Mineral Processing and Extractive Metallurgy

Transactions of the Institutions of Mining and Metallurgy

ISSN: 2572-6641 (Print) 2572-665X (Online) Journal homepage: http://www.tandfonline.com/loi/ympm21

Process simulations in mineralogy-based

geometallurgy of iron ores

Mehdi Parian, Pertti Lamberg & Jan Rosenkranz

To cite this article: Mehdi Parian, Pertti Lamberg & Jan Rosenkranz (2018): Process simulations
in mineralogy-based geometallurgy of iron ores, Mineral Processing and Extractive Metallurgy, DOI:
10.1080/25726641.2018.1507072

To link to this article: https://doi.org/10.1080/25726641.2018.1507072

© 2018 The Author(s). Published by Informa

UK Limited, trading as Taylor & Francis
Group

Published online: 09 Aug 2018.

Submit your article to this journal

Article views: 24

View Crossmark data

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=ympm21
MINERAL PROCESSING AND EXTRACTIVE METALLURGY
https://doi.org/10.1080/25726641.2018.1507072

Process simulations in mineralogy-based geometallurgy of iron ores

a
Mehdi Parian , Pertti Lambergb and Jan Rosenkranza
a
Minerals and Metallurgical Engineering, Luleå University of Technology, Luleå, Sweden; bKeliber Oy, Kaustinen, Finland

ABSTRACT ARTICLE HISTORY

Mineral processing simulation models can be classified based on the level that feed stream to Received 8 February 2018
the plant and unit models are described. The levels of modelling in this context are: bulk, Revised 22 July 2018
mineral or element by size, and particle. Particle level modelling and simulation utilises Accepted 25 July 2018
liberation data in the feed stream and is more sensitive to the variations in ore quality,
KEYWORDS
specifically ore texture. In this paper, simulations for two texturally different magnetite ores Geometallurgy; mineral
are demonstrated at different modelling levels. The model parameters were calibrated for liberation; particle-based
current run-of-mine ore and then in the simulation applied directly to the other ore. For the simulation; process
second ore, the simulation results vary between the different levels. This is because, at the modelling
bulk level, the model assumes minerals do not change their behaviour if ore texture or
grinding fineness are changed. At the mineral by size level, the assumption is that minerals
behave identically in each size fraction even if the ore texture changes. At the particle level,
the assumption is that similar particles behave in the same way. The particle level approach
gives results that are more realistic and it can be used in optimisation, thus finding the most
optimal processing way for different geometallurgical domains.

Introduction least detailed level, the entire processing circuit (black

box model) is modelled in a single operation, sections
Variation in the quality of feed to a plant has been a
(e.g. comminution circuit and flotation circuit) are used
challenge for mineral processing operations. A prior
at a moderate level, and single unit operations (e.g. indi-
quantitative understanding of the feed characteristics
vidual flotation cells) are used at the detailed level. In geo-
over the lifespan of the mining operation is the solution
metallurgy, it is common to use models, which forecast
to successful production planning, circuit design,
the full process with simple equations. For example for
optimisation, and troubleshooting. Furthermore, an
Hannukainen iron ore deposit, the iron recovery into
effective utilisation of ore body and proper risk man-
the iron concentrate has been defined with a simple
agement in the mining industry is also an interest. Geo-
equation based on iron and sulfur head grade (Equation
metallurgy utilises geological and mineral processing
(1), SRK Consulting (2014)), and similar equations have
information to capture the spatial metallurgical charac-
been proposed for copper and gold as well.
teristics of the ore body.
Mineralogy-based geometallurgical modelling RFe = 98.5 × (1 − e0.06×(Fe−6) ) × ( − 1.96
employs quantitative mineralogical information, both
× (S/Fe) + 1) (1)
on the deposit and in the process. The geological
model must describe the minerals present, give their Models describing the process section wise are also com-
chemical composition, report their mass proportions mon especially in comminution circuits when throughput
(modal composition) in the ore body and describe or comminution energy usage is being forecast. Geome-
the ore texture. The process model must be capable tallurgical models going to unit operation level are rare.
of using mineralogical information provided by the This is due mainly to the difficulty of obtaining quantitat-
geological model to forecast the metallurgical perform- ive results from geology to work at unit level, as well as the
ance of different geological volumes, such as samples, lack of an interface to connect geological data to the min-
ore blocks, geometallurgical domains or blends pre- eral processing unit level.
pared for the plant feed for different periods from From another perspective, the models used in the
hourly and daily scale to weekly, monthly and annual mineral processing simulation are based on the level
production (Figure 1). that the feed stream to the plant and unit operations
As to the level of detail, predictive models of mineral are described. The flowrates information in a stream,
processing can be divided into three categories based on from lowest to the highest level of detailedness, is
the size of the smallest block in the simulation. In the total solids, chemical elements, minerals, and particles.

