Content uploaded by Reza Moghaddam
Author content
All content in this area was uploaded by Reza Moghaddam on Mar 09, 2023
Content may be subject to copyright.
1 INTRODUCTION
1.1 Background
Mining operations generate mineral waste in the form of waste rock and process tailings. Hydro-
cyclones are commonly used throughout the mining industry to separate process tailings into a
coarse sand underflow fraction and a finer overflow fraction. The coarse underflow sands can be
used as an embankment fill material in tailings dam construction while the finer overflow fraction
is typically disposed of in the tailings facility impoundment area.
Cyclone sand tailings dams constructed and raised vertically by the centerline or downstream
method require free drainage and sufficient density in the embankment fill to maintain adequate
facility performance during the construction, operations, and closure lifecycle phases. Although
the evaluation of free drainage is not the objective of this paper, it is important to note that the
fines content (<0.075 mm) which affects permeability of the underflow tailings is one of the key
parameters in the drainage system design. In practice, the fines content can also be the main pa-
rameter used for quality control of the sand during construction and operation. As described by
Barrera et. al, (2011), lowering the maximum fines content specification (e.g., to 10%) may re-
quire double cycloning, which can significantly constrain sand production and require higher con-
sumption of water for dilution. Table 1 summarizes the range of fines content of the underflow
tailings produced in the base-metal mining (copper/gold) industry and used in the construction of
cycloned sand tailings dams in different regions. With some exceptions, the maximum fines con-
tent of the underflow tailings generally ranges between 12 and 18% with the coarse fraction of
the underflow tailings mainly consisting of fine sand, as shown in Figure 1.
Assessing dilative conditions of cyclone underflow tailings
Paula Mello Martins
WSP/Golder, Belo Horizonte, Minas Gerais, Brazil
Henrique Oliveira Alves
Lundin Mining, Brazil
Reza Moghaddam
Lundin Mining, Toronto, Ontario Canada
Andre Gagnon
Lundin Mining, Toronto, Ontario, Canada
ABSTRACT: Hydro-cyclones are commonly used throughout the mining industry to separate
process tailings into a coarse sand underflow fraction and a finer overflow fraction. The coarse
underflow sands that are used as an embankment fill material require sufficient density or shear
strength and free drainage to maintain adequate facility performance during the construction, op-
erations, and closure lifecycle phases. It has been recognized that achieving dilative conditions is
critical to eliminate/reduce the liquefaction risk of cyclone sand underflow embankments while
the stress conditions increase as the facility is sequentially raised. In this paper, the dilative con-
ditions of coarse cyclone sand from cooper tailings were evaluated using field compaction test
results in comparison with the critical state locus. A comparison between these results shows that
for up to approximately 700 kPa mean effective stress, the underflow material will exhibit dilative
behavior if compacted to more than 98% of Standard Maximum Dry Density.
Table 1. Fines content of cyclone sand tailings dams.
Study
Mine/Region
Mine Type
Fines Content (%)
Rang
e
maximum or
tested
Klohn (1984)
Brenda
Mines
Copper and Molyb-
denum
3-12
12
Barrera et al. (2011)
Las Tórtolas
Porphyric Copper
12-15
16
Torito
15-18
20
Ovejería
12-14
15
Quillayes
13-17
18
Mauro
13-16
18
Kujawa (2011)
NA
Gold-Copper porphyry
12
NA
Obermeyer & Alexieva
(2011)
Cerro Verde,
Peru
Copper Mine
15
15
Busslinger et al (2013)
Canada
Gold-Copper porphyry
17
NA
Boshoff and Reid (2015)
Australia
Gold
24
24
Valenzuela (2016)
Chile
Copper
12-18
18
Figure 1. Particle size distribution of cyclone underflow tailings from different regions.
1.2 Objective
It has been recognized that achieving dilative conditions is critical to eliminate/reduce the lique-
faction risk of cyclone sand tailings dams while the stress conditions increase as the facility is
sequentially raised.
In this paper, the dilative conditions of coarse cyclone sand underflow tailings produced in a
copper mine after placement and compaction are evaluated using field and laboratory testing re-
sults.
2 UNDERFLOW TAILINGS PRODUCTION AND OPERATION
Hydro-cyclones or cyclones typically have a feed inlet and two outlets, one lower named Apex or
spigot and the other upper one known as the Vortex Finder, as schematically shown in Figure 2.
