SAP Predictive Maintenance and Service, cloud edition: Vibration Analysis add-on
Vibration Analysis add-on is available with the Feature Pack 4 of SAP Predictive Maintenance and Service, Cloud Edition thereby significantly enhancing the solution capabilities.
Vibration Analysis is a large component of predictive maintenance programs for machines with rotating components. It is used in conjunction with other process-related measurements, such as machine speed, load, power, flow, pressure, temperature measurements, etc.
Vibration Analysis helps to evaluate the condition of a machine by detecting developing faults within the machine components and avoiding failures. Maintenance or Service Engineers can minimize unplanned downtime by scheduling needed repairs during normal maintenance shutdowns.
Below, you can see the tile design for Vibration Analysis functionality. A Maintenance or Service Engineer having expertise in Vibration Analysis can be assigned to the relevant Vibration Analysis roles and can then access these tiles in addition to the other Condition Monitoring Tiles.
To establish and utilize effective vibration-based predictive maintenance, a complete knowledge of each machine component and how they interact within the machine is necessary. The basic principle is that all rotating machines have common components, corresponding vibration frequencies, failure modes and characteristics.
The Vibration Analysis add-on allows to flexibly setup a database of key rotating components for a machine and then associate the unique Vibration Sources (frequencies), Diagnosis (failure modes) and Component Property (characteristics) via intuitive Fiori like User Interfaces.
Each Machine also has unique operating dynamics. Vibration Analysis add-on allows to create a template for a Machine type to capture it’s operating dynamics and also other details like Sensor locations, etc. as shown below.
The system can calculate the rotation speeds of each machine component based on configuration done via the formula editor (as shown below).
We can start with a driving component (like motor) and specify the formula for calculation of driven component (like bowl) based on relevant properties for intermediate components (like pulley diameter, belt length, etc.) . Based on this configuration all rotation speeds can be calculated and correlated to the appropriate frequency in the frequency spectrum.
Based on these templates, machines can be efficiently on-boarded (for example: from a business system like ERP/CRM/etc. This includes the Machine Part lists and associated design information (for example: motor speed, pulley diameter, belt length, etc.).
Once on-boarded the part lists can be viewed in Vibration Analysis Add-on. It is also possible for an Engineer to maintain certain amplitude design Thresholds of the Machine component frequencies (if these amplitude threshold ranges are known by design). Also, they can be changed as and when an Engineer gains more information from experience of the Machine.
Once this configuration setup is complete and the machine is on-boarded into the Vibration Analysis add-on, the most important task is to gather data from the Vibration Sensors. It basically means obtaining a full-spectrum vibration signature at different components of that machine.
The Vibration Signature can be recorded as either Time Waveform or frequency Spectrum files.
If the Signature is recorded as Time Waveform, the Vibration Analysis add-on can convert it into frequency Spectrum via Fast Fourier Transformation algorithm.
Once the frequency Spectrum is captured, the Vibration Analysis add-on provides the following mechanisms to create good baseline references which can be used to compare against the recent frequency spectrum, and hence get a better idea of the machine health.
1. Fingerprint based on frequency Spectrum: It is possible to run a machine under optimal conditions and classify the corresponding frequency spectrum as Fingerprint for that Machine. And as you see below the Fingerprint can then be compared against the recent frequency spectrum to see the violations.
2. Envelop Threshold Curve based on frequency Spectrum: It is possible to create an envelop curve for a complete Vibration Spectrum by entering some tolerance limits. This can be used for future Spectral comparisons to detect possible threshold violations.
Finally a Service Engineer can use the Vibration Diagnosis application to pinpoint specifically the failure mode and identify which Machine Component is degrading. For example: Below you can see a Frequency Spectrum recorded at a Sensor2 of a particular machine. The application highlights that the design threshold for a Motor frequency has been violated and the possible diagnosis (failure mode) could be looseness.
It is possible for a Service or Maintenance Engineer to further compare this Spectrum against the fingerprint or envelop threshold curve in order to confirm the Diagnosis.
With this SAP has delivered the first specific analysis add-on for Predictive Maintenance and Service, cloud edition. This is a highly relevant analysis for all manufacturers or operators of rotating machines. We see impressive interest from the market already.
As next steps we will integrate additional features like different alternatives for defining vibration amplitude thresholds and for integrating the add-on further with backend systems, e.g. to automatically generate a maintenance or service requests, when vibration issues have been detected with a machine.
We will keep you posted.