicle weight) haul truck
smaller 108 t GVW truck operating under
similar oil sands mine operating conditions.
RR is a material property, GR is set by the
geometry of mine ramps, and neither RR nor
GR is a function of the GVW over the foot-
print area of six tires (Joseph et al., 2017).
Therefore, similar TR values for the duty
cycle were adopted. It is clear that the varia-
tion in emissions for the smaller hauler were
similar to those of larger hauler relative to
mine GR and RR (Figure 9).
Regressing emission levels or fuel con-
sumption against the product of vehicle speed
and TR (km/h%) yields a family of parallel
curves that share a core relationship (Figure
10) described by equation 4. It is evident that
trucks consumed more fuel when traveling at Figure 8. Relationship between total resistance of the running surface and speed of a
higher speed and when they encountered 624 t (gross vehicle weight) haul truck

Fuel consumption or emissions =

P{Ln(v.TR)−Q} (4)
where P is a constant specific to

fuel consumption (32) or the
emission gas (0.00002–0.086); v
is vehicle speed (km/h); TR is
total resistance (%); and Q is a
constant for fuel consumption
and CO2 (3.8), for NOx, CO, and
unburnt HC emissions (4.0), and
for PM emissions (3.5).
Application of
relationships to a raw field
dataset for a 624 t GVW
hauler
The RR evaluation approach
Figure 9. Emissions from a 108 t (gross vehicle weight) haul truck for a single duty cycle during operations
in an oil sands mine; RR: rolling resistance, GR: grade resistance, HC: unburnt hydrocarbon, PM: of Joseph et al. (2017) was used
particulate matter > 2.5 µm. Data from Singh et al. (2008). to interpret the running surface
RR profile for a full duty cycle
at an oil sands open pit mine.
Suspension response data were
evaluated for a 624 t GVW
hauler in operation for 8 h, dur-
ing which 15 full duty cycles
(from shovel to dump and
return) were performed. A sam-
ple of the raw suspension pres-
sure data is shown in Figure 11.
For each of the running sur-
faces encountered by the truck,
a pressure stiffness cyclic plate
load test was performed in the
field (Joseph et al., 2017) to
generate a linear relationship
with truck speed (Figure 12) and
TR (Figure 13). By combining
the outcome of Figure 12 with
the equation in Figure 8, Fig-
Figure 10. Fuel consumption and emissions for a 624 t (gross vehicle weight) haul truck as a function of the ure 13 permits an association to
product of vehicle speed (v) and total resistance (TR) of the running surface be made between ground perfor-
mance under the action of truck
ground with higher RR (Figure 2) or GR. If we consider motions, as may be evaluated from truck onboard data
operational practices—observed by the authors during (Joseph et al., 2017), and TR commensurate with fuel con-
field work—operators of haul trucks operating on low sumption and emissions.
RR or GR surfaces (e.g., principal main haul roads The raw suspension pressure data from Figure 11 were
between pit and dump locations) tend to drive faster than reduced to force (kN) using the known front and rear sus-
in pit or dump areas or on ramps where either RR or GR pension internal active bearing areas (calculated from
is much higher. In pit and dump areas, specific road sur- internal suspension diameters of 0.40 and 0.38 m, respec-
face construction is not practiced, such that running sur- tively). The force generated by the suspension is common
faces are compacted by hauler trucks only, with far less to the entire suspension–tire–ground contact. It generated
grader maintenance performed compared to designated a tire deformation (δ) relative to the tire elastic stiffness
main haul roads. Fuel consumption is then driven by of 3.65 MN/m for a 59/80R63 E4 designated 4.0 m diam-
high TR, high operating speed, or poorly managed roads. eter (ф) tire. The tire was deformed diametrically and

elastically by 0.28 m (7% dia-

metrical strain) under a 1 / 6
nominal GVW load (i.e., load
was equally shared among the
six tires on the truck). The tire
deformation was then used to
evaluate the footprint area of
the tire generated by the tire
deformation reported by
Joseph et al. (2017) as 1.35δф.
The ratio of the force passing
through the system to the tire
footprint area then defined the
ground deformation.
The RR based on the ground
and tire deformations was then
calculated after Joseph et al.
(2017) to generate Figure 14. It
Figure 11. Sample 624 t (gross vehicle weight) haul truck speed and left front (pLF), left rear (pLR), right should be borne in mind that the
front (pRF), and right rear (pRR) suspension pressures during a duty cycle number of events measured on a
1 Hz record is also representa-
tive of the time it takes to travel the length of a given run-
ning surface. Since the mining operation has much longer
main haul roads of low RR compared to the distance trav-
elled either in-pit, on dumps, or on ramps, the number of
events recorded for the main haul roads exceeded those for
any other surface in the mine by an order of magnitude.
It is clear from Figure 14 that the constructed main haul
road, ramps, and dump and pit areas represent a broad
range of RRs that a hauler would encounter. The main haul
roads accounted for the highest number of events and had
the lowest RR, which permitted haul trucks to operate at
faster speeds, generating higher-than-expected fuel con-
sumption and emissions. The RR in ramps ranged from 6.5
to 8.5% (mean 7.5%). The added effect of 6–8% GR means
that the TR in ramps varied from 12.5 to 16.5%. In ramps,
Figure 12. Measured pressure stiffness cyclic plate load versus 624 t
(gross vehicle weight) haul truck speed achieved on surface fuel consumption and emissions were therefore as high as
would be experienced in softer running surfaces (i.e., pit
and dump areas with 8.5–11.0% RR).
Figure 15 shows the application of the relationships in
Figure 10 to the haul truck data in Figure 11. It is evident
that the majority of hauler activity—other than for slow-
ing down and speeding up for intersections, accessing
and leaving ramps, and slowing to idle during pit loading
and dumping activity—generated similar fuel consump-
tion and emissions relative to v.TR, giving rise to a simi-
lar required effective power draw. The lower fuel
consumption and emissions in-pit compared to on main
haul road reflect an order of magnitude (10×) fuel con-
sumption and proportionate emissions associated with
hauling a 363 t payload on main haul routes at faster
speeds. In pit areas, haulers move very slowly over poor
ground conditions, while on main haul routes, the ground
Figure 13. Field measured ground stiffness versus total resistance for a conditions are of higher quality (lower RR) and haulers
624 t (gross vehicle weight) haul truck move more quickly.

