Skip to main content
Captia Technology

What is Predictive maintenance

Definition

What is Predictive maintenance?

Strategy that anticipates equipment failures using models over vibration, temperature or consumption data. Captia AI applies it over data ingested by Connect and turns alerts into executable work orders through Service.

Predictive maintenance anticipates equipment failure by analysing operating data (vibration, temperature, consumption) to intervene just before it happens. Unlike corrective (repair after failure) and preventive (calendar-based) maintenance, it acts on the real condition of the asset. Its economic case rests on avoided unplanned downtime, not on fewer interventions.

What predictive maintenance is

Condition data is captured continuously through sensors connected to the equipment, part of the industrial IoT, and models ranging from statistical thresholds to machine learning and anomaly detection learn the normal behaviour of the asset and flag deviations; advanced models estimate remaining useful life. The alert only pays off when it becomes a scheduled work order.

For an SME the candidates are the few assets whose failure stops production: compressors, line motors, critical pumps. Instrumenting them is a bounded project, the scope of prediction and alerts, and the return comes from avoiding unplanned downtime, which shows up directly in the availability factor of OEE.

Related terms

Related solutions

How we apply this concept in practice:

Frequently asked questions

What is the difference between preventive and predictive maintenance?
Preventive maintenance intervenes by calendar or usage hours without looking at the real state of the equipment: it avoids some failures but replaces parts that still had life. Predictive maintenance intervenes based on the measured condition of the asset, estimating when it will fail.
What data is used for predictive maintenance?
Most commonly vibration (bearings, misalignment), temperature, pressure, electrical consumption and oil analysis. The right signal depends on the failure mode to anticipate. Data is captured with sensors on the equipment and analysed against the history of normal operation.
Do we need to digitalise the whole plant first?
No. The reasonable starting point is the critical assets: the few machines whose failure stops production or is expensive. Instrumenting them and setting up condition monitoring is a bounded project whose results justify (or not) extending the approach.
Does predictive maintenance replace preventive inspections?
It complements rather than replaces them. Some components and requirements (safety, regulation, manufacturer warranties) keep scheduled inspections. Predictive maintenance applies where condition is measurable and failure is costly, and lets you space out the remaining preventive work.

Keep reading

This term belongs to the scope of Captia AI. You can find every other definition in the full glossary.