Abstract: Manufacturing operations are dependent on continuous operation of certain critical assets. When a critical asset fails without warning, manufacturing stops. The cost of such unplanned downtime could be significant. For instance, lost production might cost as much as millions of dollars in the oil and gas industry. Failure of critical assets might result in loss of valuable raw materials. In the worst case, unexpected equipment failure can have even more serious consequences including environmental damage and health hazards to operators and to the public.

Predictive maintenance (PdM) is a relatively new maintenance approach based on monitoring the health of critical assets and predicting failures ahead of time. PdM enables plant operators to plan maintenance operations based on asset health as well as remaining useful life estimates of such assets. By minimizing unplanned asset downtime and optimizing maintenance schedules, PdM programs yield higher production capacity and reduced maintenance costs.