The distributions of all fuel burn and

emission events for the sample 8 h dataset
as a function of v.TR (Figure 16) indicate
an order of magnitude (10×) variance for
> 50% of the events measured, commensu-
rate with either higher speed or higher TR
operating conditions. This suggests that a
significant fuel consumption and emissions
reduction could be realized by reducing RR
for hauling in oil sands mining operations,
which is reflected in the fuel consumption
shown in Figure 2. Lowering RR and using
road construction capping materials
enhances the pressure stiffness of running
surfaces such that v.TR reduces the power
Figure 14. Distribution of 18,740 rolling resistance values for the left front (LF), left rear (LR), required for a haul truck to perform its duty
right front (RF), and right rear (RR) tires on a 624 t (gross vehicle weight) haul truck for 15 truck cycle.
duty cycles in an 8 h operating period in an oil sands mine (after Joseph et al., 2017)
In the example dataset used in this paper,
the haul truck only encountered GR for a
short period at low speed when entering or
exiting pit and dump areas. As such, RR is
the major component of TR in v.TR in Fig-
ure 16. The only uphill ramp activity in the
dataset is indicated by the four short periods
showing a drop in fuel consumption or emis-
sions immediately prior to main haul activity.
Road maintenance may seem an obvious
approach to reduce RR in oil sands opera-
tions. Whereas it occurs continuously on
main haul roads to ensure that principal
pathways remain trafficable, road mainte-
nance is not scheduled in pit and dump loca-
tions but occurs on an “as and if required”
basis.
Figure 15. Fuel consumption and emissions during a duty cycle for a 624 t (gross vehicle
weight) ultra-class haul truck operating in an oil sands mine Selecting wearing course road
surface materials for lower truck
emissions
The wearing course is the uppermost
layer in a mine haul road construction. It
generally comprises the strongest available
material on a mine site, that is, mine waste
that is frequently in short supply. Weaker
wearing course material requires thicker
base and sub-base material to support the
applied load (Shi, Joseph & Rasimarzabadi,
2018), as shown in Figure 17.
The Boussinesq pressure bulbs suggest
that if the width of the footprint area of a
moving tire is √A, then only 10–15% of the
surface tire footprint contact stress will be
transmitted to a depth of influence
Figure 16. Fuel consumption and emissions rates for an ultra-class haul truck (624 t gross (3.5×√A). For the 624 t GVW truck loading
vehicle weight) operating in an oil sands mine as a function of the product of vehicle speed
(v) and total resistance (TR) of the running surface; RR: rolling resistance, GR: grade a road surface over a tire footprint area of
resistance, HC: unburnt hydrocarbon, PM: particulate matter > 2.5 µm 1.5 m2 and generating a surface stress of

734 kPa, the stress at a depth of 3.5×√A (4.3 m) is 140 kPa. the resilient moduli derived from the pressure stiffness via
This stress is considered very low and would generate neg- CBR values (Figure 17). Limestone provided the thinnest
ligible ground deformation. cap layer required to permit the 100,000 t/d hauler traffic
Figure 18 shows the results of RR tests (after Joseph et for the 624 t hauler fleet of 21 units, such that minimal
al., 2017) on a lean oil sand (<6% bitumen
content) that was either uncapped or
capped with three wearing course candi-
date materials: sand, crushed limestone,
and pit run (coarse blasted waste rock).
The RR performance had a narrow range
(± 5–10%) in the uncapped and pit run and
limestone cap treatments. Both the pit run
and limestone cap had lower RR than the
oil sand without a wearing course cap.
However, the sand cap had high RR (mean
2.5× that of the uncapped oil sand), mak-
ing it a poor choice for a wearing course
layer. It is not considered further.
The pressure stiffness values are also
expressed in Figure 18 in terms of an
equivalent California Bearing Ratio (CBR
in %) using a CBR versus pressure stiffness
relationship inferred by the tests conducted
by Shi et al. (2018). The CBR values can
further be expressed as resilient modulus
(ER) values using the relationship reported Figure 17. Boussinesq-based road construction layers and stresses for a 624 t gross vehicle
by Thompson and Visser (2006). Equa- weight (GVW) haul truck loading a road surface; ER: resilient modulus of each material, √A:
tion 5 permits an evaluation of the defor- square root of the tire footprint area at any instant, t: layer thickness, z: layer depth. After Shi et
mation performance of running surface al. (2018).
layers of varying thickness given the stress
impacting that layer as indicated from the
Boussinesq approach in Figure 17.
ER = 17.63 CBR0.64 (5)
A critical strain limit of 1,750 microstrain

(µm/m or µε) was established following the
guidelines of Thompson and Visser (2006),
commensurate with running surfaces capa-
ble of trafficking at least 100,000 t/d for up
to 10 years. For 104 t loading per tire
mounted on a loaded 624 t GVW truck,
this suggests that for trucks trafficking
15 full duty cycle loads (from pit to dump
Figure 18. Rolling resistance test results for example lean oil sand (<6% bitumen content) that
to pit return) per 8 h operating shift period was either uncapped or capped with three wearing course materials: sand, crushed limestone,
(Figure 11), the road should be designed or pit run (coarse blasted waste rock); kgr: pressure stiffness, CBR: California Bearing Ratio
to handle a maximum fleet of 21 trucks,
each moving 45 loads within 24 h. Table 2. Wearing course cap thickness analysis for a set critical strain limit
The required wearing course layer
Wearing Resilient modulus Critical strain Cap thickness
thickness (Table 2) was determined by course material (MPa) (µɛ) (m)
evaluating the appropriate thickness of Oil sand 67 1,747 4.09
wearing course material on a lean oil sand Limestone 170 1,748 2.01
base and sub-base to ensure that the criti- Pit run 118 1,747 2.83
cal strain was not exceeded and relying on

deformation would not deteriorate the running surface and It is evident that to see a greater reduction in emissions,
require more road maintenance. it is contingent upon the mining operation to balance the
Figure 19 shows the relationships from Figure 10 cycle time (controlling speed) with the cost to the environ-
applied to the three wearing course materials in Figure 18. ment. Since emissions depend on the product of truck speed
The inability of the sand cap is highlighted: at a mean RR and TR, the mine planner must consider not only the pro-
of 25%, the truck was unable to generate sufficient power duction requirement, but also the impacts of greater effort
to traffic loads. Thus, no analysis was possible as no motion in road construction and maintenance and changing grade
(v = 0) negated use of the relationships from Figure 10. For due to ramps. Mine design must seek a balance to lower
all other wearing course cap options, the increase in emis- fuel consumption and emissions in a post-2000 global envi-
sions by quality (lower RR) was driven by the fact that a ronment.
hauler could move faster on a low-TR surface.
An earlier draft of this paper was published in the
Proceedings of the Maintenance, Engineering and
Reliability/Mine Operators Conference (MEMO
2017) prior to undergoing the CIM Journal peer-
review process.
Paper reviewed and approved for publication by the

