PEPT studies of heavy particle flow within a spiral concentrator
Introduction
Many devices utilise density differences to separate valuable minerals from gangue. Among these devices, spiral concentrators play a major role, especially in the exploitation of iron ore and heavy mineral sands. As demand for the associated metals increases, recovery and grade should be improved wherever possible to maximise the output of each deposit. In the past few decades, improvements of the spiral concentrator were based on empirical modelling and iterative design. Some experimentation has been conducted to understand the characteristics of the primary and secondary flow (Holland-Batt and Holtham, 1991, Holtham, 1990). Many tests have been carried out on different spiral designs, processing different types of ore and the performance of the spiral is measured by assessing concentrate, middlings and tailings. This empirical trend is still a useful method of obtaining results, but to push design further, computer aided modelling was introduced at the turn of the millennium (Kapur and Meloy, 1998, Wang and Andrews, 1994), with the aim of increasing the use of the fundamental laws of fluid dynamics and particle motion.
With this in mind, a more fundamental assessment of the spiral requires an understanding of the particle motion inside the spiral flow, and how separation of the valuable minerals from the gangue occurs. It is also important to know which parameters of the spiral interact in this selectivity. Therefore, an understanding of particle motion in a spiral concentrator is a prerequisite to improving the understanding of the mechanisms behind the separation. This understanding will eventually lead to more accurate fundamental models of separation.
The typical diameter of the particles processed in heavy mineral spirals ranges between 75 μm and 3000 μm (Holland-Batt, 2009, Matthews et al., 1998). Probe measurements have shown that the depth of the flowing slurry on the spiral is less than 5 mm in the inner diameter region and can reach more than 15 mm on the outer edge of the trough (Holtham, 1992, Richards et al., 2000). The inner section of the trough is a laminar flow regime while a turbulent regime zone encompasses the middle and outer diameter position (Kapur and Meloy, 1999). The stream speed (calculated from volumetric sampling and depth measurement) is approximately 0.1 m/s in the inner region, while middle trough flow can reach around 0.4 m/s and more than 1.7 m/s in the outer edge of the trough (Holtham, 1992). Measurement techniques used for spiral investigations should be in accordance with these values to provide accuracy and limit flow disturbance.
Observation of the fluid flow patterns on spirals has been undertaken over the past few decades (Holland-Batt and Holtham, 1991, Holtham, 1990, Loveday and Cilliers, 1994). This flow is composed of a primary downward flow and a secondary radial flow. This secondary flow has a rotational motion in the plane perpendicular to the primary flow. It is important to show how a single particle moves inside these flows, as many mechanisms take place during separation including the Bagnold effect (Holtham, 1992), sedimentation, transport mechanisms and bed phenomena (Holland-Batt, 1995). Individual tracking of particles with different sizes and densities will allow researchers to follow motion parameters (position, speed, acceleration) of valuable and gangue particles within these flows. This information can then be used to improve the understanding of the fundamental mechanisms affecting spiral performance. Especially in the finer fraction, where some designs are less efficient and result in a substantial loss of valuable minerals that remains in the waste stream of spiral concentrating plants such as in Canadian iron ore operations.
Development of fundamental models including particle motion and interactions within the slurry is the next step for spiral improvement (Doheim et al., 2013, Matthews et al., 1998, Mishra and Tripathy, 2010). Thus, the ability to track a single particle in the bulk flow will allow validation of models of particle motion.
Positron emission particle tracking (PEPT) was developed at the University of Birmingham, UK (Hawkesworth et al., 1986, Parker et al., 1993) and has been described in detail by Leadbeater, 2009, Leadbeater et al., 2012. PEPT is based on tracking an individual particle labelled with a positron emitting radionuclide. There are different methods to label particles: direct activation (as used in this study), ion exchange (Fan et al., 2006b) and surface modification (Fan et al., 2006a). In direct activation, the selected particle is placed within a cyclotron beam (35 MeV, 3He+) thus transforming some oxygen atoms of the particle to 18F radionuclides via competing capture reactions (Fan et al., 2006a). The 18F loaded tracer particle has a half life of 110 min and decays through β+ decay.
In β+ decay, proton rich nuclei (18F in this case) emit positrons. Annihilation of each positron occurs when they come into contact with an electron, producing two gamma rays (511 keV each) which are emitted back to back (180° ± 0.5°). The detection of these gamma rays by two detectors located close to the experimental set-up allow a line of response (LoR) to be generated in 3D as shown in Fig. 1.
A single LoR generation is called an event, and multiple events are recorded over a short interval of time. These events can be separated in groups of a fixed number (N) of LoRs. Some of these LoRs are corrupted events caused by scatter, random coincidence, etc. After the removal of these corrupted events by the PEPT iterative triangulation algorithm (Parker et al., 1993), the remaining fraction (f) of the LoRs intersect close to one single point (the origin of the gamma rays). This location is assumed to be the tracer particle position at this point in time. Recording the tracer positions over time allows a number of parameters to be determined; including the trajectory, speed and residence time.
This measurement technique has been applied to track particles in mixing vessels (Chiti et al., 2011, Guida et al., 2009) and fluidised beds (Laverman et al., 2012). In mineral processing, particle behaviour in mills (Jayasundara et al., 2011), flotation cells (Waters et al., 2008, Waters et al., 2009) and hydrocyclones (Chang et al., 2011) have also been investigated using PEPT.
