4
The most important of these reagents is the iron
depressant, which enables the processing of iron-rich mar-
casite ores. This has since 2022 allowed for the processing
of over 200,000 tons of marcasite high lead and high Cu
ores through Buick Mill which otherwise would not have
been as economical to process.
The other four reagents allow for finetuning the perfor-
mance of the float cells as needed to ensure good recovery
and the continuous production of saleable concentrates.
By placing these under the control of the expert system, it
can rapidly respond to variations in feed grade, concentrate
quality, or tail assays. Generally, these reagents are added
in short bursts to quickly correct for short term variations,
while the bulk reagents are used to control the process
overall.
All these reagents can be used to reduce the usage of
bulk reagents though. Of particular interest here is the iron
depressant again, as it can significantly replace the require-
ment for cyanide in Buick Mill’s normal operations espe-
cially when processing marcasite ores. The collectors also
help supplement the need for xanthate as well.
The iron depressant in the hybrid reagent system has
been shown at Buick Mill to be effective in allowing the
processing of marcasite ores, but marcasite ores are only
sporadically provided to the mill. The other, even more
important, side of the breakthrough at Buick Mill is the
development and implementation of a groundbreaking
expert control system to operate the entirety of the plant,
including the hybrid reagent system. With this expert con-
trol system, the iron depressant and the other reagents in
the hybrid system can be dosed automatically based on
automated detection of marcasite ores or other problematic
situations in the mill.
DIGITAL ONE ENTERPRISE EXPERT
CONTROL SYSTEM EVOLVED FROM
VISIOFROTH™
Controlling flotation in full is generally accepted to be a
very challenging prospect, as flotation is a widely believed
stochastic process, which in return most likely hindered
development of such a control system. The process is par-
ticularly sensitive to changes in mill feed ore grades espe-
cially for multi- metallic ores at Buick mine.
The fundamental outline of the expert control system
is guided by the key requirement that it be robust enough
to remain online and operational continuously. As a quick
overview, the layers of the system are as follows:
1. Field instruments and controllers: The physical
devices responsible for providing input and output
from the control system to the plant.
2. Distributed Control System (DCS): Aggregates
data from the field instruments and adjusts the
controllers to maintain operating set point targets.
3. VisioFroth ™ System: Receives data from DCS and
calculates new operating set point targets for it to
control the froth velocity in each flotation circuit.
Provides camera bubble data and fundamental
software framework for the expert control system.
4. Our Expert Control System: Built in the software
OCS-4D that powers VisioFroth ™ and utilizing
aggregated data from the VisioFroth ™ and DCS,
adjusts set points for target froth velocities, hybrid
reagent dosing, air flowrates and MIBC, and some
bulk reagent dosing towards reaching target lead,
zinc, and copper rougher and final concentrate
grades.
5. PI System: Receives data from all the prior lay-
ers, providing a central data hub and enabling
advanced data analytics. Also has the capacity to
provide soft sensors to the expert control system
as needed.
In essence, the DCS layer is ultimately responsible for
controlling the plant to the set points requested by opera-
tors or by the expert control. The set points provided by
the expert system or the VisioFroth ™ system are treated as
suggestions by the DCS layer and, if necessary, can be over-
ridden by human operators at any time. However, avoiding
that be coming necessary is a core objective in the design of
the expert system.
The Digital One expert control system as described
has been developed for both Rockwell DCS and Emerson
Delta V systems at separate mill sites and can be an indis-
pensable enterprise expert control and operating system for
both the Doe Run mills and wider industry.
To ensure that human operators are supported by the
expert system rather than fighting with it, several key fun-
damentals are built into the design of the system:
1. Fundamental thermodynamic and kinetic founda-
tions: The decisions which the expert control sys-
tem is allowed to make are based solely on key flo-
tation principles. These decisions can be supported
by data, by years of operator experience, and by
theoretical reasoning.
2. Avoiding encoding arbitrary rules of thumb:
When asking an operator when they might enable
a reagent, they may give specific suggestions such
as “when the Pb grade in that stream is over 1.0%.”
