Do you feel you are getting the most out of your current condition monitoring programs?

In our first blog introducing Relialytics, we eluded to our use of a combination of text and numerical data and data visualisation to analyse condition monitoring data and reduce your on-site data analysis workload.

But why use text?  Numerical data provides you with all of the information you need to manage your equipment, doesn’t it?

Well no.

A common rule of thumb is that 20% of the data produced by a business is in structured form (e.g. data warehouses where data is stored to a plan) and mainly numerical, while 80% of the data is unstructured (e.g. stored on servers without a plan) with a significant amount of text.

If you only study numbers, you are not paying attention to the vast bulk of your data and are probably missing out on valuable operational insights, costing you money.

A lot of the time, the data generated from a condition monitoring program is developed by professional laboratories and inspectors.  Each of the samples or reports produced by these professionals contains free-text detailing their thoughts, recommendations and conclusions. The text provides context to the numbers.

Text is important - but who has the time to read it all.  But if you don’t read it aren’t you are missing out on information that is potentially vital to the well being of your equipment and overall business operation?

This is why, at Relialytics, we are taking the hassle out of having to read all of your sample reports.  Our supervised machine learning models read your incoming reports, compare them with historical data and make decisions as to whether:

  • The results can be actioned without further review because the conditions described by the data are common and you already have satisfactory responses to them.
  • Your site team needs to study the results in more detail because they are outside the norm and human intervention is needed to ensure an appropriate response is developed.

Our current studies show a potential reduction in site-based data review by as much as 80% using this machine learning approach.  Review this case study for more information.

We also provide you with visualisations, built from text-based data, that enable you to view the entire condition monitoring history of a single type of equipment or component on a single diagram. These visualisations provide a new perspective on your data and, because an entire history of many components or equipment items can be seen on one diagram, allow you the opportunity to uncover insights that cannot be found with conventional analysis techniques.

Review this case study for more information.

Contact us at Relialytics to find out more. We would love to hear from you.