At The University of Birmingham, UK, and under optimal conditions, recording rates of a 1 mm tracer moving at 10 m/s are up to 1000 localisations per second with a precision of 0.5 mm (Leadbeater et al., 2012). With lower localisation rates due to sub optimal conditions, Cole et al. (2012) showed in the high performance PEPT Cape Town facility (South Africa), that it is possible to use the PEPT technique (with ion exchange activated tracers) to follow trajectories of particles of 50 μm ± 5 μm, moving at speeds of up to 2 m/s with a location error of approximately 2.6 mm. The direct activation technique used in this work allows a particle of the same composition as the bulk to be used as tracer. This is of particular importance in processes where surface chemistry plays an important role, such as the separation of minerals through flotation, and where the density of the ore minerals is the key factor in the separation, such as in spiral concentrators. The aim of the work is to validate the use of the PEPT technique and the Birmingham ADAC camera to track the motion of a particle inside a spiral. Once validation is completed, it will be possible to determine operational parameters required to investigate the primary and secondary flow as well as the minimal size of tracer that can be tracked on a spiral.
Section snippets
Spiral and bulk sample characteristics
The spiral used in this study had four turns with a 350 mm diameter trough around a 38 mm centre column and a pitch angle of 10.5° (Walkabout Spiral, Mineral Technologies). The solids used for creating the slurry was a bulk iron ore from Mont-Wright mine (ArcelorMittal Exploitation Minière Canada, Fermont, QC, Canada). The ore was composed of quartz (∼57%) and hematite (∼43%) with trace amounts of other minerals. The size fraction greater than 850 μm was removed for ease of manipulation and
Results and discussion
Combining multiple passes of a specific tracer while operating the spiral in closed circuit allowed the behaviour of the size fraction of the tracer to be represented. The camera also gives random scatter locations when the tracer is outside of the field of view recirculating in the tank and pumping system, but those are easily identified and rejected. Reference axis are x and z in the transversal direction of the spiral, while y axis is height. The x or z coordinates of the spiraling motion
Conclusions
Tracers of different sizes (53–1700 μm) and relative densities (2.7 and 4.7) have been tracked within a slurry (27% w/w solids) flowing along a spiral concentrator trough. Position with respect to time, and speed, have been used to show the general behaviour of these tracers. Speeds measured were up to 2 m/s. Analysis of two passes of the −63 +53 μm quartz tracer seem to show different layers in the flow with a behaviour that can be related to the secondary circulation. Also, slight inward and
Acknowledgments
The authors are thankful to ArcelorMittal Exploitation Minière Canada for providing the ore. The financial support of COREM and The Natural Sciences and Engineering Research Council of Canada (NSERC) through the Collaborative Research and Development Project Grant (CRDPJ 437324-12) are also duly acknowledged, as is the McGill Engineering Doctoral Award (MEDA) from the McGill Faculty of Engineering.
References (29)
- et al.
Particle flow in a hydrocyclone investigated by positron emission particle tracking
Chemical Engineering Science
(2011) - et al.
Using positron emission particle tracking (PEPT) to study the turbulent flow in a baffled vessel agitated by a Rushton turbine: improving data treatment and validation
Chemical Engineering Research and Design
(2011) - et al.
Positron emission particle tracking measurements with 50 micron tracers
Chemical Engineering Science
(2012) - et al.
Numerical simulation of particulate-flow in spiral separators: Part I. Low solids concentration (0.3% & 3% solids)
Applied Mathematical Modelling
(2013) - et al.
Enhancing 18F uptake in a single particle for positron emission particle tracking through modification of solid surface chemistry
Nuclear Instruments & Methods in Physics Research, Section A: Accelerators, Spectrometers, Detectors, and Associated Equipment
(2006) - et al.
Labelling a single particle for positron emission particle tracking using direct activation and ion-exchange techniques
Nuclear Instruments & Methods in Physics Research, Section A: Accelerators, Spectrometers, Detectors, and Associated Equipment
(2006) - et al.
Positron emission particle tracking in a mechanically agitated solid–liquid suspension of coarse particles
Chemical Engineering Research and Design
(2009) - et al.
A positron camera for industrial application
Nuclear Instruments & Methods in Physics Research, Section A: Accelerators, Spectrometers, Detectors, and Associated Equipment
(1986) The dynamics of sluice and spiral separations
Minerals Engineering
(1995)A method for the prediction of the primary flow on large diameter spiral troughs
Minerals Engineering
(2009)
Particle and fluid motion on spiral separators
Minerals Engineering
Flow visualisation of secondary currents on spiral separators
Minerals Engineering
Particle transport in gravity concentrators and the Bagnold effect
Minerals Engineering
CFD–DEM modelling of particle flow in IsaMills – Comparison between simulations and PEPT measurements
Minerals Engineering
Cited by (27)
Investigation of residence time and fluid volume in spiral concentrators
2023, Minerals EngineeringFramework components for data-centric dry laboratories in the minerals industry: A path to science-and-technology-led innovation
2022, Extractive Industries and SocietyInvestigation on the evolution of flow field stability in a spiral separator
2021, Minerals EngineeringCitation Excerpt :Holland-Batt (1995) found that the degree of particle migration in the radial direction was reduced gradually with an increase in the trough length. Boucher et al. (2014) and Atasoy and Spottiswood (1995) indicated that most of the separation occurred in the first few turns of the spiral separator. For revealing the mechanism of particle segregation thoroughly, it is necessary to investigate the particle behavior in the first few spiral turns.
Repurposing legacy metallurgical data Part I: A move toward dry laboratories and data bank
2020, Minerals Engineering