1094 XXXI International Mineral Processing Congress 2024 Proceedings/Washington, DC/Sep 29–Oct 3
research and development. Additionally, many traditional
Original Equipment Manufacturers (OEMs) treat digital
development like product development while technology
companies or start-ups lack product and process exper-
tise, generating a market gap. Bridging this gap requires
products tailored to mining industry needs. In spite of this
success, which goes beyond the product with easy imple-
mentation—excellent after-sales support, readily available
resources for installation, and rapid iterations are crucial.
Assumption of Data Availability and Instrumentation
Already in Place
During our initial days, we have commonly heard multiple
experts and companies claiming customers (mining compa-
nies) do not need more sensors added to their equipment.
A recent article by Gleeson (2022a) suggests that some
slurry pump OEMs have prioritised not accessing new data
streams but also making the most of the information avail-
able upstream and down stream of the pump. Information
such as pump flow, head, RPM is said to be already avail-
able for mission-critical pumps at key customer sites. These
available data can help create key performance indicators
and complex calculations to optimise not only the pump
operation but also the full grinding circuit. Additionally, the
article highlights that not much can be added to the pump
from an optimising of the circuit point-of-view. However,
the major problem the industry faces is the reluctance to
share the data with OEMs to carry out improvements.
We have looked at this avenue in detail before, and
our experience reveals a stark contrast. This view is echoed
by industry experts like Peter Munro, who eloquently cap-
tured this statement in his 2016 AUSIMM Mill Operators
paper (Munro, 2016):
“There is a credibility chasm between the lofty concept
of ‘big data’ and the reality I see in the current general
unsatisfactory state of plant measurement instrumen-
tation and process control systems.”
It is appreciated that some key customers and many mod-
ern plants have decent data streams for critical equipment,
but it is not echoed for all ancillary equipment like pumps.
Some critical pumps have reasonable data—many lack the
completeness, accuracy, consistency, and quality needed
for meaningful analysis. This can stem from various issues
like missing or poorly calibrated instruments, sub-optimal
sensor placement (e.g., pressure sensors on top of cyclones
reading terminal pressure not the discharge reading), mal-
functioning equipment, or even a reliance on visual read-
ings without digital recording. Moreover, numerous pumps
in a plant lack adequate instrumentation altogether due to
cost constraints.
The data disconnect creates multiple levels of prob-
lems. Starting with faulty foundations, if the underlying
data is flawed, any models and calculations built upon it
will inevitably be skewed, leading to misguided optimi-
zation efforts. Secondly, limited learning and insufficient
instrumentation on majority of pump applications restrict
the availability of data for diverse scenarios, hindering our
ability to learn and adapt for applications outside grind-
ing like thickener underflow, tailings, froth, non-Newto-
nian fluids, and more. In our recent paper on slurry pump
safety (Varghese et al., 2023), we have noted many smaller
pumps and pumps in duty-standby arrangement (where
availability is less of concern) have less instrumentation
that subsequently has resulted in major explosions, damag-
ing property, creating injury and even death to personnel.
Finally, in blind benchmarking, without well-instrumented
pumps, crucial metrics like energy utilization and sustain-
ability remains invisible, hampering industry-wide bench-
marking and progress. While in many cases, reasonable
historical data and expert insights (without digitalization)
can offer valuable guidance for improved pump operation,
they lack the guarantees and scalability sought in the digital
revolution. Companies striving for digital transformation
cannot afford to build on shaky foundations.
Therefore, the need for lower-cost sensors is critical.
It is interesting to note that Barnewold and Lottermoser
found a close correlation between implementation of digi-
tal technologies, the run-of-mine production rate, and the
parent companies’ revenue. This is an indication that the
larger and high-volume producers have been leading the
implementation. Barnewold and Lottermoser contend
that, due to IT hardware cost falling sharply in recent years,
digital systems should also be interesting and affordable to
the smaller operations, which is not reflected in the analy-
sis. It is their belief that part of this can be explained by the
affordability of sensors, limiting their ubiquitous deploy-
ment to help monitor equipment, processes and systems
(Barnewold &Lottermoser, 2020). To put it in perspective,
to monitor a pump accurately in a system pressure, flow,
power and density measurements are needed. The cost of
sensors in a slurry system can easily be double the cost of the
pump and drive system alone. We need affordable, accurate
and easily deployable sensors. Only this way can we follow
a trajectory similar to the rapid rise of electronics in the
automotive industry since 1970s. According to Figure 2A,
it is projected that the cost of electronics, as a percent-
age of total car price, will reach 50% by 2030 from 1%
in 1950 (Mathas, 2017 Placek, 2023). Lower-cost sensors
research and development. Additionally, many traditional
Original Equipment Manufacturers (OEMs) treat digital
development like product development while technology
companies or start-ups lack product and process exper-
tise, generating a market gap. Bridging this gap requires
products tailored to mining industry needs. In spite of this
success, which goes beyond the product with easy imple-
mentation—excellent after-sales support, readily available
resources for installation, and rapid iterations are crucial.
