How can unreliable KPIs in condition monitoring systems be systematically improved?

Unreliable KPIs are often caused by unsuitable filter bands or unaccounted interference frequencies. With ibaAnalyzer, both trends and raw data can be analyzed in detail to identify anomalies. Combined with ibaInSpectra, calculation rules can be precisely adjusted. This results in stable and meaningful condition indicators for wear assessment.

How can speed-dependent effects be properly considered in vibration analysis?

Speed-dependent effects can be identified by correlating KPIs with operating parameters. ibaAnalyzer enables X/Y analysis to reveal relationships between speed and KPI behavior. Based on these insights, ibaInSpectra allows precise adjustment of frequency bands. This prevents distortions in specific operating ranges.

How do interfering frequencies affect condition assessment and how can they be eliminated?

Interfering frequencies can significantly distort KPIs if they fall within defined analysis bands. ibaInSpectra allows identification and exclusion of these frequencies from KPI calculations. Additional filtering methods can be applied to remove dominant frequencies from raw data. This significantly improves the reliability of condition indicators.

How can alarm thresholds be reliably defined without a long learning phase?

Alarm thresholds can be defined using simulation-based analysis of historical data. ibaInSpectra enables playback-based calculations to determine expected KPI ranges. ibaAnalyzer supports validation against real datasets. This allows robust threshold definition without requiring a lengthy learning phase.

How does combining process and vibration data support predictive maintenance?

Combining process and vibration data enables a holistic assessment of machine condition. ibaPDA ensures synchronous data acquisition, while ibaInSpectra provides vibration-based KPIs. This integration helps identify indirect influences and external sources of faults. It forms the foundation for predictive and targeted maintenance strategies.

Optimization of a Condition Monitoring System

ibaInSpectra and ibaRotate

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.

Business Benefits The Technology Used Functions More Success Stories

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.

Automated calculation

of characteristic values

Optimum understanding

through combination of process and vibration analysis

Root cause analysis

for proactive maintenance

The Technology

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.

An x-ray style rendering shows the internal components of an electric motor.

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.

A diagram shows data flow from an electric motor to a computer, leading to analysis and user adjustments.

Test the iba System Now!

Download the demo of ibaPDA and ibaAnalyzer and start optimizing your processes today. Try it free of charge for 30 days!

Request a ibaPDA-Demo License

The unique combination of process and vibration analysis makes it easy for me to understand and analyze the vibration behavior, even with complex processes.

Christian Reinbrecht
Product Manager, iba AG

Used Functions

The following functions of the iba system are used in this success story.

Explore More Success Stories

Read more inspiring success stories of our satisfied customers.

Frequently Asked Questions (FAQ)

How can unreliable KPIs in condition monitoring systems be systematically improved?: Unreliable KPIs are often caused by unsuitable filter bands or unaccounted interference frequencies. With ibaAnalyzer, both trends and raw data can be analyzed in detail to identify anomalies. Combined with ibaInSpectra, calculation rules can be precisely adjusted. This results in stable and meaningful condition indicators for wear assessment.
How can speed-dependent effects be properly considered in vibration analysis?: Speed-dependent effects can be identified by correlating KPIs with operating parameters. ibaAnalyzer enables X/Y analysis to reveal relationships between speed and KPI behavior. Based on these insights, ibaInSpectra allows precise adjustment of frequency bands. This prevents distortions in specific operating ranges.
How do interfering frequencies affect condition assessment and how can they be eliminated?: Interfering frequencies can significantly distort KPIs if they fall within defined analysis bands. ibaInSpectra allows identification and exclusion of these frequencies from KPI calculations. Additional filtering methods can be applied to remove dominant frequencies from raw data. This significantly improves the reliability of condition indicators.
How can alarm thresholds be reliably defined without a long learning phase?: Alarm thresholds can be defined using simulation-based analysis of historical data. ibaInSpectra enables playback-based calculations to determine expected KPI ranges. ibaAnalyzer supports validation against real datasets. This allows robust threshold definition without requiring a lengthy learning phase.
How does combining process and vibration data support predictive maintenance?: Combining process and vibration data enables a holistic assessment of machine condition. ibaPDA ensures synchronous data acquisition, while ibaInSpectra provides vibration-based KPIs. This integration helps identify indirect influences and external sources of faults. It forms the foundation for predictive and targeted maintenance strategies.