The primary principle around hydro-cyclone operation is to use centrifugal and gravity forces to
separate materials with different dimensions and concentrate solids. Their conical shapes together
with feed pressure facilitate the concentration of larger diameter grains on the periphery of the
cyclone. The larger diameter grains are driven by gravitational force to the bottom of the cyclone
and then discharged through the Apex (underflow). With the positioning of these grains on the
periphery and at the bottom of the cyclone, the smaller diameter grains (overflow) and most of
the water tend to concentrate at the top and center of the cyclone and then discharged through the
Vortex finder.
For a progressive dam raise, cyclones are typically installed at equidistant points on the crest,
as shown in Figure 2. At this site, the underflow tailings are produced daily through 20-inch di-
ameter cyclones connected to a pipeline with 75% of solids content and an underflow split of
45%. The underflow tailings are placed along the downstream slope in 30 to 40 cm thick layer
using dozers. Compaction takes place using a roller adopting six to eight passes. In situ dry density
tests are performed to ensure compaction specifications are met (Section 3.3).
Figure 2. Schematic of hydro-cyclone system and field set up.
The percent solids, particle size distribution, and mass fraction in the cyclone tailings underflow
and overflow are driven by several factors. To ensure proper underflow tailings production in
accordance with the technical specifications, the operator must consider the following controls:
− Feed Pressure: Cyclones have manometers installed in their chamber that allow to evaluate
the pressure of the feed. The feed pressure can be varied and can affect the quality of pro-
duction as described below.
− Visual Assessment of Underflow Output: In the evaluation of cyclone function, the main as-
pect to be observed is at the underflow exit from the Apex. When underflow has the appro-
priate solids percentage, the output angle is similar to the cyclone angle. As the feed rate,
pressure and/or density of the tailings slurry entering the cyclone decreases, the forces act-
ing down on the slurry are small relative to the centrifugal forces within the cyclone, and the
dispersion angle increases. This phenomenon is called spray discharge, and results in an
overflow with a finer particle size distribution. As the feed rate and pressure increase, the
dispersion angle decreases. A low dispersion angle indicates a low cyclone efficiency be-
cause the separation forces (centrifuge) are relatively low. The extreme case occurs when
the cyclone is overloaded excessively, and the slurry descends vertically from the Apex (de-
load cord). The cord discharge indicates that the Vortex no longer exists; the central column
of air does not exist. In this case, the cyclone is no longer properly classifying the feed.
− Field Measurements of Underflow Solids Percentage: The percentage of underflow solids
can be verified using the dynamometer and calibrated sampling container.
Pump speed, the amount of open drains, and opening of cyclone feed valves can be adjusted to
improve the cycloning process and to obtain a better quality and higher quantity material.
3 UNDERFLOW TAILINGS CHARACTERSTICS
3.1 Index properties
The total tailings are classified as non-plastic silty sand materials with a specific gravity of 2.81.
3.2 Average solid and fines content
The cyclone sand tailings consist of fine to medium sand with the fines content (<0.075 mm)
averaging between 12 and 20%. Figure 3 presents the range of particle size distribution for the
underflow tailings produced since 2018 at the site.
Figure 3. Particle size distribution of cyclone underflow tailings produced since 2018.
3.3 Field Compaction
For dam construction, the cyclone sand tailings are placed in 30 to 40 cm lifts and compacted to
a degree higher than 95% of Standard Maximum Dry Density (SMDD). The field tests performed
confirmed that the average in situ dry density values of approximately 98% of SMDD have been
achieved over the years with a mean dry density of 1.67 g/cm³. The mean field void ratio was
estimated to be 0.69 after compaction at the surface and the average field void ratio for the areas
compacted below 95% of SMDD was estimated to be 0.74 at the surface.
4 LABORATORY TESTING
To evaluate the dilative condition for future dam raises, the critical state locus (CSL) of the cy-
clone sand tailings was established through a series of triaxial tests and compared to the isotropic
consolidation line (ICL) obtained during the consolidation stage of the triaxial testing.
