The Path from Data to Actionable Information as a Driver for the Industrial Ecosystem
P REDICTIVE
This then becomes the operational recipe
that is used for the execution of the process.
The prescribed process can be verified with
low latency feedback to ensure it is being
executed correctly. Prescriptive models also
allow us to build digital twins of the process
and simulate the outcome to a greater level
of accuracy.
Predictive analytics can be deployed using a
low latency feedback loop to stop or correct
process parameters if the grinding wheels
are not being used optimally. The possible
causes are as follows:
dressing cycles are not being done at
the correct frequency,
accommodation of material or wheel
differences, or
there is maintenance to be performed
on the machine.
R ESULTS
In production, our experience with this
methodology has allowed the customer to
gain between 10% and 52% savings based on
prescriptive process parameters and
verification of correct process execution.
The savings are illustrated in Figure 11.
Tool management systems will also utilize
the insights to correlate variations in the
wheels to the lot and manufacturer. This
takes the customer from reporting on what
was happening in the manufacturing process
to predictive analytics to identify potential
problems.
These alerts created by defining models that
can live at the enrichment stage of data
analysis are created by analyzing historical
data from multiple streams and joining it
with requisitioning of material and wheels to
the process and geometries being operated
upon. The result is the ability to stop
processes before a wheel or part is
damaged.
Figure 11: Grinding Process Improvement
The MTConnect standard was used in this
project to collect the machine tool and load
sensor data. The controller was a FANUC
controller and the syntactic transformation
was done first with an MTConnect adapter
from FANUC’s FOCUS2 protocol to the data
as key/value pairs and then into the
MTConnect semantic standard. The
remaining stages were performed using
proprietary technology for the statistical
analysis (many open source technologies
P RESCRIPTIVE
Prescriptive analytics allows us to take this
one step further and knowing the wheel
manufacturer, the machine, the wear on the
machine and the material being ground: The
system can determine the optimal process
parameters to get maximum life and
performance out of the equipment and
tooling.
- 38 -
June 2018