During the past decade, many industrial manufacturing organizations have implemented condition monitoring programs on production machinery. The objective is to track various condition and performance measurements related to a specific machine (or component of a machine, e.g., electric motor) over time in order to deduce pending machine failures.
By taking proactive action to avoid unscheduled production downtime, these organizations expect to reduce the costs of maintenance, direct labor, and waste related to the conversion of raw materials into finished goods. Such action presupposes a capability to diagnose pending failures within a sufficient lead time (depending on the contemplated action). However, development of condition monitoring technology has, in the past, focused on diagnostic aspects rather than on predicting residual machine life.