In the complex field of asset management where many elements (e.g., human and tangible and intangible resources) must interact with each other, a large amount of data is collected and accumulated. The ability to enable that data to forecast or better reduce variation is also of great importance to improve asset sustainability.To improve asset sustainability (e.g., reducing waste, increasing energy and resource efficiency, a good maintenance policy, based on date, is necessary to ensure system reliability, reduce cost, avoid downtime, and maximize the useful life of a component.
So, if the future behavior of assets can be approximated this knowledge may help decision making. To extract useful information from manufacturing data, different techniques can be used. A computing system, can create knowledge that may help a better decision
The earliest maintenance strategy is unplanned maintenance (run to failure), in which no maintenance will occur until a machine breakdown happens. In this situation, the useful life of a machine component may be increased to some extent, but unplanned downtime is unavoidable. Preventative maintenance, more widely used policy, inspects and maintains the components with periodic intervals to prevent unexpected machine breakages. However, the regular inspection/maintenance practice can incur early disposal of good components, long suspension time and therefore high maintenance cost. Because of these pros and cons, a maintenance engineer often confronts with the tradeoff situation: they need to choose between maximizing the useful life of a component (unplanned maintenance) and maximizing uptime (preventive maintenance).
While unplanned and preventive maintenances have the tradeoff scenario, predictive maintenance (PdM) is a promising technique that has an ability to break the tradeoff by maximizing the useful life of a component and uptime simultaneously. PdM is designed to monitor the condition of components, and the prediction when equipment will fail. It means that the future behavior/condition of components can be approximated, which will help to optimize maintenance tasks (e.g., prognostic health monitoring). Accordingly, the machine downtime and maintenance cost can be reduced significantly while making the maintenance frequency as low as possible.
Modeling techniques good also be applied to operation data i.e. their performance, remaining useful life of components.