CONTACT Mehdi Parian [email protected] Minerals and Metallurgical Engineering, Luleå University of Technology, Luleå SE-971 87, Sweden
© 2018 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group
This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License (http://creativecommons.org/licenses/by-
nc-nd/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited, and is not altered, transformed, or
built upon in any way.
2 M. PARIAN ET AL.

Figure 1. Modelling levels based on the feed stream definition for process simulation in concept of mineralogy-based approach to
geometallurgy.

Each of these flowrates can be further assigned in cer- was to evaluate advantages and disadvantages of simu-
tain size classes in order to reach the sized level (Table lation at the different levels decribed above and high-
1). The selection of a modelling level comes with cer- light the benefits of particle-level modelling and
tain restrictions that limit its usability when expanding simulation especifically in geometallurgical studies.
it outside the scope for what the model was created and
calibrated (Table 1). The limitation of modelling level
originates from feed properties definition used at that Materials and methods
level. For example, a bulk model based on chemical
Materials
assays is no longer valid if the distribution of elements
in minerals, particle size distribution and liberation The two feeds that used for plant simulation were from
distribution of feed changes. Luossavaara-Kiirunavaara Ab (LKAB) iron ore deposit.
This study focussed on flowsheet simulation for The first data set is actual feed to the plant that were
mineralogical-based approach to geometallurgy of collected a during plant survey and analyzed in size
iron ores. Process unit models used were selected fractions as described by Parian et al. (2016). The
based on their applicability at particle level for iron second set of data is the particles from drill core crush-
ore beneficiation and flowsheet simulation. The feed ing (Parian et al. 2017). Therefore, the first feed is from
streams for plant simulation were run-of-mine feed part of the ore body that is currently being mined and
and forecast feed from ore textural study. The aim processed and the second feed is from part of the same

Table 1. Modelling levels and their assumptions in minerals processing based on the feed stream components. The model is valid
and can be extended outside the calibration point as long as the feed stream assumptions are satisfied.
Level Feed stream assumptions Example of application
Bulk Solids Each particle will behave identically regardless of its size and Solid splitter model, equipment scale-up, basic engineering
composition
Element Element disttirbution in minerals, particle size distribution, Recovery function model for elements
and liberation distribution are fixed
Mineral Particle size distribution and liberation distribution are Mineral splitter model
unchanged
Behavioural Particle size distribution is unchanged. Similar behavioural Gravity separation model
type types will behave identically
Sized Solids Each particle of given size will behave identically Comminution model, Size classification model, equipment scale-
up
Grindability Each particle type of given size will behave identically Comminution model
type
Element Element disttirbution in minerals and liberation distribution Recovery function model in size classes
are unchanged
Mineral Liberation distribution is unchanged Mineral splitter model in size classes
Behavioural Similar behavioural types will behave identically in narrow Flotation model (fast and slow floating materials in size classes),
type size fraction Gravity separation model in size classes
Particle Similar behaviour of particles of same composition and size Principally applicable to all unit models
 MINERAL PROCESSING AND EXTRACTIVE METALLURGY 3

Figure 2. The flowsheet of magnetite beneficiation plant for demonstration of different levels of modelling and simulation.

ore body that is to be mined in the near future. From a simulation software HSC Chemistry 9.2 by Outotec.
geological point of view, the feeds have similar miner- The process unit models used in the simulation at
alogy; however, they are texturally and compostionally different modelling levels are gathered in Table 2
different. The ore mineral in both is magnetite, which is with their respective references. The unit model par-
associated with apatite, quartz, amphiboles, biotites ameters were adjusted to match experimental and
and fedspars. mass balancing results.

Plant circuit and unit models

Results and discussion
The flowsheet used here consists of comminution and
Modal mineralogy and liberation state of feed
concentration circuits and is based on the LKAB Kir-
streams
una concentration plant (Söderman et al. 1996; Sams-
kog et al. 1997). The circuit is a general form of the The run-of-mine feed to the plant (ROM) and particles
magnetite beneficiation plant that LKAB uses in Kir- from crushing magnetite drill core sample (DCS) was
una. The feed to the plant is the product of crushing characterised in details in order to be used for simu-
and cobbing plant. The beneficiation process comprises lation. The two samples are compositionally different
closed circuit autogenous grinding followed by primary (Table 3) and they also have different liberation distri-
magnetic separation and pebble milling coupled with bution (Figure 3). The DCS sample represents the ore,
hydrocyclone followed by apatite flotation and mag- which is significantly lower in head grade, and ore tex-
netic separation (Figure 2). The flowsheet was used to ture is fine-grained giving lower liberation degree at
demonstrate different levels of modelling according given particle size. The assumptions used here are
to Table 1 and was validated against known metallurgi- that composition of minerals and particle size distri-
cal performance of the plant. bution after crushing for both ores are the same. In
Mass balancing of streams at mineral and particle addition, for simplicity, different minerals of the
levels, data reconciliation, as well as simulations at same group were lumped together and the main signa-
different levels were done using the modelling and ture mineral was used for reporting grade and recovery.

