Optimization of a condition monitoring system
with ibaInSpectra and ibaRotate
of characteristic values
through combination of process and vibration analysis
Root cause analysis
for proactive maintenance
A Condition Monitoring System (CMS) is supposed to monitor the wear of two motors. The CMS calculates characteristic values at regular intervals under certain measurement conditions, which serve as status indicators for the individual machine components, and records these as a long-term trend.
After a test phase it became apparent that the standard calculations and settings of the CMS did not provide reliable characteristic values for wear detection due to the conditions on site. This phenomenon now had to be investigated. With the iba tools for vibration analysis, the recorded trends and raw data were therefore analyzed more detailed in order to configure the Condition Monitoring System to deliver reliable characteristic values.
In the long-term trends, the characteristic values increased at some points in time by a factor of more than 10. Each characteristic value was affected at different times. A temporal correlation could not be recognized. However, the representation of the characteristic values as an X/Y diagram in ibaAnalyzer showed clear conspicuities. The calculated characteristic values were displayed sorted according to the speed at which they were recorded. It could now be seen in the X/Y diagram, that certain characteristic values were increased, especially in specific speed ranges.
Therefore, the raw data of a DAT file were examined more closely for speed-independent frequencies. In fact, a high frequency including two harmonics can be recognized in ibaRotate. Although the cause was found, it was not possible to resolve the problem technically.
With the help of the marker function in ibaRotate it could be proven that the bands used to calculate the parameters intersect this high frequency or one of the harmonics - exactly in the speed ranges where the characteristic values were exceeded. The measurement conditions and calculation rules of the characteristic values were then optimized in that way that this frequency was no longer within the bands that were used for the calculation of the parameters. Finally, these adjustments required the alarm thresholds to be checked and reset. In the next step, the current states of the monitored components were therefore analyzed on the basis of the raw data. For a better overview, the speed-independent frequency and its harmonics have been eliminated using the Vold-Kalman filter. Within the scope of the analysis, an additional insight could be obtained: In the waterfall display without the dominant frequencies, a developing damage was detected on another machine from which a low vibration level had been transferred to the monitored machine.
Finally, it was necessary to determine in which range the characteristic values were to be expected from the new calculations. A simulation with ibaInSpectra based on a playback of the measured data was helpful here. With the help of the simulation, the alarm values could be reliably calculated without a learning phase. With the configuration customized this way, the Condition Monitoring System is optimally configured for the customer‘s system. The system now delivers reliable characteristic values and enables the validation of the current status.
"The unique combination of process and vibration analysis makes it easy for me to understand and analyze the vibration behavior, even with complex processes."
Product Manager, iba AG
ibaAnalyzer-InSpectra can now be added to utilize the powerful ibaInSpectra library offline. Analysis configurations can first be designed and tested offline within ibaAnalyzer and transfer to ibaPDA for real-time vibration monitoring.
Explore how iba helps you to understand and master the growing technological complexity of automated processes and mechatronic systems