Maintenance, Engineering and Reliability Society
of CIM.
Regan Schafer-Frentz graduated with a BSc (2014) and M.Eng.
(2017) in mining engineering from the University of Alberta, the
latter with a research focus in haul truck emissions. Since 2017, Ms.
Schafer-Frentz has worked for GHD Ltd. in Calgary, Alberta, in
project coordination, development of soil and groundwater
remediation programs, Phase I and Phase II environment site
assessments, and industrial approval renewal applications. Her field
experience includes groundwater monitoring and sampling,
monitoring well installation, and soil and surface water sampling.
Tim G. Joseph, PhD, P.Eng., FCIM, is associate dean of

engineering and professor of mining engineering at the University
of Alberta, and director of the Alberta Equipment–Ground
Figure 19. Evaluation of emissions for four wearing course cap materials; RR: Interactions Syndicate (AEGIS) research program. He is vice-
rolling resistance, HC: unburnt hydrocarbon, PM: particulate matter > 2.5 µm president of the Association of Professional Engineers and
Geoscientists of Alberta (APEGA), an active ambassador and
former vice-president to the CIM, the president and principal
CONCLUSIONS engineer of JPI Mine Equipment, and recognized as an expert in mine
This paper outlined a procedure for evaluating haul equipment performance related to ground–equipment interactions. He
truck emissions based on RR, GR, and truck speed. When delivers certification industry short courses and mining equipment
performance evaluations to the global resource industry, mining equipment
combined, these variables are proportional to the power manufacturers, and service providers. He is a former recipient of the CIM
required to cyclically traffic loads from pit to dump loca- Distinguished Service Medal for his contributions to the global mining
tions in an oil sands mine. The base emissions field data industry and mining engineering education and a recipient of the CIM Past
President’s Memorial Medal for early achievements in education to the
were used to establish strong empirical relationships that global mining industry. Dr. Joseph will retire from the University of Alberta
could reasonably be applied to the same hauler sizes and in 2020 but will continue to practice through JPI Mine Equipment for the
effective power capability investigated here. Although con- foreseeable future. tjoseph@ualberta.ca
tinued confirmatory work is planned in this area of opera- Mark Curley, MSc, EIT, is a graduate of the University of Alberta mining
tional research, there is an opportunity to take the increase engineering BSc (2013) and MSc (2016) programs. As a mining engineer in
in knowledge reported here and apply it to improve opera- Alberta, he works for BGC Engineering in Edmonton on international
mining projects, with considerable time spent in the Dominican Republic
tional mine planning. Thus, efforts to reduce emissions for expanding exploration targets. He has also worked directly for Dr. Joseph at
mine haul trucks may be reported with greater confidence the University of Alberta as a research assistant.
in meeting compliance with increasingly stringent provin-
cial and federal emissions regulations.

REFERENCES
Barrand, J., & Bokar, J. (2009). Reducing tire rolling resistance to save Government of Canada (2016). Pan-Canadian approach to pricing carbon
fuel and lower emissions. SAE International Journal of Passenger Cars pollution. Retrieved September 1, 2019, from https://www.canada.ca/
– Mechanical Systems, 1(1), 9–17. https://doi.org/10.4271/2008-01- en/environment-climate-change/news/2016/10/canadian-approach-pricing-
0154 carbon-pollution.html
Caterpillar Inc. (2016). Caterpillar performance handbook (46th ed.). Joseph, T. G. (2003). Large mobile mining equipment operating on soft
Peoria, IL: Caterpillar Inc.
ground. Proceedings of the 18th International Mining Congress and Exhi-
DieselNet (2017). Emission Standards Canada: Off-road engines and bition of Turkey (IMCET 2003), Antalya, Turkey, 143–147. Retrieved
vehicles. Retrieved on September 1, 2019, from https://www.dieselnet. from http://www.maden.org.tr/resimler/ekler/ad21787ac6487a4
com/standards/ca/#nonroad _ek.pdf
Djordjevic, M., Jankovic, A., & Jeremic, B. (2009). Rolling resistance Joseph, T. G., Curley, M., & Anand, A. (2017). Operational methodolo-
as the risk factor for fuel consumption. International Journal of Vehicle gies for rolling resistance evaluation. Geotechnical and Geological
Systems Modelling and Testing, 4(3), 185–200. https://doi.org/
Engineering, 35(6), 2935–2946. https://doi.org/10.1007/s10706-017-
10.1504/IJVSMT.2009.029388
0292-y
Dotto, M., Joseph, T. G., & Curley, M. (2019). Is bigger still better?
Considerations in increasing the size of haulage equipment. CIM Jour- Kecojevic, V., & Komljenovic, D. (2010). Haul truck fuel consumption
nal, 10(4), 180–186. https://doi.org/10.15834/cimj.2019.14 and CO2 emissions under various engine load conditions. Mining Engi-
neering, 62(12), 44–48.
Environment and Climate Change Canada (2019). Canadian environ-
mental sustainability indicators: Greenhouse gas emissions. Retrieved Leslie, D. R. (2000). Diesel emissions evaluation for the surface mining
from http://publications.gc.ca/pub?id=9.854099&sl=0 industry. In R. K. Singhal & A. K. Mehrotra (Eds.), Proceedings of the
Environment Canada (2001). Diesel fuel oil (Canada). Retrieved from 6th International Conference on Environmental Issues and Management
http://www.etc-cte.ec.gc.ca/databases/oilproperties/pdf/web_diesel_fuel_ of Waste in Energy and Mineral Production (SWEMP 2000) (pp. 211–
oil_(canada).pdf 218). Brookfield, VT: A.A. Balkema Publishers.
Environment Canada (2007). Regulatory framework for air emissions. O’Neill, B. (2006, September). Caterpillar new technologies. Paper pre-
Retrieved from https://www.ec.gc.ca/doc/media/m_124/report_eng.pdf sented at the SMART Learning Seminar, Edmonton, AB.
Environment Canada (2012a). A climate change plan for the purposes Shi, E., Joseph, T. G., & Rasimarzabadi, R. (2018). Crushed rock pad
of the Kyoto Protocol Implementation Act 2012. Retrieved from capability under mining equipment operation. International Journal
http://publications.gc.ca/collections/collection_2012/ec/En11-11-2012- of Geotechnical Engineering, 12(5), 429–437. https://doi.org/10.1080/
eng.pdf 19386362.2017.1289434
Environment Canada (2012b). Off-road compression-ignition engine Singh, R., Jain, R., Onder, K., & Unrau, G. (2008). Assessment of
emission regulations. Retrieved from http://publications.gc.ca/collec-
mobile emissions in the Athabasca oil sands region – CALPUFF mod-
tions/collection_2012/ec/En14-70-2012-eng.pdf
elling case study using high resolution emission estimates. Proceedings
Environment Canada (2019). National Inventory Report 1990–2017: of the Air and Waste Management Association, 6, 3543–3549.
Greenhouse gas sources and sinks in Canada. Retrieved from
http://data.ec.gc.ca/data/substances/monitor/canada-s-official-greenhouse- Soofastaei, A. (2016). Development of an advanced data analytics
gas-inventory model to improve the energy efficiency of haul trucks in surface mines
(Doctor of Philosophy thesis). University of Queensland, Queensland,
Flagan, R. C., & Seinfeld, J. H. (1988). Fundamentals of air pollution
AU. https://doi.org/10.14264/uql.2016.530
engineering. Englewood Cliffs, NJ: Prentice-Hall, Inc.
Government of Alberta (2016). Fiscal plan 2016–19. Retrieved from Thompson, R. J., & Visser, A. T. (2006). The impact of rolling resistance
https://open.alberta.ca/dataset/c341d72a-c424-4d6d-8c64-4ff250e50775/ on fuel, speed and costs. Continuous Improvement Case Study, 2(1), 68–
resource/4d67f16d-21b5-4bf6-b7d0-ec2ebfc66185/download/fiscal-plan- 75.
complete.pdf Watson, J. G., Chow, J. C., Wang, X., Zielinska, B. K., Kohl, S. D., &
Government of Alberta (2019a). Carbon levy and rebates. Retrieved Gronstal, S. B. (2013). Characterization of real-world emissions from
from https://www.alberta.ca/climate-carbon-pricing.aspx nonroad mining trucks in the Athabasca oil sands region during Sep-
Government of Alberta (2019b). Output-based allocation system tember, 2009 (Report). Prepared for the Wood Buffalo Environmental
engagement. Retrieved from https://www.alberta.ca/output-based-allo- Association. Reno, NV: Desert Research Institute.
cation-engagement.aspx Weyant, S. (2010). Regulatory impact on diesel engine emissions. Pro-
Government of Alberta (2019c). Capping oil sands emissions. Retrieved ceedings of the AISTech 2010 Iron and Steel Technology Conference,
from https://www.alberta.ca/climate-oilsands-emissions.aspx Pittsburgh, PA, 1355–1357.
200 MINERAL PROCESSING
Material handling and flowsheet selection considerations

