The Sensor Health predictive model
The Sensor Health predictive model analyzes an asset's sensor readings to help determine the likelihood that the asset will fail. If the likelihood of failure is high, you can schedule an urgent inspection of the machine.
The Sensor Health model continuously monitors the health of a machine or an asset and predicts potential machine faults in real time. The model uses historical sensor data profile values stored in the KPI tables and keeps a running status to determine the current health of an asset. The Sensor Health model can also be used to predict the future health of an asset.
Tip: If there are too many failures (for example, more
than 30% of the days, or multiple times in a day), then instead of
using the KPI tables for training, a user could use raw events from
the event table for training with appropriate filtering or treatment
of noise, if any.