On May 16th and 17th 2017, ARC Advisory Group hosted its European Industry Forum (EIF) in the Meliã Hotel Sitges in Spain. The European Industry Forum is part of ARC’s successful series of worldwide conferences in the USA, India, China, and Japan.
The forum featured a presentation by Thomas Lantermann of Mitsubishi Electric and Thorsten Schroeer of IBM on „Plant-wide Predictive Production Analytics Informed by Condition Monitoring at the Machine Level“.
Plant-wide Predictive Production Analytics
We all know about the possibilities of CMS (Condition Monitoring Systems), but often without realizing the importance - usually just before it all goes wrong!
Many production sites do not even realize there is a problem until the smoke starts pouring out! But before things have got that bad there was probably a period of when things were working but perhaps the machine or device was feeling a little warm or the occasionally bearing squeak started to become more frequent. However, unnoticed by most people, was an irregular vibration which had started months earlier. And of cours, there was a time back in the good ol’ days when everything was new and worked perfectly.
As you can guess, the earlier you can identify that something is going wrong, the better chance you have to plan and act; to arrange for a scheduled shutdown, and order replacements for suspected faulty or worn parts. Failure to do so means the impending catastrophe could be worse than necessary but also the subsequent loss in production/operation, with all the related financial penalties, could have been avoided. This kind of prewarning is delivered by CMS.
But that is the view of a single sensor on a single device. What could the consequences be? Is there an impact on the next machine in the process? Could the previous process experience a bottleneck? If it does, what happens to its previous process and the process before that? And what happens to the resources employed or keeping that all important delivery commitment?
The local PLC or controller will inevitably only be “aware“ of the machine it controls. Data can easily be passed to a MES system, so some of the questions can be answered; but again, how do we look ahead to potential issues and risks because MES only deals with the “facts of now”?
This is why the upper cloud like layer starts to become important especially as it can employ advanced predictive technologies such as AI to evaluate potential consequences and advise on the best course of action in order that the manufacturer can keep his commitments.
Maybe you know the old saying “rubbish in = rubbish out”? In order to make decisions and take actions, we need models and frameworks that help us to put our decisions in context. We also need data that is relevant and timely. But to get this "good" data we need to select the right data and source quickly, which means we need a barrier-free or open environment where the FA-IT interface becomes a conduit for accelerating our decision making.
In May 2018, ARC will host its European Industry Forum (EIF) again. For further details, please contact your Client Manager or email email@example.com.