for particle-based ore sorting
B. Foggiatto, J. T. Ross, T. Donkin, and B. Patterson
Ausenco Engineering Canada, Vancouver, British Columbia, Canada
M. I. Guthrie
Ausenco Limited, Perth, Australia
G. Lane
Ausenco Limited, Brisbane, Australia
ABSTRACT Interest in the use of particle-based ore sorting has increased significantly with recent
advances in sorting technologies and higher capacities per unit. Particle-based ore sorting has the poten-
tial to provide great value in processing lower grade ores from mines while delivering higher grades to
the mill, as well as in upgrading waste rock into ore. Typically the focus is on the amenability of the ore
to ore sorting and associated reductions in energy, water, and consumables utilization. However, layout
significantly affects the installed cost of a processing plant, and the impacts on layout from fitting an ore
sorting circuit into the plant plan need to be carefully considered. This paper examines processing plant
layout considerations that play an important role in ore sorting projects, including additional crushing
and screening requirements, transport of rejected material back to the waste stockpile (mine fleet require-
ments), and effects on the processing plant footprint.
■ KEYWORDS Cold climates, Ore sorting, Plant layout
RÉSUMÉ L’intérêt porté à l’utilisation du tri granulométrique du minerai a considérablement augmenté au

vu des récents progrès réalisés en matière de technologies de triage et des capacités supérieures par unité.
Le tri granulométrique du minerai offre une plus-value intéressante en termes de traitement du minerai à
plus faible teneur extrait des mines ; il permet par ailleurs l’acheminement vers le concentrateur de minerai
à plus haute teneur et la revalorisation des stériles dans le minerai. On met généralement l’accent sur la sus-
ceptibilité du minerai au triage et sur les réductions connexes en termes d’utilisation de l’énergie, de l’eau
et de biens consommables. Cependant, l’aménagement physique de l’équipement a une incidence non
négligeable sur le coût installé d’une usine de concentration, et les répercussions sur l’aménagement de
l’installation d’un circuit de triage du minerai dans l’usine doivent être soigneusement examinées. Cet arti-
cle étudie les considérations relatives à l’aménagement de l’usine de traitement qui jouent un rôle important
dans les projets de triage du minerai, et notamment les exigences supplémentaires en matière de concassage
et de criblage, le transport de matériau résiduel jusqu’à la pile de stockage des déchets miniers (nombre de
camions requis dans le parc minier) et les effets sur l’empreinte de l’usine de traitement.
■ MOTS CLÉS aménagement de l’usine, climats froids, triage du minerai
INTRODUCTION
Ore sorting has been practised throughout history. An Modern ore sorting is usually a dry process, in which a
early report by Agricola in 1556 documents how metallic portion of the material is rejected based on non-destructive
and non-metallic minerals were extracted from rocks by sensor measurements of mineral properties and/or grade
laboriously breaking them with hammers and picks and (mineral properties are sensed, and grade is inferred). The
hand-picking selected material with higher grade (Hoover methods used for sensing ore particles cover a large range of
& Hoover, 1950). The concept has not changed, but sorting physical properties. A variety of sensors are available that
technology has evolved significantly over the last 20 years. differ in terms of discrimination technique. These include

