Analyzing sensor data, such as vibrations and temperatures over a long time period, can provide information on how wear develops. A worn bearing will become warm, vibrate more, and generate a different running noise.
Such changes can be used to identify worn components long before they fail completely. Replacement can be planned appropriately and will not result in any system downtime. Using the sensor data in this way can save money.