Nearly everybody understands the benefits of predictive maintenance. Infrared analysis, ultrasonic testing, and insulation resistance testing are among the means of performing predictive maintenance.
We usually think of repair as something we do when maintenance fails to prevent a failure. But that is not always so.
Consider this scenario. The plant has six chemical reactors, each of which uses a sensor that’s exposed to the process. The reactors make pretty much the same thing, so the environment in each can be considered the same.
One day, Reactor Three starts making bad batches. After several hours, the responding technician isolates the problem to a failed sensor. He was able to determine this after, in desperation, calling the equipment builder (who said to check the PLC for fault codes). The tech replaces the sensor, and now Reactor Three makes good batches.
From this episode, a smart maintenance manager would predict that the sensors will soon fail on the other five reactors. He orders replacement sensors and schedules replacement during a window that operations makes available. This costs far less than a single spoiled batch.
In your facility, what repairs might predict that others are needed?
