Real Time Vibration Monitoring

The ibaInSpectra Expert module monitors and analyzes vibrations in the frequency spectrum, which was generated using an FFT analysis. It can be used for a wide range of applications thanks to the high level of flexibility and versatility of the module.

Versatile Module for Individual Vibration Analysis

In the expert module, the frequency bands to be monitored can be freely defined, both statically as well as dynamically depending on other measurands. The following parameters are determined for each frequency band as a result of the analysis: Peak, RMS (Root Mean Square), Peak frequency, Phase and Crest Factor. Freely configurable characteristic values can be calculated based on these parameters. In addition, it is possible to define two limits (alert, alarm) for characteristic values or individual band parameters. In addition to the values from the frequency domain, additional values are determined in the signal’s time domain, such as minimum, maximum, average, RMS or Crest factor.

The calculations for the analysis can be individually adjusted, saved as profiles, and reused. In particular, the sensor type, type of spectrum and FFT calculation parameters, such as the number of samples, window shape or overlapping factor, can be set. Different methods of averaging are also available, such as the option of detrending in order to compensate for a slow drift of the measured value.

Order Analysis and Envelope Calculation

In the calculation profile of the ibaInSpectra Expert module, the user can select a speed signal for re-sampling the input signal depending on the rotation speed. This speed signal can be a pulse train, a pulse counter or an absolute speed value. With resampling, the order analysis for changing speeds is displayed in a much more comprehensive way than with rescaling the x-axis. What makes ibaInSpectra unique is that it filters out the aliasing effects that occur during resampling, regardless of the rotation speed.

Moreover, signals can be preprocessed with the ibaInSpectra Expert module. Thus, the vibration signals can be filtered and mathematical precalculations can be done, for example, an envelope calculation with freely configurable bandpass filters. Parameters for calculation profiles are listed in the table.

Alert and Alarm

Alert and alarm thresholds can be configured in each defined frequency band to ensure reliable alarming. As soon as a limit value is exceeded, corresponding binary signals are activated.  These signals can also be made available to other systems for signaling or for interventions in process control via the output interfaces of ibaPDA.

The Auto-Adapting module can be used to detect damage to machines, gearboxes and motors as well as quality-related changes in process vibrations already at a very early stage. The self-learning InSpectra module offers optimal protection for systems through automatic monitoring in real time using learned reference values.

Detecting Damage at an Early Stage

The auto-adapting module is able to use a series of spectra to learn what the ideal spectrum should look like. The so-called reference spectrum can be learned for various process states, which, for example, relate to different speeds, materials or load areas, etc. Since ibaInSpectra is seamlessly integrated into ibaPDA, the full ibaPDA connectivity is available to acquire all possible process signals in a system and to be used to define the states. 

Comparing with the good condition and the quantification of the deviation allows the user to detect changes in process behavior at a particularly early stage before quality issues occur. The Auto-Adapting module also identifies the ranges with the biggest deviations.

Learning and Monitoring

Instead of having to manually configure a frequency analysis for certain ranges, all ranges of the spectrum are considered in the auto-adapting module. In the process, the spectrum can be individually divided into any number of ranges. During the learning phase, characteristic values are calculated and learned across all ranges of the spectrum, and a reference spectrum is determined from these values. In the monitoring phase, the Auto-Adapting module compares the current spectrum with the reference spectrum. Since the Auto-Adapting module is based on the ibaInSpectra Expert module, it offers the same extensive options for configuring FFT calculation parameters in profiles and signal preprocessing.

Permitted deviations can be configured individually. For alerts and alarms, percentage limits can be defined that relate to a maximum or average reference spectrum.

Convincing characteristic Values

The auto-adapting module calculates meaningful characteristic values for every state: 

• Absolute delta: Total difference between the actual spectrum and the threshold values from the reference spectrum.
• Relative delta: The relative percentage difference between the actual spectrum and the threshold values from the reference spectrum.
• Center frequency, relative difference and peak value for ranges in which the biggest differences occur between the actual and reference spectrum.

The ibaInSpectra Orbit module monitors shaft motion in plain bearings, among other things, enabling reliable monitoring and evaluation of machine condition. For a stable calculation of the parameters, even at different speeds, the input signals are sampled depending on the speed.

Calculation of Speed-Dependent Characteristic Values

This ensures not only reliable results for all conditions but also allows an averaging of the orbit over several revolutions. An average value can be calculated (linear) or the maximum values of the corresponding rotation angle are evaluated (peak hold).

The most important characteristic values calculated by the ibaInSpectra Orbit module are Orbit counter (number of calculated orbits), X/Y (current shaft position), Centerline X/Y (position of the shaft center), Peak to peak max  (S(p-p)max, maximum distance between two points in the orbit), Peak to peak max angle (angle of the S(p-p)max), Peak to peak max shaft angle (rotation angle of the shaft to which S(p-p)max occurs relative to the phase reference), Eccentricity (eccentricity of the shaft), Distance to clearance (minimum distance of the shaft to the clearance circle (bearing shell))

Better Insight Thanks to a Flexible Visualization

The orbit view offers various possibilities to adjust the visualization of the shaft motion individually. So the shaft motion can be visualized including phase reference. In addition, the movement history of the shaft center can be displayed over a time period. Another option is e.g. to display different acceleration levels during the start-up of the machine. To improve the understanding of the movement, the clearance diameter, rotation direction, sensor positions and the position of the phase reference can be displayed.

Offline Detailed Analysis With Playback Function

For a detailed offline analysis the recorded data can be opened together with the calculation profiles in ibaAnalyzer-InSpectra. Shaft motions can be analyzed subsequently and can be compared with other process data to recognize correlations. Additionally, the behavior can be reproduced at certain points of time, using the playback function.

The ONNX module can be used to apply AI models in ibaInSpectra. This allows easy deployment of ONNX models in ibaPDA and live monitoring of any measurement data.