4.1 Materials Tested
Two underflow tailings samples (Figure 3) that represent the average curve (Sample 1) and the
upper boundary (Sample 2) of the underflow tailings particle size distribution compacted in the
field were selected for two series of triaxial testing. In this paper, the results of triaxial testing on
Sample 1 as an average value are presented.
4.2 Sample Preparation
The specimens were prepared by moist tamping using similar equipment and procedures to those
described by Jefferies and Been (2016). The specimens were prepared to initial water contents of
about 5% and compacted in a split mould in six equal layers of equivalent height (refer to Sche-
matic 1). The specimens were prepared to different initial relative densities, being either very
loose (for CSL determination) or very dense (for stress-dilatancy interpretation). For specimens
prepared in a very loose initial state, under-compaction of the lower layers (with U=5%) was used
to improve specimen uniformity according to the method proposed by Ladd (1978). No under-
compaction was used for the very dense specimens. Lubricated end platens or “free-ends” were
used to minimize the effects of end friction during the triaxial tests.
4.3 Consolidation Stage
The triaxial specimens were flushed with de-aired water and saturated under a backpressure until
saturation was confirmed by measuring a B-value of 0.98 or greater. The specimens were then
isotropically consolidated to the target confining pressures while measuring volume change and
axial displacement. The target confining pressure ranged between 100 and 800 kPa.
4.4 Triaxial Loading
A total of 10 triaxial compression tests were completed on Sample 1. Seven isotropically consol-
idated drained (CID) triaxial tests and three isotropically consolidated undrained (CIU) triaxial
tests are shown in Figure 4.
Figure 4: Triaxial testing results on Sample 1. (a) stress path, (b) deviator stress vs axial strain, and (c) stress
path with p’ up to 400 kPa.
5 RESULTS
Figure 5 presents the results of triaxial testing in a e-p’ plot. A curved CSL (dashed-red line with
a=0.947, b=1.14 and c=0.402) was determined for this material. The isotropic consolidation
curves for dense and loose samples are also presented in this figure. It is apparent that the consol-
idation curve for loose samples (CID 103 and CID 104) are parallel to the CSL, which can be
interpreted as the limiting compression curve (LCC) as described by Pestana and Whittle (1995).
The isotropic compression curve (dashed-magenta curve) obtained from the consolidation stage
of the triaxial test on a dense sample (CID 107), which shows an almost flat slope is also presented
in this figure. It is known that there are infinite number of normally consolidation locus for sand-
like soils (Jefferies and Been, 2016 and Robertson, 2022). A hypothetical comparison locus
(black-solid curve) that represents the compression behavior of the underflow tailings compacted
to 98% of SMDD (see Section 3.3) was determined parallel to this curve considering the initial
void ratio of these two curves are close to each other.
A comparison between the compression locus obtained for the field underflow tailings (black-
solid curve) and the CSL (dashed-red curve) shows that up to approximately 700 kPa mean effec-
tive stress, the underflow material will exhibit dilative behaviour if compacted to more than 98%
of SMDD. Dilative behaviour can, also, be expected if the underflow material is compacted to
more than 95%, but the main effective stresses must be lower than 400 kPa.
A judgement criterion of ψ < −0.05 (ψ = state parameter = e0 - eCSL) is typically considered
in practice to identify the contractive/dilatative boundary at large scale based on Shuttle and Cun-
ning (2008). The state parameter criterion of ψ < −0.05 for the underflow tailings tested in this
study is also presented in Figure 5. A comparison between the compression locus obtained for the
field underflow tailings (black-solid curve) and the state parameter criterion indicates that the
maximum mean effective stress must reduce to 400 kPa for the underflow tailings to exhibit dila-
tive behavior. The stress level decreases further (~200 kPa) if the minimum dry density corre-
sponding to 95% of SMDD will be considered, as presented in Figure 5.
Figure 5: The CSL of the underflow tailings tested along with the compression locus of the loose and dense
samples.
6 DISCUSSION
The state parameter criterion proposed by Shuttle and Cunning (2008) is an approximate criterion
emerging from laboratory testing and three NorSand simulations of undrained stress–strain be-
havior using typical clean quartz sand properties and may not be applicable for the material tested
in this study. Shuttle and Cunning (2008) showed that something like ψ < –0.05 is necessary to
avoid a plateau in the stress–strain behavior and ensure that a flow slide will not develop through
porewater migration into a localizing shear band or zone. To evaluate this judgment criterion, it
is recommended that the NorSand model be developed for this material based on the available
triaxial testing results.
