1
25-065
Optimizing Mill Performance: Real—Time Grinding Insights
Eduardo Nunez
Molycop Canada
Dante Garcia
Molycop Peru
Joshua Morales
Molycop Chile
Karl Gugel
Molycop USA
Peter Czel
Molycop Mexico
Boris Sullcahuaman
Molycop Peru
ABSTRACT
Real-time operational excellence is achieved through the
seamless integration of various components. Key aspects
include precise instrumentation, real-time data collection,
data analytics, and an efficient optimization platform.
Additionally, to ensure success, the real-time optimization
platform must be incorporated into the site’s internal pro-
cess control system (Nunez, 2022).
This paper provides an overview of practical lessons
and best practices for autogenous (AG), semi-autogenous
(SAG), and ball mill (BM) grinding. It highlights the
implementation of advanced instrumentation, specifically
vibration sensors on the rotating mill shell. The paper also
delves into the application of advanced analytics necessary
for generating new measurements that accurately represent
real-time grinding activity in the mill. These Advanced
Analytics Measurements (AAM) are compared with well-
known grinding curves to determine the optimal grinding
point. The discussion focuses on key industry opportuni-
ties using real-time mill shell vibration results to derive new
correlations and optimize the grinding process.
INTRODUCTION
The process of reducing ore rocks into smaller particles
to liberate valuable minerals, known as grinding, has sig-
nificantly evolved with the advent of new instruments and
more efficient process control strategies. However, the mul-
tivariable and continuously changing nature of the grinding
process requires real-time understanding and monitoring of
how the mill is operating. Luckily, this is now possible due
to advancements in computing power which in turn yields
much faster processing of large data sets and application
of artificial intelligence (AI) techniques to provide novel
insight into the grinding operation as it occurs every rota-
tion of the mill.
In the last decade, multiple technologies have been
introduced. One of the most prominent systems is based
on a single microphone or an array of microphones. This
technology is able to provide an idea of when the mill is
empty or full in terms of impacts (‘metal to metal’ impacts,
for liner protection). The main limitations of this solution
are the lack of trajectory estimations and the ore location
inside the mill. On the other hand, DEM simulations have
been used to successfully understand the charge trajectory
and its effects on the grinding operation. The challenge
with DEM simulations is that they typically include only
one operational condition, e.g., different rotational speeds,
with a computing time that could take several hours for
one operational change. This prevents its use in real-time
optimization.
One of the most promising instruments for viewing
what is actually happening inside the mill is the installation
of multiple high resolution vibration sensors magnetically
mounted to the mill shell. These in turn provide vibration
profiles that open up a new world of analysis and observa-
tion due to their enhanced sensitivity to various conditions
previously undetectable occurring in the mill. These condi-
tions include changes in ore size, hardness and competency
the effect of different ball charges, filling levels, new/worn
liners slurry consistency and rotational speed.
These measurements, combined with existing instru-
mentation, create a vast new source of data that can now
be processed in real-time thanks to much faster data pro-
cessing as well as advancements in analytics and artificial
25-065
Optimizing Mill Performance: Real—Time Grinding Insights
Eduardo Nunez
Molycop Canada
Dante Garcia
Molycop Peru
Joshua Morales
Molycop Chile
Karl Gugel
Molycop USA
Peter Czel
Molycop Mexico
Boris Sullcahuaman
Molycop Peru
ABSTRACT
Real-time operational excellence is achieved through the
seamless integration of various components. Key aspects
include precise instrumentation, real-time data collection,
data analytics, and an efficient optimization platform.
Additionally, to ensure success, the real-time optimization
platform must be incorporated into the site’s internal pro-
cess control system (Nunez, 2022).
This paper provides an overview of practical lessons
and best practices for autogenous (AG), semi-autogenous
(SAG), and ball mill (BM) grinding. It highlights the
implementation of advanced instrumentation, specifically
vibration sensors on the rotating mill shell. The paper also
delves into the application of advanced analytics necessary
for generating new measurements that accurately represent
real-time grinding activity in the mill. These Advanced
Analytics Measurements (AAM) are compared with well-
known grinding curves to determine the optimal grinding
point. The discussion focuses on key industry opportuni-
ties using real-time mill shell vibration results to derive new
correlations and optimize the grinding process.
INTRODUCTION
The process of reducing ore rocks into smaller particles
to liberate valuable minerals, known as grinding, has sig-
nificantly evolved with the advent of new instruments and
more efficient process control strategies. However, the mul-
tivariable and continuously changing nature of the grinding
process requires real-time understanding and monitoring of
how the mill is operating. Luckily, this is now possible due
to advancements in computing power which in turn yields
much faster processing of large data sets and application
of artificial intelligence (AI) techniques to provide novel
insight into the grinding operation as it occurs every rota-
tion of the mill.
In the last decade, multiple technologies have been
introduced. One of the most prominent systems is based
on a single microphone or an array of microphones. This
technology is able to provide an idea of when the mill is
empty or full in terms of impacts (‘metal to metal’ impacts,
for liner protection). The main limitations of this solution
are the lack of trajectory estimations and the ore location
inside the mill. On the other hand, DEM simulations have
been used to successfully understand the charge trajectory
and its effects on the grinding operation. The challenge
with DEM simulations is that they typically include only
one operational condition, e.g., different rotational speeds,
with a computing time that could take several hours for
one operational change. This prevents its use in real-time
optimization.
One of the most promising instruments for viewing
what is actually happening inside the mill is the installation
of multiple high resolution vibration sensors magnetically
mounted to the mill shell. These in turn provide vibration
profiles that open up a new world of analysis and observa-
tion due to their enhanced sensitivity to various conditions
previously undetectable occurring in the mill. These condi-
tions include changes in ore size, hardness and competency
the effect of different ball charges, filling levels, new/worn
liners slurry consistency and rotational speed.
These measurements, combined with existing instru-
mentation, create a vast new source of data that can now
be processed in real-time thanks to much faster data pro-
cessing as well as advancements in analytics and artificial