top of page

Data Quality Over Quantity


The days of measuring vibration data directly at the machine once every couple of months are starting to disappear. There is still huge value in undertaking this activity, but it’s quickly being replaced by a wide range of IIoT sensors (Industrial Internet of Things). 



Seemingly, one advantage is gathering data at a vastly improved rate. Rather than a single reading every quarter, data can now be captured multiple times every hour. The advantages are obvious: much greater visibility on developing faults and a trail to follow to track existing known issues.


But there are also some disadvantages. IIoT devices are battery-powered, so there is a balance to be achieved in data quantity, quality, and battery life. Typically, the faster you acquire data, the lower the battery life. If not, then the quality of the data might suffer. 



It’s All a Balancing Act

In the case of condition monitoring, this might be reducing the FMax, lowering the resolution or only sending the spectrum rather than the full-time waveform.  This can sometimes be against the desires of Vibration Analysts, who want the cleanest, highest resolution appropriate for the application.


Another problem that exists with such large data sets; how do you quickly sift through all these data sets to find what you need to look at? Is too much data inhibiting you from viewing the pertinent data you need to analyse?






Get the Data You Deserve

If using these IIoT systems, make sure that you select one that provides a good interface to search through data sets.  Data views that trend key features in the vibration information so that you can quickly detect anomalies and go straight to the corresponding data set.  Focus on systems that provide a good frequent snapshot to understand a fast change in trends, but still measures high quality waveforms and spectrums suitable for analysis.






Accuracy is Still Key

One other factor to assess is the accuracy of the data you are analysing. Ask yourself these questions; How accurate is the frequency? Is the FMax suitable for your application? Is the amplitude showing the true energy?  Is the data fully processed and ready for me analyse?






There are many tools that aid in understanding your machine's health. Choose one that works best for your needs, but never underestimate how important data quality is to achieving an accurate assessment.

Kommentare


bottom of page