Table 2. The unit models used in the simulation of flowsheet at different levels.
Unit Bulk model Mineral by size model Particle level model
Autogenous/ Pebble Mill Perfect mixer Rosin-Rammler distribution with mineral adjustment HSC built-in liberation model (Remes and
(Lamberg and Lund 2012) Lamberg 2017)
Trommel/ Spiral classifier/ Mineral Efficiency curve (Napier-Munn et al. 2005) Efficiency curve
Hydrocyclone splitter
Mixer Perfect mixer Perfect mixer Perfect mixer
WLIMS Mineral Mineral splitter in size fractions Particle-based WLIMS (Parian et al. 2016)
splitter
Flotation Flotation Flotation kinetics in size fractions Particle-based flotation kinetics (Savassi 2006;
kinetics Lamberg 2010; Remes 2017)
4 M. PARIAN ET AL.

Table 3. Modal composition of samples.

Magnetite Apatite Albite Biotite Quartz Actinolite
Ore (wt%) (wt%) (wt%) (wt%) (wt%) (wt%)
ROM 85.57 2.48 3.01 3.50 0.96 3.43
DCS 64.09 7.24 12.88 10.27 5.52 0.00

100 100

90 90

80 80

70 70
Cumulative yield (%)

Cumulative yield (%)

60 60

50 50

40 40

30 30
0-38 µm 0-38 µm
38-53 µm 38-53 µm
53-75 µm 53-75 µm
20 20
75-150 µm 75-150 µm
150-300 µm 150-300 µm
300-600 µm 300-600 µm
10 10
600-1700 µm 600-1700 µm
1700-3350 µm 1700-3350 µm
0 0
0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100
Magnetite (wt%) in particle Magnetite (wt%) in particle

Figure 3. Liberation curves of magnetite in the ROM (left) and DCS (right) samples.

Process simulation at different modelling level experimental data. This also validates the simulation
against current metallurgical performance of the plant.
Plant simulation with ROM feed
It is known that in the magnetite beneficiation plant
Overall, for ROM, three levels of simulation give the
of LKAB, the comminution circuit grinds the ore finer
same grade and recovery for all minerals for the final
than the liberation size of magnetite in order to satisfy
concentrate of the plant. This was expected as the
pelletising plant requirement on fineness. Therefore, it
models at different levels were calibrated to match
is of interest to know the effect of coarse grinding. This

100

98

96

94
Magnetite recovery (%)

92

90

88

86

Bulk
84
Sized
Particle
Bulk-Coarse
82
Sized-Coarse
Particle-Coarse
80
80 82 84 86 88 90 92 94 96 98 100
Magnetite grade (wt%)

Figure 4. Recovery and grades magnetite for different modelling levels in normal grinding and coarse grinding conditions (ROM
ore).
 MINERAL PROCESSING AND EXTRACTIVE METALLURGY 5

100

98

96

94
Magnetite recovery (%)

92

90

88

86

Bulk
84
Sized
Particle
Bulk-Coarse
82
Sized-Coarse
Particle-Coarse
80
80 82 84 86 88 90 92 94 96 98 100
Magnetite grade (wt%)

Figure 5. Recovery and grades magnetite for different modelling levels in normal grinding and coarse grinding conditions (DCS
ore).

