Is predictive maintenance worth setting up before you have clean machine data?
I work at a local pet food factory and we tried rolling out predictive maintenance on the extrusion line last year, only to spend the first month just realizing how patchy our sensor history actually was. Tags had been renamed twice over the years, a motor got swapped without anyone updating the documentation and nobody could agree on what a normal cycle even looked like anymore. Some teams start with vibration, temperature and current data on a few critical assets and build from there while others spend months cleaning data before touching anything predictive. Im trying to find the middle ground where the setup is useful without turning into its own data cleanup project. How much data quality did you need before predictive maintenance started helping on your line?