The new edge computing capabilities ( Edge computing ) that allow to have process monitoring equipment analyzing data directly at the machine, means that they can be generating maps of “normal” operation at all times and can predict, from calculations involving statistical studies, Machine learning ( Machine learning ), among others, when the behavior of one of its components is outside a certain expected range, ie, when an abnormality occurs.The importance of generating a baseline for all the equipment in the plant cannot be overlooked, hopefully in its new state, or at least just receiving complete maintenance.
This of course is not new, if we see it as specific measures in time carried out in preventive maintenance routines. Where every industrial plant should go is to include the ability to monitor all these variables in real time and thus identify not only behaviors at a single moment in time (a photo of the status of the equipment), but also to know their status under different points of operation, either at production peaks or during low load points. The different providers of this type of monitoring equipment also offer the software tools that allow analyzing the overall state of the components of a plant and increasingly point to the fact of allowing prediction, based on the data, when the failure will occur of a critical piece.
Industrial Development Era
Developments in Industry are focused on making machines themselves detect these levels of “anomaly” and can autonomously decide if they need to start a process for requesting a new part, thereby initiating industrial solutions. Not only based on working hours, but on information gained with data obtained during its operation in its life cycle. And even better: the machines will be able to start detecting their optimum operating points from the point of view not only of productivity, but also of energy consumption, wear of critical parts, among others.