Material handling and flowsheet selection considerations for particle-based ore sorting 201
photometry, passive and active radiometry, magnetic sus- mechanical handling and separation systems and lower
ceptibility, microwave transmission, X-ray transmission cost, high-powered computing equipment that enables
(with or without luminescence), and ultraviolet fluorescence enhanced selection algorithms, and faster sensor and actua-
(Sivamohan & Forssberg, 1991; Wotruba & Riedel, 2006). tor responses (Carrasco-Tapia, 2013).
According to Nadolski, Klein, Hart, Moss, and Elmo
(2018), the feasibility of sorting is determined through anal- Considerations for low-grade orebodies
ysis of grade heterogeneity within the orebody at the scale The challenge for large, low-grade orebodies is to crush
of the intended sorting unit size: greater heterogeneity rep- and sort at a coarse rock size to minimize specific energy
resents more opportunity to add value to a project. consumption in comminution and decrease fine waste dis-
Ore sorting provides a great opportunity to recognize dif- posal costs. Foggiatto, Bueno, Lane, McLean, and Chan-
ferences in particle properties and reject particles at a rela- dramohan (2014) analyzed the economic impact of bulk ore
tively coarse size prior to energy-intensive comminution. sorting in a hypothetical, large mineral processing opera-
The main benefits that can be gained through the sorting pro- tion where the value of ore sorting was estimated using a
cess are higher production/revenue and accelerated returns simple techno-economic analysis. The study indicated that
on investment (Wyman, 1985), as well as smaller down- the sorting stage must be highly efficient to add value and
stream processing equipment in greenfield projects. In addi- compensate for the capital investment. The technical driver
tion, marginal or sub-economic deposits that are amenable to is the sorting performance, that is, mass rejection and metal
ore sorting may be upgraded, potentially making them eco- content in the waste material.
nomic to treat and thus improving resource utilization. Cut-off grade to the sorting system can be lowered,
expanding the reserve base of existing deposits and often
Ore sorting techniques offsetting metal losses in the existing resource. Economic
The two main ore sorting techniques currently practised drivers include revenue from improved head grade and costs
are bulk and particle-based sorting. Bulk ore sorting uses a associated with sorting infrastructure capital, operations,
combination of an analyzer and diversion system whereby maintenance, waste material handling, and waste disposal.
barren gangue from a fully loaded conveyor belt or shovel Additionally, layout significantly affects the installed cost of
bucket is rejected based on the grade as measured or inferred a processing plant; layout impacts from fitting an ore-sort-
from a sensor measurement. The bulk sorting unit size is ing circuit into the plant plan are typically overlooked.
directly related to the practical response time of sensors and
diverters—typically in the range of 10–200 t (Nadolski et al., Objectives
2018). During particle-based ore sorting, particles are ini- This paper examines layout considerations that play an
tially examined by sensors to measure properties and com- important role in particle-based ore sorting projects, includ-
pare them to site-specific criteria. Afterwards, individual ing crushing and screening requirements, transport of
particles are separated through an external physical force into rejected material back to the waste stockpile (additional
different products according to the measured properties.
The major advantage of particle-based sorting is the
ability to be highly selective in terms of waste rejection on
a particle-by-particle basis. For applications where particles
have a narrow size range, interparticle heterogeneity is
required, as is a significant mass rejection of subeconomic
waste. In comparison, bulk sorting can provide higher
capacities, requiring fewer units for higher throughputs.
The existing particle sorting machines working in the range
of 10–100 mm typically require closely sized feed for effi-
cient operation and depending on the sensing method, par-
ticles may also need to be washed prior to sorting. Bulk
sorting does not have the same restrictions.
Salter and Wyatt (1991) proposed that particle-based
sorting can be divided into four interactive stages: particle
presentation, particle examination, data analysis, and parti-
cle separation (Figure 1).
Particle-based ore sorting technologies have been used
predominantly on small projects, enabling the upgrade of
marginal ores and/or mineralized waste. More recently,
effective particle ore sorting at relatively high tonnage rates Figure 1. Particle-based ore sorting stages and interactions; adapted from
has been drawing attention as a result of improvements in Salter and Wyatt (1991)

202 B. Foggiatto, J. T. Ross, T. Donkin, B. Patterson, M. I. Guthrie, and G. Lane
mine fleet versus conveying), and effects on waste stock- tence) must be understood in order to identify the stages of
pile and processing plant footprints. crushing and technology required, which in turn determine
the layout considerations.
FACTORS THAT INFLUENCE PARTICLE ORE The layout of the multi-stage crushing circuit prior to
SORTING PLANT LAYOUT AND COST ore sorting should emphasize compactness and a small
The design of particle sorting systems is in essence a footprint. Cognizance needs to be given to site topography,
materials handling exercise requiring multiple conveyors with a focus on minimizing the number of conveyors and
and intermediate storage. The circuit needs to be designed transfer points while retaining the functionality of the pro-
to deliver the maximum amount of material in the selected posed process flowsheet.
size range for sorting in order to maximize the value of Cone crushers and mineral sizers are typically consid-
sorting. The main focus is the sorter feed preparation cir- ered for secondary/tertiary crushing duty. Sizers are gener-
cuit: the key to successful design and operation of an ore ally restricted to processing less competent and less
sorting facility is the feed presentation to the sorter, which abrasive ores. Both mineral sizers and cone crushers can
should provide effective screening of fines and oversize minimize fines production when operating conditions are
material to produce a narrow particle size distribution tailored to allow them to do so (example in Figure 2).
(PSD) with a maximum top size to bottom size ratio of 3:1
(e.g., 30:10 mm). The financial benefits of sorting can be
negatively affected where fines generation or ore mineral-
ogy vary significantly. The crushing circuit design must
include allowances for ore variability, allowing some ore
types to bypass the ore sorting facility.
The choice of location for an ore sorting plant requires
thorough analysis. Retrofitting an ore sorting circuit into a
brownfield plant presents far greater challenges than
designing an ore sorting circuit for a greenfield plant. Sites
are not generally constructed with excess earthworks avail-
ability, which may ultimately affect the design and layout
of a particle-based ore sorting circuit. For example, a circuit
at a brownfield site may require longer conveyors and
transport distances based on the space available for new
plant construction pads. Figure 2. Particle size distributions of products from a cone crusher and a
mineral sizer processing phosphate ore
Feed preparation
The role of the primary crushing stage is to reduce particle • The cone crusher design should aim for a low reduction
size to allow transport by conveying and provide a suitable ratio. Minimizing fines production and maximizing pro-
feed for secondary and possibly tertiary crushing stages, duction of intermediate product can be achieved by
which yield a product suitable for screening and subse- reducing cone speed, using a coarse liner (wider feed
quently sorting. Minimizing fines production is a key goal of opening, wider closed size setting), or not choke-feeding
all crushing stages to limit fines grade dilution and maximize the crusher.
production of intermediate (10–75 mm) particles, which are • Wear rates of sizers in secondary/tertiary crushing duties
subsequently fed to the sorting stage. Particle sorting is only can be very high and not cost effective, thus requiring
efficient within certain feed particle size ranges. If particles thorough ore characterization. An advantage of the
are too fine, throughput decreases and specific costs rise; if duty/standby secondary arrangement for mineral sizers
particles are too coarse, selectivity decreases. The design of is that the sizers are mounted on rails and are used on a
the crushing stages should aim to produce the correct particle roll in/out basis. Implementing a second sizer unit that
size while keeping fines production to a minimum. can be easily rolled into position allows the total sizer
Ore sorting performance can also be negatively affected if plant availability to increase in line with the ore sorter
particle size and shape vary significantly. Ore variability can units and other plant equipment.
result in varying PSD of the crushing circuit product and thus Screening efficiency also plays an important role in feed
in the proportion of material suitable for ore sorting. This is presentation to the sorting stage. Insufficient screen area
considered an important factor affecting the overall project will likely result in fine particles reporting to the screen
financial performance. Material characterization and crush- oversize, which reduces sorting performance. Incorrect use
ing testwork improves confidence in crushing technology of sprays on screens for particle surface washing and dust
selection based on material-specific characteristics. The control can adversely affect the downstream performance
material characteristics (e.g., strength, hardness, compe- of the material handling systems and ore sorting to a degree