In addition, Jefferies and Been (2016) stated that in the CIU tests, the tests start from isotropic
initial conditions so that the entire stress path is undrained. For in situ field conditions, it would
be very unusual to encounter K0 = 1 with loose soils with the general expectation being in the
range 0.5 < K0 < 0.8. This range of in situ K0 translates to perhaps only two-thirds of a potential
loading path being undrained (Jefferies and Been, 2016), and thus less excess pore water pressure
at the onset of instability: looser states than the laboratory tests will generally be stable. The in
situ K0 should be estimated for this material and be considered with other factors such as changes
in phreatic levels for future dam raises.
7 CONCLUSIONS
Cyclone sand tailings dams constructed and raised vertically by the centerline method require
sufficient density or shear strength and free drainage in the embankment fill to maintain adequate
facility performance during the construction, operations, and closure lifecycle phases. It has been
recognized that achieving dilative conditions is critical to eliminate/reduce the liquefaction risk
of cyclone sand underflow dams while the stress conditions increase as the facility is sequentially
raised. In this paper, the dilative conditions of coarse cyclone sand underflow after placement
and compaction were evaluated using field and laboratory testing results within the critical state
soil mechanics framework.
The cyclone sand tailings consist of fine to medium sand with the fines content (<0.075 mm)
averaging between 12 and 20%. For dam construction, the cyclone sand tailings are placed in 30
to 40 cm lifts and compacted to a degree higher than 95% of SMDD. The field experiments
confirmed that an average 98% of SMDD has been achieved over the years. To evaluate the dila-
tive condition for future dam raises, the CSL of the cyclone sand tailings was established through
a series of triaxial tests and compared to the isotropic compression locus of the underflow tailings.
A comparison between these results shows that up to approximately 700 kPa mean effective stress,
the underflow material will exhibit dilative behaviour if compacted to more than 98% of SMDD.
Considering a judgement criterion of ψ < −0.05, which is typically used in practice to identify the
contractive/dilatant boundary at large scale, the maximum mean effective stress will reduce to
400 kPa.
8 REFERENCES
Barrera, S., Valenzuela, L., and Campaña, J. 2011. Sand Tailings Dams: Design, Construction and Opera-
tion, Proceedings Tailings and Mine Waste 2011 Vancouver, BC, November 6 to 9, 2011.
Boshoff, J. and Reid, D. 2015. Assessing the flow liquefaction susceptibility of cyclone underflow material,
Paste 2015 – RJ Jewell and AB Fourie (eds) © 2015 Australian Centre for Geomechanics, Perth, ISBN
978-0-9924810-1-8.
Busslinger, M., Plewes, H. And Parkinson G. 2013. Testing of Cyclone Sand Tailings at High Stresses,
GEO- Montreal, September 29 to Thursday October 3, 2013.
Jefferies, M.G., and Been, K. 2016. Soil liquefaction – a critical state approach. Taylor & Francis, London,
ISBN 0-419-16170-8 478 pages.
Klohn, EJ. 1984. The Brenda Mines' Cycloned- Sand Tailings Dam, International Conference on Case
Histories in Geotechnical Engineering. 46.
Kujawa, C. 2011. Cycloning of Tailing for the Production of Sand as TSF Construction Material, Proceed-
ings Tailings and Mine Waste 2011 Vancouver, BC, November 6 to 9, 2011.
Obermeyer, J and Alexieva T. 2011. Design, Construction and Operation of a Large Centerline Tailing
Storage Facility with High Rate of Rise, Proceedings Tailings and Mine Waste 2011 Vancouver, BC,
November 6 to 9, 2011.
Pestana, J.M. and Whittle, A.J. (1995) Compression model for cohesionless soils. Geotechnique, 45 (4),
611–633.
Shuttle, D.A. and Cunning, J. (2008) Reply to discussion: Liquefaction potential of silts from CPTu. Cana-
dian Geotechnical Journal, 45(1), 142–145.
Valenzuela, L. 2016. Design, construction, operation and the effect of fines content and permeability on the
seismic performance of tailings sand dams in Chile, Obras y Proyectos 19, 6-22.
