XXXI International Mineral Processing Congress 2024 Proceedings/Washington, DC/Sep 29–Oct 3 485
7. Use a Hertzian contact model to update the
mechanical forces and moments.
Rasera et al. (2022a) and Cruise et al. (2022) identify two
crucial design criteria for effective tribocharging: first, the
maximisation of particle-wall contacts is essential for pre-
dictable charge transfer since interactions between particles
are inherently random second, it’s important to limit the
number of particles in a given charging space, as a higher
particle count suppresses both bulk saturation charge and
charge transfer. For the purposes of design optimisation, we
simplify the algorithm to track only the net particle-particle
and particle-wall contact areas and use these as proxies for
charge transfer (Rasera et al., 2022b). This simplification
reduces greatly the computational cost by removing the cal-
culation of the local electrostatic field at each point of con-
tact at every timestep. With respect to the second criteria,
we maintain a constant particle inlet velocity of 0.08 m/s.
The residence time of each particle within the domain is
then tracked and compared lower residence times imply
fewer particles within the domain at any given time.
TRIBOCHARGER DESIGN
OPTIMISATION
Baseline Design
Two basic designs were considered: angled slats within a
tube (at both 30° and 45° with respect to the vertical), as
well as a twisted baffle design employed with success by
Trigwell et al. (2009, 2012) and Quinn et al. (2013).
Different slat spacings were tested to ensure that par-
ticles would not build up within the tribocharger domain.
This consideration was necessary to ensure that the propor-
tion of particle-wall contacts would remain high relative to
the particle-particle contacts, whilst also minimising the
volume fraction of particles within the domain. Figure 1
presents a comparison of the 30° slat design with a spacing
of 1.0 cm and 1.25 cm. This difference is explained by the
still images from the simulations found in Figure 2: the nar-
rower spacing reduced greatly the throughput and partially
blocked the particle flow, resulting in 22-fold increase in
particle-particle contact area.
The ratio of particle-particle to particle-wall contacts
was made for the three preliminary designs the simulation
outputs are found in Figure 3. There were no statistically
significant differences between the net particle-particle
contact areas for each design. However, the twisted baffle
generated significantly greater net particle-wall contact area
than the slat designs. Further, it was the only design where
the net particle-wall contact exceeded the net particle-par-
ticle contact area.
A subsequent study was performed to determine how
the segmented design of Trigwell et al. (2009, 2012} and
Quinn et al. (2013) would perform against a continuous
baffle. The two 90° segments were compared to a single 180°
design. The net particle-particle and particle-wall contact
areas output by the simulations are found in Figure 4. The
segmented design produced greater overall particle-particle
and particle-wall contact areas, however the relative ratio
between them was low (PW/PP =1.03). In comparison,
Figure 1. A comparison of the net particle-particle and particle-wall contact areas for the
30° slat design resulting from 1.0 and 1.25 cm slat spacings
7. Use a Hertzian contact model to update the
mechanical forces and moments.
Rasera et al. (2022a) and Cruise et al. (2022) identify two
crucial design criteria for effective tribocharging: first, the
maximisation of particle-wall contacts is essential for pre-
dictable charge transfer since interactions between particles
are inherently random second, it’s important to limit the
number of particles in a given charging space, as a higher
particle count suppresses both bulk saturation charge and
charge transfer. For the purposes of design optimisation, we
simplify the algorithm to track only the net particle-particle
and particle-wall contact areas and use these as proxies for
charge transfer (Rasera et al., 2022b). This simplification
reduces greatly the computational cost by removing the cal-
culation of the local electrostatic field at each point of con-
tact at every timestep. With respect to the second criteria,
we maintain a constant particle inlet velocity of 0.08 m/s.
The residence time of each particle within the domain is
then tracked and compared lower residence times imply
fewer particles within the domain at any given time.
TRIBOCHARGER DESIGN
OPTIMISATION
Baseline Design
Two basic designs were considered: angled slats within a
tube (at both 30° and 45° with respect to the vertical), as
well as a twisted baffle design employed with success by
Trigwell et al. (2009, 2012) and Quinn et al. (2013).
Different slat spacings were tested to ensure that par-
ticles would not build up within the tribocharger domain.
This consideration was necessary to ensure that the propor-
tion of particle-wall contacts would remain high relative to
the particle-particle contacts, whilst also minimising the
volume fraction of particles within the domain. Figure 1
presents a comparison of the 30° slat design with a spacing
of 1.0 cm and 1.25 cm. This difference is explained by the
still images from the simulations found in Figure 2: the nar-
rower spacing reduced greatly the throughput and partially
blocked the particle flow, resulting in 22-fold increase in
particle-particle contact area.
The ratio of particle-particle to particle-wall contacts
was made for the three preliminary designs the simulation
outputs are found in Figure 3. There were no statistically
significant differences between the net particle-particle
contact areas for each design. However, the twisted baffle
generated significantly greater net particle-wall contact area
than the slat designs. Further, it was the only design where
the net particle-wall contact exceeded the net particle-par-
ticle contact area.
A subsequent study was performed to determine how
the segmented design of Trigwell et al. (2009, 2012} and
Quinn et al. (2013) would perform against a continuous
baffle. The two 90° segments were compared to a single 180°
design. The net particle-particle and particle-wall contact
areas output by the simulations are found in Figure 4. The
segmented design produced greater overall particle-particle
and particle-wall contact areas, however the relative ratio
between them was low (PW/PP =1.03). In comparison,
Figure 1. A comparison of the net particle-particle and particle-wall contact areas for the
30° slat design resulting from 1.0 and 1.25 cm slat spacings