that depends on the amount of water added, air temperature, throughput circuits where multiple ore sorters are required
and ore characteristics (e.g., pre-existing water content). have a single long bin or several feed surge bins; evenly
Additional moisture introduced from screen sprays in cold distributing feed among them can be challenging. Selection
environments and the presence of more fines and clays will of feed surge bin arrangement depends on site conditions,
cause material buildup on screen decks, conveyor compo- with impacts on both capital and operating expenditures.
nents, and transfer points, resulting in blockages and more Table 1 compares three types of feed surge bins for a mul-
downtime events. Thus, it is important to correctly select tiple ore sorter arrangement. Note that bin height is an
screen type, area, and aperture. In addition, a suitable important design criterion because the lower the material
enclosure over the screen (in particular at cold climate drop height, the fewer fines generated in the sorter feed.
sites) can help to mitigate these problems. The undersize The arrangement of the ore sorter units should be con-
fraction of the screened material can be directed to the final sidered because this can be a large capital cost driver, espe-
concentrate or waste storage as dictated by the metallurgi- cially in cold climate environments where an enclosure
cal and operations requirements at the site. would be required. In a multiple ore sorter arrangement, the
two main approaches to layout are in-line and staggered
Feed presentation and particle separation (locating the sorters on opposite ends of the bin). The latter
The feed presentation to ore particle sorting is very can be a viable option at sites in warm climates and if the
important because to work effectively, ore sorter units waste and product are being stockpiled and collected by
require a single layer of sized particles. The PSD needs to be mobile equipment. However, in nearly every other sce-
within a narrow optimal range for the operating settings of nario, staggering the ore sorters will drive up capital costs
the equipment sensors and mechanical sorting mechanisms because multiple buildings and additional transfer convey-
(i.e., compressed air nozzles). The optimal size range is gen- ors are required. Further, a staggered arrangement limits
erally a 3:1 ratio of top to bottom size. For example, a top access to equipment for maintenance.
size of 75 mm should have a target bottom size of 25 mm.
Large amounts of oversized material presented to ore sort- Waste and concentrate transport
ing units can result in inefficient sorting and increased Transport of material identified as waste or concentrate
equipment wear—if not damage. The presence of undersize by the ore sorter units can be a key economic driver of an
material that cannot be separated can contaminate sorting ore sorting project. Selecting from among the various prod-
products, decrease crushing circuit availability, and dramat- uct transport methods may largely depend on site constraints
ically increase the manpower required for material handling. that limit product transport choices if an inadequate foot-
For these reasons, careful consideration is required to print is available or if operational constraints exist (Table 2).
upstream comminution, screening, and material handling
processes and equipment selection. Considerations for cold climates
Depending on the size fractions to be sorted and in order The extra challenges associated with including an ore
to provide continuous feed to the sorters, separate feed stor- sorting circuit in a cold-climate project will likely increase
age for the different size ranges is required upstream of the capital costs. If considerations are not made, circuit opera-
sorting units. The goal is the shortest practical sorter tions will be negatively affected, and remedial fixes can be
bin/stockpile feed retention time (based on the crushing cir- costly. Buildings may be required to ensure suitable opera-
cuit availability) to limit project capital expense require- tion of equipment. Depending on the harshness of the site
ments and create a smaller plant footprint based on the bin conditions, the feed preparation circuit may not require
feed conveyor. A steady feed rate to the ore sorting units is building enclosures but at a minimum, the ore sorter units
required, necessitating provision (generally a feed bin) for should be enclosed because some sorter components are not
surge capacity. Low-throughput ore sorting circuits have a rated for use below 5°C. In addition to the building require-
simplified arrangement, with a belt conveyor discharging ment, other capital and operating cost impacts to consider
into the feed surge bin. By comparison, layouts of high- include:
Table 1. Advantages and disadvantages of three types of feed surge bins
Radial stacker conveyor Reversing shuttle conveyor Single belt conveyor
Advantages Low overall height Flexibility to feed any number of bins Low capital cost
Low capital cost Central bin discharge Little mechanical equipment
Low operating cost
Disadvantages Safety concerns with continuously High capital cost Greater bin feed height required as
slewing conveyor High operating cost a result of transfer height
Cold climate concerns with slewing base Greater bin feed height required as a result Unable to feed a large number of
Unable to feed a large number of bins of single chute discharge to multiple bins bins