has been done by increasing product particle size in simulation give more plausible results than the mineral
grinding circuit models (pebble mill and hydrocy- and mineral by size level. This is because the par-
clone). The simulation result reveals that even in coarse ameters are calibrated based on the liberation of par-
product magnetite recovery and grade is still preserved ticles instead of general calibration to the ore at bulk
(Figure 4). and size level. Decreaese in recovery of magnetite at
particle level can be linked with loss of magnetite in
Plant simulation with DCS feed tailing due to fine-grained magnetite texture that
For the simulation of DCS, the established unit model affects liberation. At the particle level, the assumption
parameters used at different levels for ROM feed in is that similar particles behave in the same way whereas
coarse and normal grinding were used. The simulation at the mineral level the assumption is that minerals
for DCS ore shows that at bulk and sized levels, mineral behave identically regardless of texture and particle
recoveries and grades have similar results. In fact, the composition.
mineral recoveries are almost the same as ROM feed.
This is due to a property of the model which defines
Conclusions
the behaviour of minerals on unsized or sized based:
regardless of the plant feed the mineral recoveries The mineralogical approach of geometallurgy requires
(and distribution in the full process) is similar, but quantitative mineralogical data for geological model to
the grades change. When the head grade decreases, be able to perform in its best form. It is common that
the concentrate grade drops as well. for estimation of plant metallurgical response black
At the liberation level, the result is, however, differ- box models or simple equations are used. However,
ent. Similar particles behave in the same way as in the this approach is not sensitive to material changes. In
base case, but because the mass proportion of particles the best case, they can adopt the changes in the head
(thus liberation distribution) is different in the plant grade but not the variation in ore textures. Within
feed for the ROM and DCS ores, the final result will this study, it was shown that simulation at mineral
differ for both grades and recoveries. The particle levels with parameters calibrated to the ore feed is
level simulation forecasts lower recovery for the DCS only valid for ore having similar liberation character-
than for the ROM feed (Figure 5). istics. However, parameters calibrated at particle levels,
Comparison of simulations for ROM and DCS that are based on the liberation of particles, are capable
samples demonstrates the difference between mineral of giving more reliable estimates. In addition, the par-
and particle level modelling and simulation. In the ticle level enables optimisation, thus finding optimal
DCS ore simulation, the results from the particle level processing conditions, e.g. grinding fineness, for
6 M. PARIAN ET AL.

different textural ore types and geometallurgical Lamberg P, Lund C. 2012. Taking liberation information
domains. This also allows deveopment of platforms into a geometallurgical model: case study Malmberget,
for establishing geometallurgical modelling and to Northern Sweden. Preprints of Process Mineralogy;
Cape Town, South Africa.
acheieve plausible results in forectasing metallurgical Napier-Munn TJ, Morrell S, Morrison RD, Kojovic T. 2005.
performance of geometallurgical volumes. Screens and hydrocyclones. Mineral Comminution
Circuits: Their Operation and Optimisation; JKMRC,
The University of Queensland. p. 273–330.
Acknowledgment Parian M, Lamberg P, Rosenkranz J. 2016. Developing a par-
ticle-based process model for unit operations of mineral
The authors acknowledge the financial support from the
processing – WLIMS. Int J Miner Process. 154:53–65.
Hjalmar Lundbohm Research Centre (HLRC) and thank
doi:10.1016/j.minpro.2016.07.001.
LKAB for providing samples and analysis. The authors
Parian M, Mwanga A, Lamberg P, Rosenkranz J. 2017. Ore
kindly thank the support from Kari Niiranen from LKAB.
texture breakage characterization and fragmentation
into multiphase particles. Powder Technol. doi:10.1016/
j.powtec.2017.12.043.
Disclosure statement Remes A. 2017. Sim mineral processing unit models. HSC
No potential conflict of interest was reported by the authors. Chemistry® User’s Guide; Outotec. p. 98.
Remes A, Lamberg P. 2017. Sim mineral processing. HSC
Chemistry® User’s Guide; Outotec. p. 38.
Funding Samskog PO, Björkman J, Söderman P, Broussaud A, Guyot
O. 1997. Model-based and expert supervisory control at
The authors acknowledge the financial support from the Kiruna LKAB concentrators - Sweden. Proceedings of
Hjalmar Lundbohm Research Centre (HLRC) and thank the XIX International Mineral processing Congress
LKAB for providing samples and analysis. (SME-AIME). p. 217–223.
Savassi ON. 2006. Estimating the recovery of size-liberation
classes in industrial flotation cells: a simple technique for
ORCID minimizing the propagation of the experimental error. Int
J Miner Process. 78:85–92. doi:10.1016/j.minpro.2005.09.
Mehdi Parian http://orcid.org/0000-0002-5979-5608 002.
Söderman P, Storeng U, Samskog PO, Guyot O, Broussaud
A. 1996. Modelling the new LKAB Kiruna concentrator
References with USIM PAC©. Int J Miner Process. 44–45:223–235.
doi:10.1016/0301-7516(95)00038-0.
Lamberg P. 2010. Structure of a property based simulator for
SRK-Consulting. 2014. Technical report on the
minerals and metallurgical industry. Proceedings of the
Hannukainen iron-copper-gold project; Kolari district,
51st Conference on Simulation and Modelling; Oulu,
Finland.
Finland.