Machines in production must be maintained. Besides the question of how to do maintenance the best way possible stands, one other important question: When to do maintenance and when to cause interruption of production.
Traditionally, recommended maintenance timelines get provided by manufacturers or things only get repaired if they break. If done this way, the timing of maintenance arises from the experience of the manufacturers. Regularly, planed and time-based maintained is called Preventative Maintenance. Its aim is to take action before things break. Fixing things, only after they already went broke, is called Reactive Maintenance.
Even though those ways of maintenance are most common, they still involve some problems. For Preventative Maintenance lots of planning, record keeping and scheduling has to be done. Also, maintenance might be done too often, because the risk of failure would be fatal if not done enough. This leads to more downtime than really necessary. For Reactive Maintenance unexpected interruptions caused by production failure can also lead to unnecessary downtime.
Traditional means of maintenance, lack a perfect timing. If done too late, things might get worse and unplanned downtime arises. If done too often, losses in production result of unnecessary maintenance stops.
This is where Predictive Maintenance comes in handy. It is only performed when it’s required.
But how is this done? Sensors gather information about the manufacturing process in real-time to track conditions. Based on this Information, the status of production can be monitored. This way you are able to tell when maintenance will be necessary and therefore only perform maintenance when it is required. This way maintenance isn’t done too often, but also, more importantly, not too late.