204 B. Foggiatto, J. T. Ross, T. Donkin, B. Patterson, M. I. Guthrie, and G. Lane
Table 2. Advantages and disadvantages of three waste and concentrate transport methods
Loadout bin station Stockpile and mobile equipment Overland belt conveyor
Advantages Low operating cost Low capital cost No need for mobile equipment operation
Limited double handling of product Small plant footprint
material Flexibility for stockpile expansion if truck
Little dust relative to open stockpiles cycle time increases
Few safety concerns from continuously
operating mobile equipment
Disadvantages High capital cost Double handling of product material High capital cost
Large plant footprint due to feed bin Safety concerns from operating mobile Complications to existing plant tie-in
conveyor equipment beneath discharging conveyor Challenges with conveyor arrangement on
Impact on sorter operation if bin sized brownfield sites
inadequately for truck cycle time
Potential for rat-holing* and material
hang-up in bin
* Feed forms a hole or narrow channel, while the remaining feed is compressed against the bin wall
• heating, ventilation, and air-conditioning costs to sup- performance, as well as identify and quantify losses.
plement the equipment latent heat and ensure suitable Thus, sampling of waste streams is necessary and prod-
operating temperatures; uct stream sampling is advised. Cross-stream samplers
• foundation design to prevent/mitigate frost heave; should be considered to take samples at a suitable time
• bin design to prevent ore from freezing; interval set by the operator in the control room. Large
• belt conveyor snow/wind covers; sample masses are needed to account for the heterogene-
• site-run piping insulation and heat tracing; and ity of the material (high fundamental sampling error),
• potential cover for vibrating screen deck. which requires a dedicated handling system.
Aside from the obvious capital cost impacts from oper- • Material handling testwork is needed to identify poten-
ating in a cold climate, there are operational factors to con- tial conveyor alternatives, such as high lift or cleated
sider: conveyors, to yield a smaller circuit footprint or to fit
• Cold temperatures and excess moisture in a poorly within a constrained brownfield site.
designed bin can result in rat-holing and material • Suitable layout of access platforms and crane around the
buildup, severely shortening retention times. ore sorters is needed to ensure safety, accessibility,
• Cold temperatures can prevent the use of water sprays ergonomics, and maintenance access without the use of
for circuit-wide dust mitigation. mobile equipment.
• Frozen surfaces can be a problem if sensors require
material to be washed prior to sorting. CONCLUSIONS
• Uncovered material stockpiles can freeze and require The most significant economic benefits from particle-
mobile equipment intervention. based ore sorting are reaped when sensor-based ore sorting
converts uneconomic waste rock into ore and when very
Other considerations high upgrades and recoveries are achieved. Key to the suc-
Other important considerations include: cessful design and operation of an ore sorting facility is
• Additional air services (e.g., compressed air of instru- optimal feed presentation to the sorter, as well as mainte-
ment air quality) must be considered for ore particle nance of the sensor system, with the intent of providing
sorting. The compressed air circuit comprises oil-free homogeneous feed to ore sorters while sustaining consis-
centrifugal compressors, air dryers, and air receivers. tent sorting performance. Key to the efficient performance
Plant air is used to service the ore sorter units, as well as and operation of particle ore sorting is selecting a layout
several utility stations for maintenance. that does not reduce the overall availability of the system.
• Dust generation in multiple stages of crushing, screen- The main selection criteria include the material prepara-
ing, and sorting needs to be controlled, especially in tion (i.e., crushing/sizing selection), material screening,
enclosed buildings. Dust extraction points at the sorting surge capacity, material presentation to the ore sorter
machine are required. unit(s), and storage of the concentrate and waste streams.
• Local regulatory and environmental permits may be The type of crushing technology, number of crushing
required due to dust and noise generation. stages, and subsequent layout of the crushing area are func-
• X-ray permitting may be needed depending on the sen- tions of the need to ensure minimal fines generation and a nar-
sor types in the ore sorting unit. row PSD for effective particle sorting, with a maximum top
• Metallurgical accounting is critical for sensor monitor- size to bottom size ratio of 3:1. Effective screening is required
ing, sensor data verification, and to improve ore sorting to remove all material smaller than the minimum particle size

required for the sorter feed. Focus should be on minimizing An earlier draft of this paper was published in the
the footprint and the number of transfer points for the crush- Proceedings of the 2019 Annual Meeting of the Canadian
ing and screening areas that precede the ore sorting stage. Mineral Processors, prior to undergoing the CIM Journal
Enclosing the screening area should be considered if peer-review process.
operating in freezing conditions to avoid problems arising
from increased material moisture from spray systems. In Paper reviewed and approved for publication by the
addition, the sorter units must be housed in a weather-pro- Canadian Mineral Processors Society of CIM.
tected building warmer than 5°C and free from dust to Bianca Foggiatto has more than 14 years of experience in process
ensure sensors integral to the sorter units operate properly. engineering. She completed her PhD at the University of Queensland,
Orientation and placement of multiple sorting units will be Australia, specializing in modelling and simulation of processing circuits,
comminution circuit design, and energy efficiency. Her experience includes
dictated by the necessity to ensure even feed and the asso- projects related to most key mining commodities. She has been responsible
ciated capital costs for the building footprint. for comminution circuit design, plant debottlenecking, and process
Surge capacity is required prior to feeding the particle optimization. bianca.foggiatto@ausenco.com
sorting units to ensure consistent, even feed. Feed surge Jordan T. Ross is a lead designer with over 13 years of professional
bins can be used; however, the feed method to the feed experience. He currently works for Ausenco as a key member of the Minerals
surge bins must ensure even distribution of material if mul- & Metals team. He specializes in plant layout and mechanical design and
successfully leads multi-disciplinary design teams to deliver cost-effective
tiple ore sorting units are installed downstream. Several plant designs.
methods of storage and transport of the concentrate and
waste streams are currently available. The choice of method Tim Donkin has a degree in mechanical engineering and 8 years of
experience working for Ausenco in the Minerals & Metals team. Tim has
is influenced by site-specific requirements, such as mobile been involved in the delivery of mining projects from the prefeasibility stage
equipment needs, climate, topography, and proximity to the through to engineering, procurement and construction management. He has
mine stockpiles and processing plant. The storage and experience in construction and commissioning, as well as operations.
transport method selected will factor into the location, ori- Ben Patterson has over 16 years of experience in management, engineering,
entation, and layout of the sorting plant. construction, maintenance, and operations within the mining and oil and gas
sectors. His core skills include project management, commissioning, planning,
Ore sorting can yield significant economic benefits to an field engineering, materials handling design, construction management, cost
operation and reduce sensitivity to fluctuating metal prices. estimating, pre-contracts, contract management, and procurement.
However, sensitivity to ore variability may increase if gen-
Malcolm I. Guthrie has 10 years of mineral processing and extractive
eration of fines in the crushing circuit varies and large metallurgy experience, ranging from testwork management through to
amounts of material bypass the ore sorting circuit. Highly engineering, procurement, commissioning, and operations. He has also
variable ore mineralogy negatively affects sorter perfor- served as the project engineer for various projects in which his experience
extends to the construction phase. Malcolm’s core expertise includes
mance and requires recalibration of the sensor system. commodities such as gold, copper, nickel, and iron ore, with a focus on
production management, engineering design, and technical support.
ACKNOWLEDGMENTS
Greg Lane has over 30 years of experience in the field of engineering,
The authors wish to thank Ausenco for permission to design, and operation of minerals processing plants. He is a director of NQ
publish this paper. The Canadian Mineral Processors Soci- Minerals PLC and chief technical officer for Ausenco. Greg is an
ety volunteer peer reviewers are also acknowledged for internationally recognized expert in concentrator design and has prepared
and presented numerous papers on minerals processing and concentrator
their contributions that improved the quality of this paper. design, including many on the selection of efficient processes and capital cost
effectiveness in design.
REFERENCES
Carrasco-Tapia, C. E. (2013). Development of geometallurgical tests to the Fourth International Symposium on Block and Sublevel Caving (pp.
identify, rank, and predict preferential coarse size by size Au deportment 133–140). Perth, Australia: Australian Centre for Geomechanics.
to support feed preconcentration at Telfer Au-Cu mine, Newcrest Western
Salter, J. D., & Wyatt, N. P. G. (1991). Sorting in the minerals industry:
Australia (Master’s thesis). The University of Queensland, Brisbane, Aus-
tralia. Retrieved from https://espace.library.uq.edu.au/view/UQ:310904 Past, present and future. Minerals Engineering, 4(7–11), 779–796.
https://doi.org/10.1016/0892-6875(91)90065-4
Foggiatto, B., Bueno, M., Lane, G., McLean, E., & Chandramohan, R.
(2014). The economics of large scale ore sorting. Proceedings of the Sivamohan, R., & Forssberg, E. (1991). Electronic sorting and other
XXVI International Mineral Processing Congress (IMPC 2014). Santi- preconcentration methods. Minerals Engineering, 4(7–11), 797–814.
ago, Chile: Gecamin. https://doi.org/10.1016/0892-6875(91)90066-5
Hoover, H. C., & Hoover, L. H. (1950). Georgius Agricola de re Metal- Wotruba, H., & Riedel, F. (2006, March). Sensor based sorting of met-
lica [Georgius Agricola subject metals]. Translated from the Latin edi- alliferous ores – An overview. Paper presented at Sensorgestützte
tion of 1556. New York: Dover Publications. Sortierung, Aachen, Germany.
Nadolski, S., Klein, B., Hart, C. J. R., Moss, A., & Elmo, D. (2018). An Wyman, R. A. (1985). Sorting by electronic selection. In N. L. Weiss
approach to evaluating block and panel cave projects for sensor-based (Ed.), SME Mineral Processing Handbook. Englewood, CO: Society for
sorting applications. In Y. Potvin & J. Jakubec (Eds.), Proceedings of Mining, Metallurgy and Exploration, Inc.
CIM JOURNAL
Volume 10, Number 4, 2019
Contents
Breaking down energy consumption in 157 J. Bouchard, G. LeBlanc, M. Levesque,

