XXXI International Mineral Processing Congress 2024 Proceedings/Washington, DC/Sep 29–Oct 3 55
segregation-dispersion equation of Kennedy and Bretton
(1966), the analytical form in this paper provides a simpler
insight into the underpinning mechanism governing grav-
ity separation using the Reflux Classifier.
In turn, this paper also briefly explores the broader
technology platform which now includes coarse particle flo-
tation in the CoarseAIR, the Reflux Flotation Cell, and the
Reflux Graviton. All these devices share a common thread,
the synergy arising from a system of inclined channels at
one end of the device, and fluidization at the opposing end.
THEORETICAL
Partition Surface
The partition surface of a separator describes the probability
of a particle of a given size and density reporting to a specific
exit stream. Recent work (Galvin et al., 2020 Rodrigues
et al., 2023) has helped to establish an empirical descrip-
tion of the partition surface of the Reflux Classifier. The
problem however is that the relationship is not expressed in
terms of the actual system hydrodynamics. That limitation
is addressed in this paper.
The primary relationship describes the variation in the
separation density, D50, with the particle size, d. That is,
d
d
o w o t
twh
-
-
=
-m ^D50
^t c h m (1)
Here, ρw, is the density of the water, ρo an adjustable
reference density, do a reference particle size of 300 mm,
and m, an exponent value. Unfortunately, Equation 1
does not fix the relationship between the buoyant density,
D
w 50 t t D =-^h, and the particle diameter. Put simply,
Equation (1) only indicates the scaling,
d m +t -(2)
where m is deemed a constant. Galvin et al (2020) obtained
a very low value of m=0.28 using a chromite ore covering
a size range from 38–300 mm using an inclined channel
spacing of z=3 mm. Rodrigues et al (2023) later obtained
a very consistent set of values of m=0.26±0.02 across 12
experiments, confirming the original result of Galvin et
al (2020). Rodrigues et al also conducted an experiment
at a much lower flow velocity (Run 1), obtaining a much
poorer separation with m=0.46.
The separation density, D50, is used in the Whiten
Equation (Scott and Napier-Munn, 1992) to obtain the
partition number, P. That is,
exp
P
E
Dh
1
1
p
50
=
+
-
dln^3h
^D
n
(3)
where Ep is the Écarté Probable. Galvin et al (2020) found
the Ep to be largely independent of the particle diameter,
Figure 1. Geometrical arrangement of the REFLUX™ Classifier showing the system of
inclined channels above the vertical fluidization system
segregation-dispersion equation of Kennedy and Bretton
(1966), the analytical form in this paper provides a simpler
insight into the underpinning mechanism governing grav-
ity separation using the Reflux Classifier.
In turn, this paper also briefly explores the broader
technology platform which now includes coarse particle flo-
tation in the CoarseAIR, the Reflux Flotation Cell, and the
Reflux Graviton. All these devices share a common thread,
the synergy arising from a system of inclined channels at
one end of the device, and fluidization at the opposing end.
THEORETICAL
Partition Surface
The partition surface of a separator describes the probability
of a particle of a given size and density reporting to a specific
exit stream. Recent work (Galvin et al., 2020 Rodrigues
et al., 2023) has helped to establish an empirical descrip-
tion of the partition surface of the Reflux Classifier. The
problem however is that the relationship is not expressed in
terms of the actual system hydrodynamics. That limitation
is addressed in this paper.
The primary relationship describes the variation in the
separation density, D50, with the particle size, d. That is,
d
d
o w o t
twh
-
-
=
-m ^D50
^t c h m (1)
Here, ρw, is the density of the water, ρo an adjustable
reference density, do a reference particle size of 300 mm,
and m, an exponent value. Unfortunately, Equation 1
does not fix the relationship between the buoyant density,
D
w 50 t t D =-^h, and the particle diameter. Put simply,
Equation (1) only indicates the scaling,
d m +t -(2)
where m is deemed a constant. Galvin et al (2020) obtained
a very low value of m=0.28 using a chromite ore covering
a size range from 38–300 mm using an inclined channel
spacing of z=3 mm. Rodrigues et al (2023) later obtained
a very consistent set of values of m=0.26±0.02 across 12
experiments, confirming the original result of Galvin et
al (2020). Rodrigues et al also conducted an experiment
at a much lower flow velocity (Run 1), obtaining a much
poorer separation with m=0.46.
The separation density, D50, is used in the Whiten
Equation (Scott and Napier-Munn, 1992) to obtain the
partition number, P. That is,
exp
P
E
Dh
1
1
p
50
=
+
-
dln^3h
^D
n
(3)
where Ep is the Écarté Probable. Galvin et al (2020) found
the Ep to be largely independent of the particle diameter,
Figure 1. Geometrical arrangement of the REFLUX™ Classifier showing the system of
inclined channels above the vertical fluidization system