Better thickener control
An Australian nickel refinery upgraded its pressure acid
leach (PAL) thickener by installing a new SmartDiver ®
system from Precision Light and Air Australia and
integrating the output signals to their DCS to enhance their
thickener control strategy
Feedwell
Flocculant
Overflow Level
Clear Layer
Interface Layer
Interface Level
Mud Level
Mud Layer
Cross section diagram of the PAL
thickener with SmartDiver ® system
leading Australian nickel refinery needed
to improve thickener control to increase
underflow density, decrease flocculant
usage and increase clarity. The installed
SmartDiver ® system from PLA provides
A
measurements of clarity, hindered interface
level and mud level as well as density trends of
the compacting mud layer, allowing the
thickener control strategy to be enhanced
dramatically.
By monitoring the two independent layers of
interface and mud, reliable process data was
Overflow
Rake
Underflow
provided for control of the underflow pump to
optimise the underflow density, and control of
the flocculant dosing system. The control
strategy also implemented feed forward
control for the mud level and model predictive
control for the flocculant addition.
Flocculant usage was dramatically
decreased and underflow density increased as
did clarity. Thickener operational down time
was reduced. Flocculant usage dropped from
an average 121.78 gt-1 to 69.94 gt-1 resulting in
a cost saving of $80,000 over six months,
assuming a flocculant cost of $350 per kg. An
added benefit of
improved control was an
increase in average flow
rate, an increase in
production due to less
downtime and tighter
operational parameters.
thickener was the discharge rate of the
underflow pump system to maintain the dense
slurry compact bed at the optimum level in the
thickener.
The PLA SmartDiver ® mud level
measurements were used in a basic slow
acting PID feed-back control loop to
manipulate the mud level. With this acting as a
verification of mass balance control loop,
feedback went directly to the VSD on the
underflow pump.
There was a maximum mud level run in the
tank, based on historical data of rake torque.
This was set, and the SmartDiver ® provided the
process variable for the controller to control
the underflow pump speed. Maximum mud
level, rake torque and underflow density were
also used as inhibits.
The second process that was automated on
the thickener was the control of the hindered
layer settling zone or interface layer. Interface
changes can happen much faster, so a PID
feed-back control was used, but on a faster
acting loop. The feedback from the
SmartDiver ® interface reading trims this gram
per tonne ratio. Here the flocculant flow was
manipulated using differential interface
control: the difference in mud and interface
level (ie adjusting the flocculant addition to
Table 1: Decreased flocculant usage
Project in-depth
The first process that was
automated on the
Graph 1: Falling flocculant dosage after change
in contro strategy
Photo of the PAL
thickener with (inset)
close up of the PLA
SmartDiver ® system
P6 International Mining | MARCH 2018 Supplement
maintain a constant interface height above the
mud).
The changes made to the control strategy
reduced flocculant usage dramatically as
shown in Table 1 and Graph 1.
Flocculant usage was reduced by around
40% over six months with underflow density
being retained. Additional improvements to be
looked at include reducing flocculant
consumption even further, and controlling the
PAL thickener with increased mud levels, with a
view to increasing underflow densities.