The Path from Data to Actionable Information as a Driver for the Industrial Ecosystem
The raw data, Figure 8, is almost unintelligible because of the high frequency fluctuations in the
load over a large number of parts and machine cycles. The controller data is also translated into
MTConnect and the machine states are used to identify the periods of time the machine is
engaged in grinding a part. This is illustrated in Figure 9 where data are constrained to these
periods (the colors only indicate a change of parts).
Figure 9: Individual Production Bands
In Figure 10, we compose the loads with the process periods and utilize some statistical analysis
and signal processing during data enrichment, we are now able to recognize the load patterns
and begin to identify periods when the loads are increasing in an orderly fashion (right-hand side)
and in a highly variable way (left). The wheel changes are identified as bands, where each color
change indicates a new wheel.
Figure 10: Loads Composed with Wheel Changes
The enrichment process will also identify the process parameters used as well as the dressing
cycles to compute the frequency and duration against the process plan. When the process is not
being performed correctly, the ecosystem integration will feedback the anomaly to the process
planning and maintenance systems indicating something is wrong with the machine or the wheel.
IIC Journal of Innovation
- 37 -