However, an operator’s experience and apprecia-
tion for context exceeds what can ever be encoded
The most important of these reagents is the iron
depressant, which enables the processing of iron-rich mar-
casite ores. This has since 2022 allowed for the processing
of over 200,000 tons of marcasite high lead and high Cu
ores through Buick Mill which otherwise would not have
been as economical to process.
The other four reagents allow for finetuning the perfor-
mance of the float cells as needed to ensure good recovery
and the continuous production of saleable concentrates.
By placing these under the control of the expert system, it
can rapidly respond to variations in feed grade, concentrate
quality, or tail assays. Generally, these reagents are added
in short bursts to quickly correct for short term variations,
while the bulk reagents are used to control the process
overall.
All these reagents can be used to reduce the usage of
bulk reagents though. Of particular interest here is the iron
depressant again, as it can significantly replace the require-
ment for cyanide in Buick Mill’s normal operations espe-
cially when processing marcasite ores. The collectors also
help supplement the need for xanthate as well.
The iron depressant in the hybrid reagent system has
been shown at Buick Mill to be effective in allowing the
processing of marcasite ores, but marcasite ores are only
sporadically provided to the mill. The other, even more
important, side of the breakthrough at Buick Mill is the
development and implementation of a groundbreaking
expert control system to operate the entirety of the plant,
including the hybrid reagent system. With this expert con-
trol system, the iron depressant and the other reagents in
the hybrid system can be dosed automatically based on
automated detection of marcasite ores or other problematic
situations in the mill.
DIGITAL ONE ENTERPRISE EXPERT
CONTROL SYSTEM EVOLVED FROM
VISIOFROTH™
Controlling flotation in full is generally accepted to be a
very challenging prospect, as flotation is a widely believed
stochastic process, which in return most likely hindered
development of such a control system. The process is par-
ticularly sensitive to changes in mill feed ore grades espe-
cially for multi- metallic ores at Buick mine.
The fundamental outline of the expert control system
is guided by the key requirement that it be robust enough
to remain online and operational continuously. As a quick
overview, the layers of the system are as follows:
1. Field instruments and controllers: The physical
devices responsible for providing input and output
from the control system to the plant.
2. Distributed Control System (DCS): Aggregates
data from the field instruments and adjusts the
controllers to maintain operating set point targets.
3. VisioFroth ™ System: Receives data from DCS and
calculates new operating set point targets for it to
control the froth velocity in each flotation circuit.
Provides camera bubble data and fundamental
software framework for the expert control system.
4. Our Expert Control System: Built in the software
OCS-4D that powers VisioFroth ™ and utilizing
aggregated data from the VisioFroth ™ and DCS,
adjusts set points for target froth velocities, hybrid
reagent dosing, air flowrates and MIBC, and some
bulk reagent dosing towards reaching target lead,
zinc, and copper rougher and final concentrate
grades.
5. PI System: Receives data from all the prior lay-
ers, providing a central data hub and enabling
advanced data analytics. Also has the capacity to
provide soft sensors to the expert control system
as needed.
In essence, the DCS layer is ultimately responsible for
controlling the plant to the set points requested by opera-
tors or by the expert control. The set points provided by
the expert system or the VisioFroth ™ system are treated as
suggestions by the DCS layer and, if necessary, can be over-
ridden by human operators at any time. However, avoiding
that be coming necessary is a core objective in the design of
the expert system.
The Digital One expert control system as described
has been developed for both Rockwell DCS and Emerson
Delta V systems at separate mill sites and can be an indis-
pensable enterprise expert control and operating system for
both the Doe Run mills and wider industry.
To ensure that human operators are supported by the
expert system rather than fighting with it, several key fun-
damentals are built into the design of the system:
1. Fundamental thermodynamic and kinetic founda-
tions: The decisions which the expert control sys-
tem is allowed to make are based solely on key flo-
tation principles. These decisions can be supported
by data, by years of operator experience, and by
theoretical reasoning.
2. Avoiding encoding arbitrary rules of thumb:
When asking an operator when they might enable
a reagent, they may give specific suggestions such
as “when the Pb grade in that stream is over 1.0%.”
However, an operator’s experience and apprecia-
tion for context exceeds what can ever be encoded