industrial grinding mills P. Radziszewski, and D. Georges-Filteau
Hydraulic excavator versus electric rope 165 R. Andreev, T. G. Joseph, J. Sammut,

shovel performance and M. Curley
Performance prediction of a full-size rapid 173 R. K. Soni, T. G. Joseph,

dampening suspension in an ultra-class and M. Curley
hauler application
Is bigger still better? Considerations in 180 M. Dotto, T. G. Joseph, and M. Curley

increasing the size of haulage equipment
Impact of operating conditions on emissions 187 R. J. Schafer-Frentz, T. G. Joseph,

from mine haul trucks operating in the oil and M. Curley
sands region of Alberta, Canada
Material handling and flowsheet selection 200 B. Foggiatto, J. T. Ross, T. Donkin,

considerations for particle-based ore sorting B. Patterson, M. I. Guthrie, and G. Lane
Published by the Publié par l’Institut

Canadian Institute of canadien des mines,
Mining, Metallurgy de la métallurgie
and Petroleum et du pétrole

CIM Journal Vol. 10 No. 4

Uploaded by

Copyright:

Available Formats

You might also like

CIM Journal Vol. 10 No. 4

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

CIM Journal Vol. 10 No. 4

Uploaded by

Copyright:

Available Formats

CIM JOURNAL

Editors International Advisory Board

© 2019 Canadian Institute of Mining, Metallurgy and Petroleum

Published quarterly by the Canadian Institute of Mining, Subscription rates

157 Breaking down energy consumption in industrial grinding mills

165 Hydraulic excavator versus electric rope shovel performance

173 Performance prediction of a full-size rapid dampening suspension in an ultra-class hauler

200 Material handling and flowsheet selection considerations

Breaking down energy consumption in industrial grinding mills

CIM Journal | Vol. 10, No. 4, 2019

CIM Journal | Vol. 10, No. 4, 2019

Quantifying work output

CIM Journal | Vol. 10, No. 4, 2019

CIM Journal | Vol. 10, No. 4, 2019

Table 2. Carnot efficiency data

Canadian Malartic 34,623 30,739 36 8.41

New Afton 9,917 8,865 18 2.75

CIM Journal | Vol. 10, No. 4, 2019

CIM Journal | Vol. 10, No. 4, 2019

Jocelyn Bouchard is currently an assistant professor in the Department of

CIM Journal | Vol. 10, No. 4, 2019

Hydraulic excavator versus electric rope shovel performance

R. Andreev, T. G. Joseph, J. Sammut, and M. Curley

CIM Journal | Vol. 10, No. 4, 2019

Comparison of dig performance

CIM Journal | Vol. 10, No. 4, 2019

The HE excavating approximately 76 t per 60 s pass and

CIM Journal | Vol. 10, No. 4, 2019

CIM Journal | Vol. 10, No. 4, 2019

In the case of availability, there were no discernable dif-

RESULTS AND ANALYSIS

CIM Journal | Vol. 10, No. 4, 2019

Productivity be needed to match the 25 year life point capability of a

CIM Journal | Vol. 10, No. 4, 2019

Figure 10. Relationships from Figure 9 applied to operating plus

Figure 12. Productivity and maintenance costs (Canadian dollars) versus

CIM Journal | Vol. 10, No. 4, 2019

Tim G. Joseph, PhD, P.Eng., FCIM, is associate dean of engineering and

Performance prediction of a full-size rapid dampening suspension

Research basis for the paper

CIM Journal | Vol. 10, No. 4, 2019

CIM Journal | Vol. 10, No. 4, 2019

Scale laboratory tests

Rapid dampening strut stiffness

CIM Journal | Vol. 10, No. 4, 2019

For comparison, Figure 5 shows the stiffness curve of

correlated, and the relative performance ratios were applied

CIM Journal | Vol. 10, No. 4, 2019

column in the cylinder and the external force

where P is pressure (kPa), V is volume (m3) at

FMOD,F = 106P –1.39FMOD,S (6)

where P is the percentage closure and FMOD,F and

CIM Journal | Vol. 10, No. 4, 2019

figuration (Figure 11). The difference between the

CIM Journal | Vol. 10, No. 4, 2019

Is bigger still better?

CIM Journal | Vol. 10, No. 4, 2019

Table 1. Hauler specifications; GW: gross weight (Caterpillar, 2012)

Table 2. Construction material specification