Environmental Engineering Reference
In-Depth Information
At 362 min, an increment of ore hardness to the rod mill is simulated, causing a
reduction to product size. This is regulated through a reduction of fresh ore feedrate
and an increase to water addition to the sump box, reducing the percentage of solids
to the battery of hydrocyclones, satisfying the volumetric balance around the sump
box. Therefore, the speed pump remains at its upper constraints, achieving a more
stable grinding circuit, one without oscillations.
As the control strategy maximizes the fresh ore feedrate at all times, it is neces-
sary to constrain the maximum value of this variable, as it can lead to overloads in
the rod mill.
5.6.2 Supervisory Control for SAG Grinding Circuits
The control of SAG grinding circuits has historically been done with expert sys-
tems [28], where a set of controllers are activated by operating conditions. Early
approaches considered a combination of logic based on signal selectors and feed-
back controllers. Reference [29] describes an application of this approach at the
Paddy's Flat gold plant of Dominion Mining Ltd Western Australia. With the ad-
vent of expert system shells, the logic was enhanced and coded as rules, which were
easier to adapt to the different requirements of each installation. Some examples
can be found in [30], where an application in Boliden Minerals Aitik concentrator
is described. Additional examples can be found in two gold ore grinding circuits in
South Africa, [31] and [32].
Recently, Thorton [33] has summarized the main challenges faced by control
strategy design for SAG mills. He has proposed the use of both feedback and feed-
forward controllers to account for the effects of major plant dynamics and distur-
bances. Their implementation has led to a successful application (95% operator uti-
lization) of conventional control techniques at two SAG milling circuits at North-
parkes Copper Mines.
In recent years, model-based strategies have gained attention due to their suc-
cesses in many process-based industries, such as the petrochemical industry. Early
work on the application of model-based controllers to SAG mills can be found in
[34] and [35]. Recent industrial applications of this approach have been carried by
Ideas Simulation and Control [36].
The following sub-section is based on work carried out by Garrido [37]. It
demonstrates the versatility of a simulator for studying process dynamic and ad-
vanced control strategies at SAG mills.
5.6.2.1 Open Loop Responses
As a first step, the simulator was configured for a simple SAG mill circuit, as de-
picted in Figure 5.14.
Search WWH ::
Custom Search