Assumption of Data Availability and Instrumentation
Already in Place
During our initial days, we have commonly heard multiple
experts and companies claiming customers (mining compa-
nies) do not need more sensors added to their equipment.
A recent article by Gleeson (2022a) suggests that some
slurry pump OEMs have prioritised not accessing new data
streams but also making the most of the information avail-
able upstream and down stream of the pump. Information
such as pump flow, head, RPM is said to be already avail-
able for mission-critical pumps at key customer sites. These
available data can help create key performance indicators
and complex calculations to optimise not only the pump
operation but also the full grinding circuit. Additionally, the
article highlights that not much can be added to the pump
from an optimising of the circuit point-of-view. However,
the major problem the industry faces is the reluctance to
share the data with OEMs to carry out improvements.
We have looked at this avenue in detail before, and
our experience reveals a stark contrast. This view is echoed
by industry experts like Peter Munro, who eloquently cap-
tured this statement in his 2016 AUSIMM Mill Operators
paper (Munro, 2016):
“There is a credibility chasm between the lofty concept
of ‘big data’ and the reality I see in the current general
unsatisfactory state of plant measurement instrumen-
tation and process control systems.”
It is appreciated that some key customers and many mod-
ern plants have decent data streams for critical equipment,
but it is not echoed for all ancillary equipment like pumps.
Some critical pumps have reasonable data—many lack the
completeness, accuracy, consistency, and quality needed
for meaningful analysis. This can stem from various issues
like missing or poorly calibrated instruments, sub-optimal
sensor placement (e.g., pressure sensors on top of cyclones
reading terminal pressure not the discharge reading), mal-
functioning equipment, or even a reliance on visual read-
ings without digital recording. Moreover, numerous pumps
in a plant lack adequate instrumentation altogether due to
cost constraints.
The data disconnect creates multiple levels of prob-
lems. Starting with faulty foundations, if the underlying
data is flawed, any models and calculations built upon it
will inevitably be skewed, leading to misguided optimi-
zation efforts. Secondly, limited learning and insufficient
instrumentation on majority of pump applications restrict
the availability of data for diverse scenarios, hindering our
ability to learn and adapt for applications outside grind-
ing like thickener underflow, tailings, froth, non-Newto-
nian fluids, and more. In our recent paper on slurry pump
safety (Varghese et al., 2023), we have noted many smaller
pumps and pumps in duty-standby arrangement (where
availability is less of concern) have less instrumentation
that subsequently has resulted in major explosions, damag-
ing property, creating injury and even death to personnel.
Finally, in blind benchmarking, without well-instrumented
pumps, crucial metrics like energy utilization and sustain-
ability remains invisible, hampering industry-wide bench-
marking and progress. While in many cases, reasonable
historical data and expert insights (without digitalization)
can offer valuable guidance for improved pump operation,
they lack the guarantees and scalability sought in the digital
revolution. Companies striving for digital transformation
cannot afford to build on shaky foundations.
Therefore, the need for lower-cost sensors is critical.
It is interesting to note that Barnewold and Lottermoser
found a close correlation between implementation of digi-
tal technologies, the run-of-mine production rate, and the
parent companies’ revenue. This is an indication that the
larger and high-volume producers have been leading the
implementation. Barnewold and Lottermoser contend
that, due to IT hardware cost falling sharply in recent years,
digital systems should also be interesting and affordable to
the smaller operations, which is not reflected in the analy-
sis. It is their belief that part of this can be explained by the
affordability of sensors, limiting their ubiquitous deploy-
ment to help monitor equipment, processes and systems
(Barnewold &Lottermoser, 2020). To put it in perspective,
to monitor a pump accurately in a system pressure, flow,
power and density measurements are needed. The cost of
sensors in a slurry system can easily be double the cost of the
pump and drive system alone. We need affordable, accurate
and easily deployable sensors. Only this way can we follow
a trajectory similar to the rapid rise of electronics in the
automotive industry since 1970s. According to Figure 2A,
it is projected that the cost of electronics, as a percent-
age of total car price, will reach 50% by 2030 from 1%
in 1950 (Mathas, 2017 Placek, 2023). Lower-cost sensors